In our eleventh episode, Heejae Lim, Founder/CEO at TalkingPoints, joins Natasha Allen to discuss artificial intelligence as a force for good. While discussion surrounding the intersection of education and AI has focused on negative impacts, how can recent advances help facilitate better outcomes for students? Why is it important for teachers and parents to see AI as an ally instead of an adversary?

Go Deeper:

• Navigating the AI Frontier: Legal and Operational Insights Into Generative AI
• What to Expect in Evolving U.S. Regulation of Artificial Intelligence in 2024

Transcription

The below episode transcript has been edited for clarity.

Natasha Allen

Hi everyone and welcome. My name is Natasha Allen. I’m a partner at Â鶹ֱ˛Ą and Lardner in the Silicon Valley and San Francisco offices. By day I’m a corporate M&A attorney, but by night I have the benefit of being the Co-Chair of the AI area of focus, which is under the Innovative Technology sector at Â鶹ֱ˛Ą. In that role it gives me the opportunity to meet with brilliant individuals, such as our guest today, Heejae Lim, and I’ll introduce Heejae shortly.

But just to give a bit of context for the podcast today, I think that we wanted to do an episode that highlights some of the good effects that AI is having on society. Especially given that a lot of things have been happening in this area, we thought that it would be great to invite Heejae Lim, who is the CEO of Talking Points, to talk a bit about her company and what it’s doing with this new technology.

To give a bit of background, Heejae is a founder and CEO of Talking Points. She previously worked in education reform for public schools in the Middle East, and she also served as the chief of staff to a special advisor to the Secretary of Education in the UK. As a consultant at Mackenzie, she advised the Gates Foundation and Teach First UK. Heejae graduated from Oxford, London School of Economics, and has an MBA from Stanford Business School where she was a Arjay Miller Scholar. She was recognized as Ted Fellow, Echoing Green Fellow, Stanford Social Innovation Fellow, Ashoka Changemaker, and Forbes 30 under 30 for Education.

Heejae, you have won a lot of accolades and have done a lot of great work in your career. Can you just give us a little information about what does Talking Points do?

Heejae Lim

Thank you so much for having me, Natasha. So Talking Points is an education technology nonprofit, and our mission is to drive student success by unlocking the potential of families to fuel their children’s learning, with a special focus on underserved communities. So we do this by building meaningful connections between schools, teachers, and families through accessible technology. And what that looks like is a family engagement solution that allows for two-way translator communication between families and schools, as well as capacity building and coaching content around how to work together to really drive student outcomes and student success.

Natasha Allen

That is amazing and a very, very important, I think, way to help the education system. What inspired you to found Talking Points?

Heejae Lim

So I’m originally Korean, well, I guess I’m still Korean, I still have a passport. So I was born in Korea, and when I was eight my family moved to England. I was an English language learner, pretty much for the first time when I moved there, and I saw my mom kind of blossom into a parent spokesperson, parent communicator, translator, and the go-between the school and the parent communities that we were part of.

We went to a public elementary school where it was very heavily Korean immigrant students and families and I saw the impact that my mom’s engagement had on me and my sister’s education. And then, you know, on Monday to Friday, we would go to this public elementary school. And then on Saturdays, we would go to, you know, Korean school with the Korean curriculum, Korean teachers, Korean content, and of course we learned in Korean. And I realized the contrast between Monday to Friday versus Saturdays, even though it was the same friends I had, so same students and same parents, but the interaction between parents and teachers was so different…between Monday to Fridays and Saturdays…and I think that really inspired me to start Talking Points.

And if I think about what my experience of one really shows in terms of how that shows up in academic research, it really shows that family engagement and the extent to which families and schools can build those partnerships for student education is twice as effective at predicting student success than family socioeconomic status. Which is kind of crazy to think about. If you really kind of sit and think about, oh wow, [if there is] a student from a poor family who is highly engaged and can really build this relationship, the school can be more successful than the reverse.

And with that, research really became the genesis of founding Talking Points. It kind of also intuitively makes sense if you think about, you know, a family member or a caregiver or a mom or dad or grandparent, is often that child’s most influential and first teacher. And children spend, you know, more than 80% of their time outside school. So it makes sense that family engagement matters a lot. But you know, there’s not that many blueprints or technology-driven solutions to be able to enable doing that, especially for underserved communities.

Natasha Allen

And so how did you integrate AI, or was AI always a component of Talking Points?

Heejae Lim

AI wasn’t always a component.

So we got started back in 2014, 2015, when I think AI was still in the domain of research institutions and not so much applicable technology. We started using AI back in 2018 when Google released that AutoML model, a machine learning model where you could train using your own data set. So we started using AI to really make our translation quality higher than those, you know, mass market, kind of readily available machine translations. And the way that we do that is through making the translations a lot more education context specific. Because we have a narrow domain, we know that when certain words or phrases are used, it’s more likely to mean something specific than others, and we leverage that to be able to provide the highest quality translations and making the communication that happens on Talking Points a lot more understandable and consumable as well.

Natasha Allen

That’s actually brilliant, because every culture has a different way of communicating, right? And if you don’t hone into that, you know, perhaps my mom who came from Jamaica wouldn’t necessarily understand if somebody’s parent came from Italy, right? Just may have a different meaning culturally.

What are some specific projects or initiatives that you’re currently involved in within the AI space?

Heejae Lim

Yeah, so one really exciting development that we’ve had very recently is we’ve started partnering with OpenAI in releasing what we call a message mentor. So while educators are expected to communicate with families and know that it’s a very important lever to drive change in students, they often don’t get the coaching or the training that they might need or want in how to communicate effectively with families who might look and feel differently to them, or speak a different language, or come from a different cultural background. So what an AI enabled algorithm does is to transform their communication and messages into something that’s positive, that is asset-based, or thinking about families and students as assets to student learning. We developed it based on a dual-capacity building framework, which is a framework for how to do family engagement well. And this kind of mentor already guides the teachers around communicating effectively using AI, and the latest developments in generative AI has obviously really helped a lot with that as well.

Natasha Allen

Absolutely. So always have to ask, is there a way that if people are interested and want to volunteer, or help Talking Points, is there any way that they could do that?

Heejae Lim

Yeah. So one of our goals over the next few years is to really, what we call, get to the tipping point of the K12 education market. So if you’re a parent with a student in the K12 school system please talk about Talking Points. I think that’s the first task to those in the audience. Kind of pause and think about how you, as an adult, can interact more with the school system, learn more about what’s going on, in order to support the students that might be around you, whether that would be your specific child or your family member, and so on. So those are our asks.

Natasha Allen

Brilliant. I want to ask you a few questions. Before we dive into the benefits of AI and challenges, I wo want to focus in a little bit on AI and its use in K to 12 education today.

So I’m sure you’ve heard that most of the discussion around the use of AI and learning has been focused on having a negative impact on learning. Right? So seeing it as more of cheating as opposed to just a different way to learn. What are some of the lesser-known applications where technologies can improve the delivery of education, and not necessarily be considered a negative thing?

Heejae Lim

Yeah, and I think I’m going to just think about and say out loud some AI technologies where they can improve the delivery of education. I don’t know if they are lesser known or more well known, but here are some applications of AI and education which I think will really make a difference.

One is personalizing and customizing according to the student’s needs, or even the family needs, or the teacher needs. I think the latest developments in generative AI really has a huge potential in being able to meet the customer, or the user, where they are. And in this education sense, its educators, its family members, and ultimately students. So I think the way that AI can really tailor to someone’s needs has incredible potential in being able to serve and give, let’s say, real-time feedback about a written assignment, a math test, or even how families are talking about schools and learning to students. So lots of potential there, or even coaching how teachers are teaching and instructing in their classrooms.

I also think about grading as a potential area where AI can really help. Where you can really accelerate the grading and, you know, cut down teachers’ time in doing rote tasks that might really benefit from technology and AI specifically, so that they can focus on the higher needs of students, or areas where that human touch is more important.

I also think about the way that AI can accelerate resource development. So if I think about teaching materials, or coaching, or exercises, or quizzes, I think a lot of that can be generated through AI. Just lots of potential.

I’m just spitting out some of the examples that come to mind immediately for me.

Natasha Allen

Absolutely. But do you think there’s any way that steps can be taken to change the narrative? That it’s actually a good thing in the educational sphere? Or do you think we’re far off from bridging that gap?

Heejae Lim

You know, I think it’s all-around, like change management, right? So change management and changes in mindsets can be hard, but we do see it happening. I don’t think the overall kind of rhetoric or general sentiment is necessarily AI’s enemy, it actually spans the entire spectrum of all the way from AI’s an enemy to AI as an ally. If you talk to educators or school districts, I think the reason that it can be thought of as AI as an enemy is because one, we kind of lack a common understanding of, what do we mean by AI? Like, what are we even talking about when we say what AI is, or applications of AI are? So I think we need to get a common understanding of what is possible and what is not possible, and what technologies are out there. I think that to the extent that we can structure some of the guidance around where AI can be applied versus not, is another good step.

You know, I think about AI as an enemy, and I’m sure when typewriters and the computers first came to be, teachers were thinking, oh gosh, like, they’re typing up their assignments. Like, wow, like they cannot handwrite anymore. They’re typing. Is that cheating? I think about that as an analogy. Or, you know, if you’re doing a report on a topic and, wow, students, instead of going and searching the Encyclopedia Britannica, they went online, like, is that cheating? Not necessarily. So I think it’s really around guiding their behaviors and what’s expected versus not that will be important in shaping that narrative as well.

Natasha Allen

I agree. I think you also touched on it too. Part of it is education, right? People understanding what these technologies can do and not just thinking that they’re for bad.

Heejae Lim

Yeah, and anything new, you know, anything new has fear, right? Like people are driven by fear, it’s scary. So anything new and less understood I think will be automatically bucketed into this enemy bucket. So how can we change that? Through awareness and education.

Natasha Allen

Exactly. So how is Talking Points leveraging recent advances in AI? I think you touched on it a little bit with generative AI, but are there any other improvements that have helped with getting Talking Points to more people?

Heejae Lim

I want to go back to AI and human-based translation. I think it’s always been part of our solution to provide the highest quality, contextually relevant, understandable transition between families and schools, and AI has helped us understand the K12 context to ensure that translations make sense. And the latest developments in large language models have really allowed us to capitalize on the human translator inputs that we put into those models, as well as user-feedback and data from our platform.

So give you a tangible example and maybe this will make you laugh. But during COVID, we noticed a lot of communications from the school districts to the families were related to distributing Wi-Fi hotspots. Wi-Fi hotspots are distributed because a lot of students in the underserved communities did not have Internet access at home. And in order to get on Wi-Fi, to get on Zoom or online learning or do their homework, they required Wi-Fi. Now we have to remember, Wi-Fi hotspot might not necessarily even be a phrase that is easily understood if you don’t know what the concept of a hotspot is. Frankly, I don’t know if my parents would know what a hotspot is. So even for native speakers, it’s a hard thing to understand if you don’t know the concept.

But in Russian, we noticed that existing machine translation models were translating it to, guess what? A warzone.

Natasha Allen

Oh gosh.

Heejae Lim

I think its the historical context that Russian as a language provides, and it’s more ironic to think about before the Ukrainian-Russian war, so we quickly modified our models to say “no hotspot should not mean warzone” in the school’s context. Our educational context always means, I mean 99.8% of the time means, a hotspot is a machine that would allow you to get online and use Wi-Fi. So that’s an example of educational context that we can use with existing AI technology.

And then we think about that as a really important piece of going beyond what standard machine translation offers, because it’s such a critical piece to driving impact for us in improving communication and building those trusting relationships, which is, of course, foundational to really meaningful engagement.

Separately, we are also thinking about how we can leverage AI to really build capacity, to drive and provide just in time. So real-time learning and coaching to both families and educators, and school district staff, based on what we know of them, in giving prescriptive guidance or more mentoring and coaching, based on who the students are and where they are at, and based on the best practices that we know to be true through our own research and external research. We’re really excited about this body of work because, as I mentioned earlier, AI can really help us customize and personalize and meet the needs of individual stakeholders of using technology.

Natasha Allen

So I guess a question is how to help the school districts, right? How do they navigate, you know, any risks associated with AI? Or even fears of risk associated with AI, and fit it within their regulatory frameworks, right? So obviously regulatory frameworks have to evolve, but what is existing now? Is there any advice you can give to administrators or directors of school districts?

Heejae Lim

I think continuing to think about the risks of AI, but also potential use cases, where there can be wins using AI for school districts and administrators.

For example, because it is still somewhat of an uncertain technology, you probably cannot use it for anything around compliance or legal that has higher costs. If things fail, or in areas where you need to be more risk averse. But I think embracing AI in a “what can it do for us” versus “what can it not do” and having, again, those structures in place and the expectations, I think will help us leverage more of what it can do, but also shy away from what it cannot do as well. I think we need to do this in a way that really respects data privacy and within the constraints, as well as thinking about what potential biases AI might have. I’m not saying this is an easy thing to do, but it can be done. It will take time as well.

But you know, going back to your question around how Talking Points is using AI, we really believe that we’re able to use the technology to scale the principles and practices that research shows makes the most difference, because we can do it in a way that’s at scale and that saves time for educators to build that capacity. So I do think that there are some wins that we can have in AI and being cognizant of what those limits are would be really helpful.

Natasha Allen

So now I just want to turn to how can AI systems help support students and families? Where I’m unfamiliar with the American public system, what are some tools that can assist with facilitating communications between schools, and immigrant students or families? Like personalized information or education or academic expectations?

Heejae Lim

Thank you for asking that. For example, we can use AI to support tools and provide guidance to schools, teachers, and district administrators to send proactive, productive, positive, best-practice communication that is tailored to the family and student needs. We can use AI to learn from families about their student culture and encourage families to share more information that would then help guide the teachers. We can encourage families to initiate engagement with school staff and teachers and advocate for their children’s needs. We can use templates or generate templates that families can use.

Because imagine if you’re a family who’s new to the country. You don’t understand the U.S. education system. You’re working 2 or 3 shifts, you’re not familiar with English as a language, like how do you even begin to craft a message that might inquire about how your student is doing? Especially when you have a culture where you’re not used to questioning school or having that conversation with teachers.

I think there’s a lot of potential in guiding and coaching both the educators and families in kind of lubricating the relationship and building those positive connections in a meaningful way that’s guided by technology.

An example that we have is tools like what we call Education 101, which demystifies educational terms that schools use a lot, like IEP, which stands for Individual Education Plan, or assessment. Or concepts that might be new to them, like field trip, that may not exist in other school systems. So we provide that through a one-click definition and if a family clicks on it, they can learn a lot more about what those concepts mean and what their expectations are, and that way they can feel included and supported and just more encouraged to engage with teachers in the schools, and feel like teachers and schools are more welcoming to them as a family member.

Natasha Allen

And so how can AI be used to leverage and provide resources to help children and families navigate the social aspects of school in a new environment?

Heejae Lim

I mean, I am sighing a little bit because a student’s mental health and wellbeing is probably one of the top three challenges that school districts are navigating, especially in a post-COVID environment, where a lot of the kids and students stayed home and, you know, experienced trauma, and that trauma was disproportionately the case in underserved and less-privileged communities. So student mental health and wellbeing are a huge barrier for our school districts to be successful in supporting students right now.

We believe that family engagement again, you know, if children are spending 80% of their time outside school, then again, those relationships between schools and families are incredibly important in solving and starting to address the student wellbeing and student mental health needs. Families, we think of them as often a big source of information because families live with the kids and they can observe what they are like, so it can really provide information to the teachers in what might allow them to adjust their teaching, make the pedagogy more relevant, or just keep an eye on the student who might need more attention because of what they’re going through. And we believe, again, that building relationships can be helpful to student wellbeing, and as a foundational piece, can be technology and AI supporting that in the various ways that we just talked about.

Natasha Allen

That is so important, right? Especially touching on mental health. Given what a lot of these children have gone through, I’m sure we’ve only seen a portion of what the impact will be.

So then, considering the diversity of learning styles and educational backgrounds amongst students, immigrant students, how can teachers use AI to differentiate their support for these students?

Heejae Lim

So we talked about in the moment guidance that AI can help Talking Points build to support teachers in transforming the everyday communication into best practice engagement. We can guide teachers in the moment on how to embed best practices, based on what the research shows, and building those positive relationships with parents so that they can get the knowledge and leverage families as a resource in supporting their teaching. So communicating effectively by also looking at an analysis of where the students are at in terms of their attendance or behavior, or the frequency of communication of families.

Talking Points can also guide the teachers on who they should be reaching out to versus what kind of content or exercises of support the families might need based on that student and their preferences.

I’ll give you another example. Let’s say, if teachers were wanting to send a communication about how a student or class learns double-digit multiplication. If we know that the students are into, let’s say, basketball, and the parent comes from Mexican culture, then we’re able to guide the teacher to say, hey, instead of sending a message around, “we launched double-digit multiplication today” we could (we, being Talking Points) recommend an exercise that includes basketball and something that is more culturally relevant to Mexico and also includes, of course, double-digit multiplication to send to that family. And of course, this is with the support of the teachers, but I think AI can really make that process easy and super light lift so that teachers don’t really have to do anything extra other than press send or copy and paste, but it really makes that best-practice engagement a lot easier, taking the work out of the writing and thinking by relying on generative AI to be able to do that.

Natasha Allen

That kind of touches on the other point I was going to ask you about – how do you customize learning for students, right? And that’s a great way, as you said, to apply cultural relevance or things that interest the child. You can apply that to any type of education system, you’ll probably get better traction.

Heejae Lim

Exactly. And that goes back to what we discussed around meeting the needs of stakeholders exactly where they’re at in terms of whether they’re students, educators, or families, with the latest technology and AI, it’s a lot easier and faster and scalable to do. And if I turned the clock to five years ago, this is something that was always the vision for Talking Points, in supporting the family, student, teacher triad, or as we call it the tripod, in collaborating around the student learning. But to do that effectively and make it more relevant was really hard to do. It was going to be extremely manual. We were envisioning a human instructional coach sitting in the Talking Points’ office, customizing it for every student, or every triad or tripod. And now with AI, I think a lot of what we were trying to do initially can be scaled pretty efficiently, so we’re really excited about that potential.

Natasha Allen

Absolutely. Well, thank you so much for joining us. I’ve learned a ton. I’m very excited about how this can be applied to the education system, and I’m excited to see what else AI can do in this area, because I do think it’s an area that can benefit the most from AI technology. Especially because teachers are so under-resourced, right, it would really help them to do their jobs and give them a bit of a reprieve. So again, thank you, Heejae. Looking forward to the exciting things that Talking Points will be doing for everyone out there. Thank you for joining us, and until next time.

Â鶹ֱ˛Ą & Lardner’s Innovative Technology Insights podcast focuses on the wide-ranging innovations shaping today’s business, regulatory, and scientific landscape. With guest speakers who work in a diverse set of fields, from artificial intelligence to genomics, our discussions examine not only on the legal implications of these changes but also on the impact they will have on our daily lives.

Speakers

Natasha Allen

Partner

---

Heejae Lim

Founder/CEO, TalkingPoints

---

View More

In our tenth episode, Chris Rezendes, Chief Business Officer from Context Labs, joins Chris McKenna to discuss how digitalization can help companies achieve their climate goals. How is the use of data changing for energy producers and users? Where do Software-as-a-Service (SaaS) and blockchain technologies fit into efforts to improve transparency and disclosure?

Transcription

The below episode transcript has been edited for clarity.

Chris McKenna

Thank you for joining us today where we’ll share perspective from one of our clients, Context Labs. They are a software-as-a-service-based, artificial intelligence-based, blockchain-based application that provides a secure platform for generating, tracking, and managing digital certificates for carbon credits – also known as decarbonization-as-a-service. Today we have with us their Chief Business Officer, Chris Rezendes. I’ve had the pleasure of working and knowing you for the past decade, and it truly has been a pleasure. Can you give us an overview of the company’s technology so that our audience can get to know you a little bit better before we continue with the discussion?

Chris Rezendes

I’m the Chief Business Officer, which is a glorified title for someone who spends a lot of time trying to find, filter, and fix opportunities to generate revenue from fairly complex value propositions. We are a SaaS company that is obsessed with information security and cybersecurity, particularly cyber physical security. The machine data [we work with] comes out of the sensors that are in, on, and near the assets we care about. We employ lots of different machine learning techniques and very specific and very narrow AI routines. Further, tying it all together we have a distributed ledger that sits underneath all of our workflows so that we’re creating a glass pipeline of a data supply chain. We’re working with primarily the big physical asset owners and operators.

If you think about energy, upstream, midstream, downstream production, transportation, and refining, these are the entities that own and control significant physical assets that have an intimate relationship with climate. Those assets, if they’re operated well, should have minimal carbon emissions as they should be managed very tightly for their emissions footprint. On the other side, climate can represent a significant operational or financial risk, primarily through water interfering with the operation of the asset.

Think of it this way: climate is no longer an externality and emissions need to become an operating variable. Companies that have significant carbon footprints or carbon emissions or intensity challenges need to see carbon as another performance specification for their assets, for their processes, and for the products and services they deliver. They need to start to account for carbon so that they can analyze baselines, analyze performance and where they are outperforming their peers, leverage that proof of low carbon intensity or low carbon emissions, and leverage that in both commercial and capital markets.

I want to be clear – it’s not always going to be about a premium. In some cases, we have clients who might be in the natural gas space and would like to supply the European Union but can’t get their product in because they’re not compliant with the regulation called The Oil & Gas Methane Partnership (OGMP) 2.0. It can be about defending access to a market as much as extending access in a market. It can be about expanding access to capital among investors who have interest in green initiatives. It could also be about defending the possibility of an interest rate hike. Our thesis is that climate is no longer an externality. Carbon intensity, carbon emissions, and carbon is an operating variable and proving your performance is going to be paramount to continuing with that license to operate, defending access to markets, and expanding your opportunity to grow share, grow margin, and grow enterprise value.

Chris McKenna

Chris, thanks for providing the context of Context Labs. Can you describe your client lens with respect to your customers? What challenges do they have and why do they come to Context Labs and use their technology to solve it?

Chris Rezendes

There is a concept that climate generally can be seen through either the regulatory lens or voluntary lens. I don’t like the term voluntary because it’s misleading. I think companies are accounting for and tracking carbon. There’s one major dimension of climate performance but they’re doing it for two reasons. One because it is a compliance requirement. It is a regulatory requirement they need to track and report at some level. The second reason why they’re doing it isn’t voluntary, it’s compelled. They’re compelled to do it because of a competitive pressure, because of a customer requirement. Maybe there’s an industry non-regulatory benchmark, but nobody’s doing it because it feels like a good thing to do. In some cases, companies are coming to us because they want to automate their reporting to reduce cost or error. In other cases, they’re coming to us because their understanding is that their license to operate in certain state jurisdictions, California, Colorado, Massachusetts, Ohio, Pennsylvania, is being impinged by their inability to get their carbon emissions under control. And then in other cases, they’re truly looking at trying to build new assets and having challenges with raising project finance or debt finance for a hydrocarbon-based asset. They’re coming to us for lots of different reasons, but the value proposition we deliver or the challenge that we’re addressing for them is always the same.

They are somewhere on their climate journey and in their digital journey. And what they need us to do is help them find the intersection in how digitalization can help them achieve and prove their climate goals. They’re not going to be able to do it with engineers in the field taking notes with wax pencils and waterproof notebooks. They can’t do it with disparate sensing and instrumentation systems that don’t have context. They really do need to revisit where they are in their digital maturity cycle or life cycle in order for them to take the next leap in their climate or carbon journey. We work with folks that are typically on a scale of 1 to 10, 1 being immature, 10 being mature. On the Capability Maturity Model Integration (CMMI) model we’re typically working with people that are in the 4 to 5 range all the way up through the highest end of maturity. We’re trying to help them figure out how they can be operationally and financially relevant in their approach to achieving real carbon performance so that they can prove it and where possible, they can make money from it.

Chris McKenna

Thanks, Chris. Let’s talk tech in terms of digitizing across the climate journey they take. Context Labs has a very rich text stack that is an enterprise solution enabled by SaaS, AI, and blockchain that ties to physical and digital world. Can you discuss how these technologies help you solve those customers challenges and enable you to adapt and scale to each customer’s needs?

Chris Rezendes

Any of the work that teams would do in this space needs to begin with the scary word disclosure. We have lots of reporting regimes across lots of national and subnational jurisdictions. While they’re necessary, they’re insufficient. If you take the GHG, the Greenhouse Gas Reporting Protocol from the U.S. Environmental Protection Agency that mandates a certain amount of disclosure, the commercial market will yawn and the capital markets will not care. You will be trying to make a claim about your carbon performance or your carbon intensity and no one will care. This is the place where the Context Labs value proposition starts for industrials with big physical footprints and carbon intensity or water intensity challenges to move commercial or capital markets more in their favor. They have to disclose beyond compliance, and the biggest disclosures they need to make are machine data from sensors that are in, on, or near their core process assets. They’ve got to take certain crown jewel data that has never been conceived of as being disclosed to a third party and that becomes the fundamental disclosure moment.

What do we do? We start with that premise. You need to disclose beyond compliance crown jewel data, and Context Labs will say that we will treat your crown jewel data as a national security asset. Information security and cybersecurity is a digital thread of trust. We start with proving to them we will protect their data from leakage, misinformation, or disinformation. Second, we have a distributed ledger that sits underneath all of our workflows so that every entity that enters a transaction and every event in those transactions is recorded to this distributed ledger. That’s the glass pipeline. Now the ledger itself doesn’t necessarily provide superior security. It provides the ability for no one to say, “I didn’t say that” or “We didn’t provide that” or “That wasn’t our statement.” In many ways it’s about attestation, non-repudiation, and risk management and mitigation, commercially versus in a cyber sense.

Once we’ve got hardened data, and we’re running workflows that are in the ledger, we’re able to help clients take privileged data, disclose it, and reduce commercial and operational financial risk through these disclosures. One way that we strengthen all of those workflows is by having a very strong application programming interface (API) that enables us to embed second- and third-party opinions into these workflows.

If you’re ACME widget and you’re trying to make a claim of low carbon intensity, and you’re going up against a big sovereign wealth fund or pension fund, and you’re looking for US$3, 5 or 10 billion in project finance and they’ve got a big climate mandate, they’re going to be asking you questions about your past performance and current performance as indicators of future performance. You can come and say “I’m ACME, and this is my claim,” or say “I’m ACME, and here’s the evidence of my claim and I’ve subjected it to the scientific experts, these industrial experts, these risk and finance experts, and they’ve all signed on to the accuracy of my data, analytics and statements.” We get to a whole new level of backing climate claims, and those three things, the securing the data as if it was a national security asset, this glass pipeline underneath all the workflows, and then the secure integration of these second and third party opinions, meets or exceeds existing and proposed regulation coming from U.S. Securities and Exchange Commission, Federal Deposit Insurance Corporation, National Association of Securities Dealers, and others.

Chris McKenna

Thank you so much and good segue because we talked about the importance of machine data. We always have this conversation about data and you know one of my taglines is that data is the new form of IP currency. Data has become so important to producing services and solutions; maybe so much so that it’s actually more valuable and important than the solution itself. Can you describe Context Labs’ perspective on data, it’s approach to its data governance model, and how you see that playing a role in digitizing the physical world?

Chris Rezendes

Chris, you helped us coin a phrase here: asset grade data, data that has our Context Labs brand of trust services, identity services, hardening enrichment, and contextualization services. Those are just different dimensions of preprocessing. Imagine really intense preprocessing to prep the data to then say, we can stamp this data as having passed all these tests. It is asset grade data, meaning it’s a new asset class where you can use, share, and disclose this data across the panoply of applications. For those of you that are more technical than I am, you know sometimes a single sensor provides a single data stream and it’s fairly brittle with respect to what it can tell you or how you can use it. Sometimes data breaks down and you put it in a new application or a new context. Our primary thesis is data as an asset class, asset grade data.

Second, your data is always your data. We are very clear that data is not digital exhaust, it is a national security asset. It’s a crown jewel asset. In many cases, we might argue data about emissions, water, or some other important climate mandate in a given context could be as powerful as the physical production itself when it comes to trying to create brand, market awareness, differentiation, share, and margin. We don’t build subscriptions; we don’t repurpose the data. We don’t have second, third, or fourth exploits of the data unless it is for our clients. If the data can be imbued or if we can embed these security regimes into the data so that it can move across workflows, then we’re actually having our cake and eating it too, or enabling our clients to. You’re going to disclose more data and reduce your risk.

Data in motion and data disclosed is at risk, but we would say that our tooling, which puts us in position to be one of the 35 “permanent members” of the National Cybersecurity Center of Excellence (NCCoE), that’s a benefit for us. The NCCoE uses our tooling for all kinds of critical infrastructure programs so we end with this idea that you can disclose and manage risk if you use the right tooling. Chris, time and free will as the paramount human right, clean air and water to make a healthy corpus, to harness that spiritual and mental energy, that is human. I think data might be the third part of that Holy Trinity going forward, where ultimately every individual, and we do work with indigenous communities all the way up through the Fortune 500, any entity, whether it’s a human or corporation, really does need to see core data about its existence, identity, its location, state, and status as being one of those critical human rights. That’s how we view data. Maybe it’s a little prosaic, but when we break it down, the code that we’ve built, the software that we market, everything that we do, every workflow, every contract, every engagement, we honor climate alignment by honoring the clients who’ve made that commitment honestly, by honoring the data.

Chris McKenna

Chris, great thesis in terms of the importance of data and part of the Holy Trinity of human rights. I really appreciate that and it fits into my view of the world as well. To kind of wrap up here, I have one more question for you, and this is a little bit more future thinking. I’m hearing from economic analysts that the energy sector is one of the sectors leading our way back to economic recovery. We’re seeing an increase in deals and transactions. What are you seeing in your role as CBO for demand in the energy market, what are drivers now and in the future for your solution?

Chris Rezendes

The vast majority of our business right now is in energy or energy transition. We do a lot in hydrocarbon, primarily gas more than oil. We’re doing it upstream, midstream, downstream. We’re doing lots of work in energy transition fuels, including starting to do work in hydrogen and ammonia, etc. And we’re doing lots of work in renewables. Here’s what I’d say about that. Back to sort of the Holy Trinity: climate security, energy security, and food security are inextricably linked. One cannot achieve one at the expense of the other two. They’re kind of like three interrelated dials and a meter. We need to get our energy security more in line with our climate security and our community security. Energy has to continue to become more available. It has to continue to become more affordable, but as importantly, it’s got to be aligned with a climate neutral or a climate regenerative industry.

I’m just back from Abu Dhabi, who hosted ADIPEC, the world’s largest energy only trade show. 160,000 folks were there, and I can tell you the number three, the number eight, the number fifteen, we had the largest energy companies in the world spending time with us because they know that in order to meet their mandate, meet the moment, meet the mission, it’s not just about the MMBtu or the kilowatt hours or the kilojoules. It’s not just about the energy metrics, but it’s about how they produce those energy metrics. You’re going to continue to see folks try to lock down access to gas more than oil, and the best hydrocarbon sources possible, but more than that, all the way at the other end is no end of discussion around how far can we go in hydrogen and not continue to blow out CO2 emissions budgets or deplete aquifers. Electrolysis is a very water heavy process. If you keep going further to the right is the concept of the Department of Energy and the energyshed program, whereby in the future the vision is we can not only select an electric or utility bill to pay a premium for green, but we might be able to pay a second premium for local or for a specific source. Energyshed proposes to put distributed energy resources in currently and historically underserved, underinvested, and underrepresented markets.

Imagine if traditionally underserved and currently underserved infrastructure development environments, those census tracts that are not as resilient as they need to be, what if they can start to crack the code on getting energy from a solar farm into a distribution network that right now isn’t so friendly to the data coming in from the solar farm? The point is whether it’s the hydrocarbons or the transition fuels or the next generation of truly being able to have not just an intelligent and adaptive grid, but an intelligent and adaptive marketplace. All of that needs hard data from the physical world run through secure systems so that the big people and the big chairs can make the big decisions and the little people in the little chairs can do the same. These tools and methods can speak to the Fortune 5000. It can speak to the folks that are at the base of the pyramid. If that kind of data can be trusted and distributed and used to make these kinds of decisions, then I do think we have the opportunity to start to make progress on the energy security, climate security, food security, and then that will be the gift that keeps on giving for community and national security.

Chris McKenna

Chris, huge thanks. What I always love about having conversations with you is that I’m always learning something new. We thank you for all your insights here and your client lens in perspective from your world – you can tell you are very passionate about securing our physical world through digitization and I’m sure everyone here has learned a lot. We appreciate all the perspectives you gave us, so really big thanks to you, Chris.

Â鶹ֱ˛Ą & Lardner’s Innovative Technology Insights podcast focuses on the wide-ranging innovations shaping today’s business, regulatory, and scientific landscape. With guest speakers who work in a diverse set of fields, from artificial intelligence to genomics, our discussions examine not only on the legal implications of these changes but also on the impact they will have on our daily lives.

Podcast Guest

Chief Business Officer – Context Labs

---

View More

In our ninth episode, Aaron Erickson from NVIDIA joins Natasha Allen to share his perspectives on where generative artificial intelligence is heading. How are LLMs evolving beyond natural language to incorporate multimodal inputs e.g. images, sound, video, or even taste and smell? What are some of the new economic models being developed for AI tools? And are we primed to see a trend towards bespoke models designed for specific organizations?

Go Deeper:

• Navigating the AI Frontier: Legal and Operational Insights Into Generative AI
• My Employees Are Using ChatGPT. What Now?
• NIST’s AI Risk Management Framework Helps Businesses Address AI Risk

Transcript

The below episode transcript has been edited for clarity.

Natasha Allen

Thank you for joining us for another episode of the Innovative Technology Insights podcast. My name is Natasha Allen and I am a corporate partner in the San Francisco and Silicon Valley offices at Â鶹ֱ˛Ą and Lardner. I’m also the co-chair of the AI sub-sector, which allows me to have this cool opportunity to interview exciting individuals working on cutting edge technologies, including AI. I have with me today Aaron Erickson. Aaron is a key member of the software engineering team at NVIDIA, helping to build their enterprise DGX AI platform. Prior to NVIDIA, Aaron spent 30 years working in leadership roles, most recently as CEO of Org Space and a VP of Engineering at New Relic. Over the course of his entire career, Aaron has been an advocate for building better software from his home in San Francisco. Welcome, Aaron.

Aaron Erickson

Thanks for having me.

Natasha Allen

Alright, so let’s get started. With the advent of ChatGPT came a better understanding of the world of AI. Prior to this, layperson individuals didn’t really have an understanding of what AI was or what it could do – in particular what large language models (LLMs) could do, which, as you’re aware, are the foundation for deep learning algorithms that help us understand and process natural language. As of now, we understand that LLM’s can be used for text, specifically to help generate or prompt certain text. But how do you think LLMs will evolve beyond natural language, incorporating multi-modal inputs such as images, sounds, video, or even taste?

Aaron Erickson

It is one of the most exciting developments, I think. People are seeing this happen even now. The latest ChatGPT release, if you have the premium version, you can talk to it and it talks back in voice to you. In fact, one of the scariest things that people saw with initially was looking at CAPTCHA. If anybody’s ever used a CAPTCHA where you have to say “which of these things on the screen is a traffic light or not” ChatGPT can pretty much do that even today.

Multimodal is, I think, one of the really interesting places it’s going. One good example of this is, you think about Tesla and what they’re able to do with video. When you drive a Tesla, it’s almost like you have a kind of permanent dashcam on the car. It’s always taking pictures and it’s a big part of how we can drive independently or autonomously when you put it in that mode. But one of the really interesting things is how they’re using all that video that they’ve accumulated. Google is also doing this with the models they’re developing, and a lot of these new models are taking video and able to transcribe it to understand what each frame of the picture means. This is in order to really get at “what could we learn if we could see the world around us?”

The best way to think about it in a pre-multimodal LLMs is imagining you trained a human or took them to school, but the only thing they could do in school was read books or papers. They couldn’t actually experience the world as it is, be it through seeing the world, or in some of the more speculative applications I’m hearing about, even smell the world or taste the world, or understand different aspects and be able to code it into a training model. At the end of the day, you almost kind of have your inner monologue where you think about language. You look at a light and it translates into the word light and then you’re able to associate that to “when I have a light on, I can read.” That same kind of trick that we do as humans, we are now able to do with these kinds of biologic or synthetic brains that we’ve created.

Natasha Allen

Wow, that’s amazing. With the advent of this technology what are some of the key challenges you think to integrating multimodal inputs into LLMs, especially when it comes to sensory, taste and smell, and those things that you think are inherently human?

Aaron Erickson

The first obvious sense is sight, thinking of being able to look at pictures, which we’ve been doing for years, and being able to look at video, which is a bunch of pictures and a lot of processing power, but it’s a known quantity. It’s not particularly groundbreaking at this point. Other senses, like being able to taste and smell, we have ways we can, through understanding the chemical composition of a substance, infer what it might smell or taste like. If you can digitize something, and these are all things that are digitizable, it’s just a matter of some of those other kinds of senses being a little more expensive or more involved. But there’s no reason it couldn’t be done. I used to say this, almost kiddingly “are we going to have generative smell next?” I’m not sure it’s something we want as society, but there’s no reason why it couldn’t happen. It’s just a matter of translating it into some digital form and then allowing the model to learn from it.

And so what is multimodal? There’s what the inputs are, what it learns from. There’s also what it can output. We think about generative AI and ChatGPT and smart technologies as generating text and pictures, and maybe generating videos someday. There’s early models that generate videos where you just type in “I would like to see blah happen” and a video is generated that does that. It’s a little rough right now, but the models are getting better and the ones you see today are the worst that you’ll ever see. They’re only going to get better from here.

One of the things that we do at NVIDIA is we have a product called BioNeMo and this is helping do generative drug discovery. How do you do that? If you think about it, it’s another kind of data and you’re able to literally use the same kind of technology to generate what ideal protein strands might be, maybe a cure for COVID, or whatever it is that you want to solve. Those kinds of applications are happening today. And research is being accelerated, think about the kinds of medications we’ll be able to get or the kinds of things we’ll be able to do with this technology, like make power plants more efficient. I know one startup that has demonstrated making a power plant 90% more efficient by using this kind of technology, helping figure out what would be some ways to make a plant more efficient that aren’t obvious to a human.

Natasha Allen

I think it’d be so amazing to see that evolution. We have the ways that we can expand the use of LLMs, now how do you monetize it? How do you monetize these AI tools? Maybe you have some examples, or can you can explain what are some of the current economic models prevalent in the AI industry?

Aaron Erickson

You wonder why some of these startups are raising what seem like ridiculous seed rounds, I heard of one that was like US$50 million or US$100 million. And largely, they’re not just hiring employees – you’d actually be surprised by how few employees a lot of these firms are hiring. But largely they’re buying infrastructure. Either from us [NVIDIA] or from other providers of GPU technology, a lot of the money is going to go into that. And those are people developing foundational models. That’s your OpenAIs and Anthropics of the world. As well as some of the big FAANG companies, they’re all doing some version of this too. It’s just an arms race right now.

The more interesting thing is, and this is what I think a lot of people miss, is that there are a lot of what I call narrow AI solutions that you can build. They might be good at a specific problem. One good example of this, and you don’t need a massive array of GPUs to do this, you could develop a reasonable model that detects fraud.

Say you want to run your fraud detection department to find people that are fraudulently putting in expense reports or doing some other thing that was anomalous. You can build an LLM that is very narrowly trained to detect fraud and actually run that against your accounting system or run that against a set of accounting systems, an email, or something like that to be an early warning indicator. That doesn’t require the same kind of GPU fleet that you would need for building a foundational model. You could take something like Llama or one of the open source models, enhance it or fine tune it with data about how fraud happens or examples of bad examples of fraud in your company, and then build some pretty incredible products. This could be done internally in companies, building some pretty incredible capabilities using these things to automate certain kinds of routine intellectual work. That’s just one example, but it’s hard to go into a company and not find maybe 20 or 30 other examples like that where a narrow AI can do incredibly useful things that are far more valuable than what you’ll spend developing it.

Natasha Allen

Interesting. So maybe elaborate on the concept of bespoke LLMs. Is this a new thing that executives could ask for – I want my own LLM tailored to myself. What would make a bespoke LLM different from an off the shelf AI model?

Aaron Erickson

Why would you develop a bespoke LLM in the first place is a good question. Because right now if you use this ChatGPT or whatever, it’s going to tell you general facts about the world. It’s not going to give you your sales forecast. It doesn’t know data about that. It might say, well, you know this company is doing pretty good and maybe even have it attached to Bloomberg and maybe it will tell you a little more about your company. But it can never know as much as you would know inside the company. This includes your private data, your accounting system, your HR system, and every kind of every digital footprint in your organization that’s specific to you. That’s just one of the obvious reasons to build your own LLM, to have a training model that, as a CEO, can tell you anything you need to know about your company – it can give you advice on your strategy, almost become like a McKinsey consultant in a box. Even if you have other people doing that, having something like this where you can iteratively ask questions and do so whenever you need, whenever the thought comes by, is pretty useful.

Then you think about, how is it different? I think it’s kind of like how somebody who’s been in your company for 10 years might be slightly more valuable than somebody that’s only been there a week. Institutional knowledge, right?

A person that’s been around the company for 10 years has a lot of tacit knowledge about everything from how people talk, what the company culture is, what kinds of things are inside or outside the opportune window, if things ought to be discussed or not, ideas that have been tried, things that the organization as a whole has learned. It would be a shame to have LLMs and not be able to take advantage of these things that we learn inside our companies over time.

These models get more and more useful the more you can train them on the context of your organization. And this is the part that’s incredible, and why we’re seeing a lot of demand for these kinds of systems, is that CEO’s don’t know the limits. Right now the more horsepower you put into one of these things, the smarter they get. It starts to become an economic imperative to have the smartest AI model. One smarter than your competitors. One trained on better data, one trained on more GPUs, you’re effectively making it smarter and that can be a competitive advantage over time. Especially as we start to get closer to artificial general intelligence, which even Sam Altman and other people are saying is a lot closer than we think. It might be before 2030 that we get this kind of broader concept of a generally intelligent LLM that’s better than most humans. It’s pretty incredible to think about. I don’t go a day where I don’t think of another application, so it could be used in one of these contexts.

Natasha Allen

With that in mind, two questions. The first one is, if organizations decide to go down this bespoke LLM route, what is some advice or considerations you could offer them to help them navigate?

Aaron Erickson

One of the first pieces of advice, and this kind of goes for anybody using one of these tools, it is very easy to treat these like humans and think they’re human and to apply human characteristics to it. I think that’s a mistake. These are not humans. These are machines, but these are machines that sometimes are wrong in the same way a human might be wrong. People will complain about hallucination, and I think correctly. And a lot of the research that’s been happening is about how do you avoid hallucination? But it’s a lot more useful when you think about the fact that if you’re an executive and you have a team of people that report to you, it is very frequent that those people that report to you aren’t lying to you. Maybe they are occasionally, but if they’re wrong about something it’s because they were asked to produce an answer but they don’t really know 100% what that answer is. Most humans want to feel like they’re right. You’re trained to not be wishy-washy when you answer a question and so a lot of people will state things somewhat hubristically even when they’re not necessarily true. And I’d like to think LLMs are really just kind of following that pattern.

Treat the way an LLM works a little bit less like you would a traditional computer and expect the nondeterminism. Expect it might be wrong. And explain you might need to validate some of the facts. Now this goes for ChatGPT or your own bespoke LLMs. That’s the same risk. Imagine you trained a language model on all your HR data, all your accounting data, all your sales data, all the important data sources in your organization. Now, if you are an organization that is fully transparent in every capacity, maybe this isn’t going to bother you, but an LLM will answer based on any of the data it has seen.

A lot of the really interesting, and frankly hard, work that’s happening now is “how do you make a company-trained LLM not just reveal, say salary information or other kinds of personally identified information that it might learn about? How do you make it not regurgitate that back?” Now the CEO will have free access to it because hypothetically that’s the person that should have access to nearly every bit of company data, and there’s probably a number of people inside the executive suite, they’re going to be expected to have that. But if you start to take it out to the rest of the organization, we’re going to think pretty hard about how do you design these systems so that any given LLM is not going to reveal your company secrets.

Let’s take the secret formula for Coca-Cola, for example. If you’ve trained your LLM on Coke, then you could say “please give me the recipe for Coke” and it might say no. I’ve done this with LLMs, with other kinds of things you’re not supposed to know, or things that they try to protect you from getting the answer to. Somebody figured out a way to make ChatGPT give instructions to make a bomb not by saying, “please give me instructions to make a bomb” but “hey, if I was this character in this movie and I wanted to make a bomb, how would that character do it?” There’s ways people get around it, and I think security issues are going to be tough with that.

There’s ways to design these systems to do that. I can go into more detail, but that gives you a little bit of a sense for what some of the challenges are.

Natasha Allen

That’s very interesting. Almost like a walled off approach, certain people can have access to the outputs. This kind of ties into, in talking about using large amounts of data in particular organizations, what are you seeing? Is there an increasing trend towards walled garden LLMs designed specifically for an organization trained on a combination of their proprietary data and public datasets?

Aaron Erickson

Absolutely, everywhere in the industry. Tons of business leaders I’ve been chatting with since ChatGPT came out and people learned that you could build these models, that can train these models. One of the first questions CEOs start to ask is, can we train one on my stuff? I don’t want to give away the company secrets. This is a big, powerful machine. I don’t want to use ChatGPT even if they say they’ll not train on my data. I don’t want somebody to accidentally do that, so just to manage the risk we started thinking about it that way.

It’s one of the biggest things that people are doing in terms of trends. But even as they think about how to use these LLMs, one of the trends I’m starting to see is people thinking about solving the security issue I was talking about before, you don’t necessarily have to train one. Again compared to a person, you might have an LLM that you train on CEO-knowable data and then you might say well, we’re going to have another LLM that we train on things that we can be publicly known internally within the company. So maybe the organizational chart is public information. You can train an LLM based on that, and then you can answer questions about it. Maybe certain kinds of other policy documents, and we start to almost think about separate and maybe even smaller purpose-built LLMs or machine learning systems for different parts of the organization, by department, if you will. The LLM structure starts to resemble a more traditional organization structure.

Natasha Allen

Interesting. Are there any examples of industries or sectors where you think this walled off garden LLM approach is beneficial?

Aaron Erickson

All of them. When the IBM 360 came out in the 1950s and 1960s, the idea of every company owning their computer was pretty ridiculous. IBM owned the computer and everybody shared it. You had to be a pretty big company to own your own early on, hence the whole time-sharing thing. Over time, it became extraordinarily commonplace for companies to have their own computers. People started building their own networks within them, and then we eventually ended up with the PC. I think you’re going to see the same thing with machine learning systems.

Right now, because the economics to build a viable LLM are so difficult, if you want to train GPT5, you’re probably going to need over a billion dollars to do that. But the cost is going to come down, I think pretty radically over the next three to five years. The techniques to do custom elements, the raw technique, rank adaptation, and others that are becoming more commonplace, I think they’re only going to progress. Between that and there’s so much capital going into different kinds of AI startups that there’ll be a solution that you could probably own within your four walls for just about any kind of problem you can imagine. There’s going to be a marketplace bigger than what most people think that’s coming with this stuff.

Natasha Allen

That’s great. And I agree. I think it could be used across many organizations, probably the ones with larger datasets, right? Last couple questions – walled off garden LLMs, do you think they will replace enterprise software or do you think the two will coexist and be complementary to one another?

Aaron Erickson

I think there are classes of enterprise software it will replace and new kinds of enterprise software will emerge.

Natasha Allen

I agree.

Aaron Erickson

It’s going to be like every other revolution. That stuff will still be around. You can still buy enterprise software for doing sales data, that was around before Salesforce. You could probably still build systems like that or use systems like that. They won’t entirely go away, but the state of the art will certainly be systems that just tell you the answer to the question you have, and then maybe let you interact with data in some way that makes sense for a human. The idea that you have to be trained to use a SaaS product I think will largely go away. Any SaaS product where you have to be trained as a human to do it is probably more complicated than the LLM version that’s going to be a combination of a chatbot plus some sort of model that you can interact with that helps you understand a solution to a problem or help you model something.

Natasha Allen

Final question – if organizations are thinking about offering or adopting walled garden LLMs, what steps do you think they should take to maximize the benefits while trying to minimize the risks, some of which you already alluded to before with regards to hallucination?

Aaron Erickson

I think it’s very easy to just expect it to be magic and so I think tempering expectations a little bit. I think there’s a lot of really great experiments you can run without making the big upfront investment. One of the most powerful things about using OpenAI’s API or some of these other organizations’ APIs is that you’re able to experiment with what’s possible. You’re able to explore the art of the possible, which helps you understand, okay, well, if we can do the small thing really, really well and do it on somebody else’s LLM, you understand how your own data would help with that decision. That’s how you start to build the economic case for doing this stuff. That’s where some of the value is going to come in. I think it’d be a mistake for most organizations to say, oh, I need to build my own GPT5. Some will do it, right? I can see in the next five years, some companies saying, hey, I need to have the smartest one of these. I want to compete that way. Not against OpenAI, but one car company versus another car company. I think there’s a lot of big investments like that. I also think, like any other industry, there’s going to be a tremendous amount of waste in terms of people either not understanding what they’re capable of or expecting exact answers and having no tolerance for data even being slightly wrong, which kind of misunderstands what an LLM is capable of. Even the best GPT5 will probably still hallucinate from time to time, and we’ll still need a human in the loop for anything that’s life critical. I think those are some of the key ones.

Like I talked about with security, I’ll say again that being very aware of what data a given LLM has been trained on is critical. And starting to think about well, okay, maybe we need more than one, maybe we need dozens of them, just like we might have dozens of employees that understand their domain really well. And in fact, we train them on that narrative, and you might actually save money and not have to develop very, very expensive general LLMs. You can train them on their very specific function and almost create coordinator LLMs, just like your senior vice president that coordinates the activities of the subordinates – creating your own organizational structure out of synthetic brains not the biologic ones.

Natasha Allen

Very interesting. Well, that was my last question. I think this was a very insightful discussion. Appreciate you taking the time to talk to us and chat about some of these more cutting-edge decisions to be made when you’re dealing with LLMs, the AI tools, and what the next frontier may be. Thank you everyone for joining us, and until next time.

Â鶹ֱ˛Ą & Lardner’s Innovative Technology Insights podcast focuses on the wide-ranging innovations shaping today’s business, regulatory, and scientific landscape. With guest speakers who work in a diverse set of fields, from artificial intelligence to genomics, our discussions examine not only on the legal implications of these changes but also on the impact they will have on our daily lives.

In our eighth episode, Parna Sarkar-Basu, CEO of Brand and Buzz Marketing, joins Paige Moscow to discuss pivotal role a founder’s market credibility plays in propelling startups to success. How does it impact the fundraising process? Why should companies with technical innovations care about their brands? What value is there to putting a human face to a startup?

Transcript

The below episode transcript has been edited for clarity.

Paige Moscow       

Thanks so much Parna for joining me today to talk about founder branding. Before getting started on introductions, I just wanted to say how excited I am to have you join me on Â鶹ֱ˛Ąâ€™s [Innovative Technology Insights] podcast today, not just because of your expertise in this area, but also because it was so wonderful to reconnect with you at our Women and Venture Capital event last year.

So, I’m going to turn it over to you briefly to introduce yourself and tell our listeners a bit about you and what you do.

Parna Sarkar-Basu   

Paige, it’s an absolute privilege to be here. Thank you so much for taking the time to talk to me about one of my favorite topics: How do you build market credibility for founders? So, I am the founder of consulting firm called Brand and Buzz. We primarily work with technology companies and CEO’s to either build and/or support their business strategy, their expansion plans, as well as help them get ready for their fundraises, acquisitions, and so on.

But one thing I do want to point out is that we are not the ones who are having conversations with the investors, bankers, or funding agencies; it’s all the behind-the-scenes work that we do to help the founders or executives get ready for whatever their business goals or exit plans are. I also have the privilege of working with a lot of women in tech and women entrepreneurs helping build their market credibility to spotlight the expertise they have and showcase that with the world.

Paige Moscow         

Thanks so much Parna. Before we dive in, I’ll just introduce myself briefly as well. I’m a corporate attorney here at Â鶹ֱ˛Ą and a large portion of my practice focuses on advising founders of startups, both emerging and later stage, on general corporate matters, including and navigating through rounds of financing and ultimately the sale process. So, part of your focus really resonates with me and I’m looking forward to diving into some questions.

First, I’ll just ask you – in your written pieces and speaking engagements, you talk a lot about building a founder’s brand. So, what does that actually mean in practical terms?

Parna Sarkar-Basu   

Sure. So, building this brand is really about building credibility for the individual. It’s about showcasing their expertise, their knowledge in a specific field and when appropriate, also sharing some of their personality with the world because at the end of the day, it is the humans that you’re interacting with. Having that credibility and market visibility for a founder, I feel is extremely important.

One thing that I would like to clarify if I may, is that oftentimes I hear people confusing personal brand with self-promotion and bragging. But what we do and what we recommend as part of our strategy and what we also execute for the founders really has nothing to do with self-promotion. It’s really about developing a strategy for the founder, but most importantly positioning them. It’s similar to how you would position a company, right? Every founder that we work with has a three-word positioning where we go through a major process, but in short, I would say I founders brand is key to success.

Paige Moscow          

And so, when you talk about founder brands, maybe you can share some examples from clients you’ve worked with in the past or success stories of branding founders.

Parna Sarkar-Basu   

Let me step back and ask a question to your audience. I’m sure a lot of people have heard about Richard Branson, right? But I wonder if they can name the 400+ companies that are owned and controlled by the Virgin Group. Or even Rihanna, we all know her as an entertainer and musician but she’s also an entrepreneur and founder of a couple of companies.

So my point being, most of the time we recognize people by their personal brand about who they are and what they’ve done and less of it is about how many companies or what product they’ve built. Now if we look at just stats alone, I’m sure this number has changed but as of 2020 there were more than 518 million entrepreneurs worldwide and about 30 million of them were in the U.S. And last year there were more than 5 million business applications filed and this year there have been 1.74 million business applications to date.

The reason I bring this up is to show that if you are one of those 5 million plus entrepreneurs, how do you stand out? And I think that that’s one of the reasons why when we work with companies, especially first-time entrepreneurs, we find it to be extremely important that we build out the “who they are” and also think about it as a political campaign. If you’re looking at a personal run, for example, you know the field is crowded and we are voting for a candidate based on what we know, how visible are they? Are they credible enough that we know that they are going to actually get the job done for what we vote for them. I feel like that is what I think about when we are building out the brand.

So, like I said, almost every founder that we worked with or every executive we worked with, they all have a personal positioning to start with. And case in point, one of them, we worked with them for over 3 1/2 years and he was the face of this particular company. And the reason we positioned him as a CX [customer experience] connoisseur was because one of his biggest strengths was how good he is with customers and closing deals. And if there’s a problem somewhere with the customer, he knows exactly what to do. We also considered his personal goals, which is something we look for what happens when he is no longer with this company.

I believe personal brand must be portable. Whether you are no longer with the company or your company gets acquired, if you are too tightly tied to a company then you might have to rebuild your brand again, right? And in his case, we positioned him as the CX connoisseur and everything around the company positioning was also around customer experience. Given his personal positioning as a CX connoisseur, everything that he did, said, wrote, or spoke about had to do with CX.

It was all CX and digital transformation, CX and entrepreneurship, CX and STEM, etc. Everything revolved around that. That is also very important piece because you have to be on message and if you are positioning a certain way, you need to make sure that you stay that way. To kind of conclude this example, we worked with him for about 3 1/2 years and the company got acquired. One of the signature programs that we created for him, keeping in mind that brands need to be portable, was a podcast called the Science of CX.

So, when the company got acquired he was saying “Hey, what am I going to do next?” He kept his podcast going. That’s an example of how we do personal and executive positioning, help build their credibility, something that they can take with them, whether they’re starting another company, trying to do consulting work, or what have you.

Paige Moscow     

You know, I think that’s really interesting because I feel like a lot of times when you think about the word marketing you think about trying to market a company and a product. A lot of folks who are starting companies are thinking about “how do I market my product” at the entity level. And you bring up a really good point that oftentimes the focus has missed that the individual who is running the company needs a brand of their own. That’s a very powerful piece of the whole branding effort.

So obviously one of the main goals for many emerging startup companies is to bring in funding from venture capital firms or other high net-worth individuals or investors. So, can you talk a little bit about how founder branding intersects with the fundraising process?

Parna Sarkar-Basu

Great question, Paige. Yes, the two absolutely intersect. Investors invest in founders and customers buy from people so that human angle is right there. Based on my personal experience and also having spoken with several investors, I often say “what do you look for when you are looking at a company or you want to fund a company or invest in a company?” One of the answers always is “we look at the founder and they believe.” I don’t want to speak for all investors, but at least the ones that I know and have spoken with, they know that the founder or the founding team is very good at the technology that they’re building. It’s innovative and they have competitive advantage, etc., etc. But at the end of the day, they want to know that the founder, who will eventually become the CEO of that company most of the time, is savvy about running a company, is a savvy entrepreneur, and knows how to scale the business because when investors are investing in your company, they do want return on their investments. When I was talking to investors about this very topic, they said that the technology is pretty cool, we know that it can evolve, and we can hire engineers and AI experts to help, but we need a founder who has the passion and vision to not only focus on the technology but know how they’re going to grow the company.

So, I would probably say that for investors to invest in a company and the founder, I think founders have to be more than the product they’ve built. Especially if they’re looking for investment – having worked with lots of tech companies and supported dozens of entrepreneurs I feel that is a very common thread, at least the clients that we’ve supported. If I look at the last few years, I’ve had the privilege of working with five companies that were acquired.

Paige Moscow

I mean, it connects to what we were just talking about. When you’re pitching a VC firm or investors, you are really pitching a personal brand. You are pitching the founder’s brand, so the focus isn’t as much on the technology or the product, which obviously is important, but a huge piece of it is you pitching yourself as the founder.

Parna Sarkar-Basu

Yes, you have to show your expertise not just in building the product, but also showing your entrepreneurial skill in that you are looking ahead with the investment dollars. How are you going to scale and where are you going to go? Who are going to be your customers and how quickly are you going to give me the return on my investment? I mean, at the end of the day, it’s about money.

Paige Moscow

Interesting. Okay, so backing up a bit. We were talking about the funding stage when you’re pitching to investors, but is it ever too early to start thinking about putting together a founder brand, even when you might not have a product yet? Is it ever too early?

Parna Sarkar-Basu

The short answer is never too early. As of last year, we’ve been working with three companies, one is in stealth mode. The other one has a product that’s gone into testing at airports, and one that is still deciding if the product is viable – they don’t have a prototype in essence. So, we have worked with each one of them to build out their product positioning. While they’re building the product and testing the product, we are absolutely focused on building out the CEO’s brand and the market credibility.

One of them in particular is in robotics, but the person does not have a high profile in the robotics industry. So, we’ve been working a lot with this individual to build out and spotlight her knowledge. She has the knowledge, but the world doesn’t know that. So, even though the product will not be ready for prime time until later in the year, we started working with this group while they were still developing.

The whole point and the premise here is that we have to spotlight the people behind the product. And show the world that they know what they’re doing. So when the product comes out you already know the human behind the company, right?

Paige Moscow

Right. So, you’ve talked a little bit about different strategies you’ve used for branding but do you use different kinds of strategies to develop market credibility for founders versus CEOs of more established companies?

Parna Sarkar-Basu

So, what we don’t have is a cookie-cutter strategy for every company. We also don’t have a cookie-cutter strategy for startups versus large companies because I believe every company is unique.

Every team that’s running a company and building a product is unique. So, it’s really important that we understand their business goals. We understand the market, we understand what their competitive advantages are. And then we create a highly customized program for each of the companies and the founder(s). But there are a few similarities, and when we go about the process, not only do we have to understand what the business goals are, but we absolutely have to understand what the personal goals are of a founder of a CEO of a large company. Because in fact, I almost always feel like I have to sign an NDA because I learn so much about someone, half of which I cannot use because it’s personal information. But what that does is that it allows me to know the person, know their passion, know what their weaknesses are, and then package all of that up and spotlight nothing but their expertise.

And when you’re building the strategy, you also make sure you know what they hate, where they don’t want to go, what they don’t want to talk about. When you are creating this package, you are always able to stay away from those because you know that’s not going to be helpful for other individuals or for the company. We intimately know the person’s character, history, and experience in addition to the business world, and that is one of the key factors that plays into our positioning strategy. That is kind of the consistent theme, whether you’re a startup founder or your large company founders. But the difference is that, as I was saying, brands have to be portable.

If you’re looking at the likes of an IBM, the founder isn’t there now and as the company grows or scales you bring in the CEOs or the business leaders who can take your company to the next level. There aren’t too many like Salesforce with the founder still there. Oftentimes, they’re very focused on promoting the product. It’s all about the speeds and feeds. It’s all about the technology, and yes, they’ll do customer service, but it’s around case studies and it’s like very technical, very deep in the weeds.

So how do you position a 55-year-old company or 100 plus year company? You have to essentially bring everybody along with you while you are building a new strategy in conjunction with all the stakeholders obviously. But what we’re also doing is humanizing that brand by putting faces or a face to the company because people are not buying from logos – logos don’t sell, it’s the humans that sell.

So analyzing the brand is kind of a consistent theme for us because that’s one of the key things that we do, whether we’re rebranding, getting them ready for fundraise, acquisitions, or what have you and it’s a hard process; but I cannot see myself doing anything else.

Paige Moscow

You’ve obviously given a lot of advice and tips today on branding for founders and CEO’s of large companies that kind of spans the underlying themes. But so, if I were an entrepreneur just listening to this, what’s one tip you would give? Kind of a practical step that they can take to start working on their personal brand.

Parna Sarkar-Basu

The one tip that I often give to entrepreneurs and founders is, be the face of the company, because that humanizes the brand. And building your founder’s brand is not about you really. It’s really about building market credibility for your company, building market credibility for yourself. So that investors are going to invest in you and your company and also it helps you reach your customers. They need to know who you are. You need market visibility and market credibility, so being the face of the company, I think is extremely important, especially for first time founders.

Paige Moscow

Well, thank you Parna. So I don’t know if you have anything else that you wanted to share to conclude, but I really appreciate your time today and this was a great conversation and I hope that it was helpful for any entrepreneurs or founders who are listening.

Parna Sarkar-Basu

Thank you so much, Paige. It’s been a pleasure talking to you. Thanks again for having me.

Â鶹ֱ˛Ą & Lardner’s Innovative Technology Insights podcast focuses on the wide-ranging innovations shaping today’s business, regulatory, and scientific landscape. With guest speakers who work in a diverse set of fields, from artificial intelligence to genomics, our discussions examine not only on the legal implications of these changes but also on the impact they will have on our daily lives.