The Founders Guide to Leveraging AI
Eight CEOs outline how they’re adapting to the GPT Era.
🌟 Hey there! This is a subscriber-only edition of our premium newsletter designed to make you a better investor and technologist. Members get access to the strategies, tactics, and wisdom of exceptional investors and founders.
When historians reflect on our era, there’s a good chance they’ll spend a great deal of time on November 30, 2022. That was the day OpenAI launched ChatGPT and sent AI hype into overdrive. It is not an exaggeration to say that the world has not been the same since, roiled by successively brilliant foundation models and the applications leveraging them.
For startup founders, the Age of GPTs presents extraordinary opportunities but also introduces severe risks – many of which may not yet be well understood. Essentially overnight, CEOs have had to get to grips with this new technological breakthrough, chart a path to incorporate it and devise the construction of fresh moats. In a few years from now, we’ll get to see who managed this period best.
To help readers understand the issues at play and how to adapt best, today’s edition of The Braintrust focuses on these topics. We’ve asked eight unicorn CEOs and founders to share what they make of the AI revolution, how they’ve adapted to it, and what they’re doing to plan for the future.
I hope you like it! To unlock the full guide, join Generalist+, our premium newsletter tier, for just $22/month. You’ll unlock all of our founders’ guides, along with a host of other exclusive features designed to make you a sharper investor and technologist.
Brought to you by Vouch
For startups, the recruitment journey is fraught with challenges as critical as they are complex. Missteps can create serious legal and financial repercussions for your company. Vouch is more than a business insurance provider; we understand the challenges facing scale-stage startups go beyond common risks, reaching into the nuances of hiring.
Vouch offers your startup a reliable partner, including Employment Practices Liability coverage to shield your company from the hazards of recruitment mistakes. Whether it's wrongful termination or discrimination issues, we protect your business, freeing you to build a team that reflects your company's values and aspirations.
Ready to tackle the complexities of recruiting? Learn how to protect your business as you scale your team.
Quick hits
If you only have a few minutes to spare, here are eight highlights from The Braintrust’s contributors on leveraging artificial intelligence.
Mathilde Collin outlines how Front quickly integrated AI to avoid being disrupted.
Jack Altman describes how AI actually opens the door for more human connection.
Trae Stephens explains why the current AI boom is more applicable to Palantir than Anduril.
Pedro Franceschi shares why finance will be heavily impacted by AI.
Immad Akhund warns executives not to get too wrapped up in the hype cycle.
Christina Cacioppo discusses the startling efficiencies of LLMs for Vanta’s business.
Avlok Kohli doesn’t worry about the cost of GPUs, given AngelList’s business model.
David Hsu talks about the importance of customization in benefiting from AI systems.
Read on to discover the full list of tactics and strategies.
“Seat-based software is fundamentally at risk”
Mathilde Collin, CEO and Co-founder at Front
There’s not a single day where I don’t talk and think about AI. AI has the potential to disrupt a lot of industries, and there are a few industries that will DEFINITELY be disrupted. Customer service (where we operate) is one of them, so we had to make sure we were investing in AI early on. For Front, this looked like:
Acquiring AI talent. We acquired Windsor.io to accelerate the speed of our AI innovation. Acquiring a team that has spent the past few years immersing themselves in the AI space (with the added bonus of proving they know how to work well together) will help us deliver new AI capabilities even faster than before.
Shuffling teams to bring more people to our AI engineering team. In addition to the acquisition mentioned above, we reconfigured our team in a few important ways. We had many ideas on how to harness the latest developments in AI to use within our product, but we wanted to confirm that AI 1) worked on the datasets and use cases that our customers cared about and 2) leveraged the product we had already built. As a result, we had our co-founder and CTO act as the product owner so that we had a strong technical leader who could utilize his deep understanding of our existing product as we explored our AI use cases. This allowed the team to move very quickly and learn fast.
Infusing AI in every single team. After our AI engineering team built several features end-to-end and we confirmed answers to the above, we then moved to the next phase of the team’s evolution: cementing these capabilities into a common platform. The AI engineering team formalized many of the capabilities we confirmed work across the Front use case and made them available to all of the feature teams across the product development org. This means that any team can build AI-enabled features and leverage AI capabilities without having to go through the difficult and time-consuming R&D work that the initial team had to do. Of course, every team includes those outside of engineering. Our company runs on Front, so every employee is a beta tester for our features. We are focused on operational efficiency, and AI plays an integral role in unlocking that efficiency by automating routine tasks. For example, Front’s support team uses AI tagging to automatically categorize messages based on the content of the conversation. (We then use tags to track trends, trigger automatic workflows, etc.)
Changing our messaging to incorporate AI. We’re seeing that buying decisions are already being influenced by the availability of AI today, so it’s important to show your product is meeting (and ideally succeeding 😉) the expectations of the market. We did this in phases. Last spring, we built up anticipation for our AI features with a teaser demo and a signup list to beta-test them, quickly followed by the launch of the features themselves. Then, in the fall, we held our inaugural Frontlines event, where we launched more features and shared our roadmap that highlighted the additional ways we were incorporating AI into our product.
Rethinking pricing and packaging. Seat-based software is fundamentally at risk. The market shift has resulted in pressure for businesses to do more with less, and advancements in AI mean that this is possible in ways that it wasn’t before. Even if the market recovers, there’s no guarantee that companies will have the same headcount as before. Leaders need to think about how they’re going to incorporate AI as well as how to implement a business model that doesn’t rely solely on seat-based pricing.
“AI doesn’t just make processes more efficient; it enhances our ability to connect and engage on a human level”
Jack Altman, Co-founder and Executive Chairman at Lattice
On some level, the debate around AI can be boiled down to a difference in your fundamental perspective on the technology’s impact. Do you believe AI will mainly allow us to do the same things we currently do more efficiently, and end up replacing many of the jobs that exist today to do those things in the process? Or, do you believe that AI can bring superhuman productivity to people in a variety of roles – and rather than replacing these jobs, productivity will actually skyrocket, and we’ll be able to do things we couldn’t before?
The reality is that this is still an open question, one we’re answering in real time and one that will vary field by field and industry by industry. We can’t say exactly what will happen because the technology is still developing, and we can’t clearly see where it’s headed.
Personally, I’m an optimist, especially when it comes to technology and technology’s impact on humanity. So I tend to believe that in more cases than not, as AI technology develops, it will greatly increase people’s ability to accomplish things – and in some cases, we’re going to see people do things we never thought were possible before.
I think this is especially true in HR – Lattice’s 2024 State of People Strategy Report actually found that HR teams are actively looking for ways to implement AI to help their organizations be more effective, especially in augmenting managers. It may seem counterintuitive, but I believe AI can actually empower HR teams to provide a more personalized and human experience, especially as these teams are increasingly expected to achieve more with fewer resources. AI doesn’t just make processes more efficient; it enhances our ability to connect and engage on a human level.
That’s a big driver behind Lattice’s own investment into AI. Just this month, our new CEO, Sarah Franklin, announced a slate of new AI products that the company will roll out over the coming months. The first feature we announced, AI Engagement Insights, will synthesize every employee survey comment in a matter of seconds – not days. Once your survey is closed, you’ll receive key trends, summaries, and recommended actions, allowing HR teams to more quickly and effectively drive positive change for their people.
We are currently just looking at the tip of the iceberg. There are so many ways AI can augment teams, and I, for one, am here for all of it.
“It took some companies up to thirty years for the internet to eat their world”
Trae Stephens, Executive Chairman and Co-founder at Anduril
It would be disingenuous to say that anyone was fully prepared for what GPT-3 would do to the tech market. Even OpenAI was projecting a tiny fraction of the revenue they ended up generating in the 2023 calendar year. However, it’s been broadly understood that there was going to be a turning point with new developments in AI. It was a matter of time until we “crossed the Rubicon,” and AI was good enough that it had day-to-day utility. With the public launch of GPT-3, it became clear that AI was going to be a major part of the future, and it was a matter of figuring out the right initial opportunities for the tech.
The most interesting edges of AI for Anduril are very different from the most interesting edges for a company that’s working with natural language since Anduril is most focused on reasoning models, computer vision, and sensing. These use cases aren’t as natural of fits for where we are today with language models, but we’re getting closer. The AI boom of the last year has been more interesting for companies like Palantir and Scale since the LLM infrastructure that has pushed the edges of AI in the last 18 months is more easily transferable to their core businesses.
Though AI in its current state is more applicable to some companies than others, every industry needs to ask themselves questions about how they can leverage the state of the art in AI to build efficiencies into their businesses. AI will cut across every industry in a similar way to how the internet did. It took some companies up to thirty years for the internet to eat their world (and maybe it will take that long for AI to do the same), but everyone should be on the lookout for that to happen again, and perhaps more rapidly.
As for Anduril, like any place that employs a lot of talented software developers, many people are drawn to investigate how new technologies can make them more efficient and effective at their jobs (and, in our case, how this new technology can positively impact the warfighter). Finding those kinds of efficiencies is constantly on our minds. We’re looking forward to seeing where these models go, particularly as they get better at the perception, sensing, and reasoning use cases that are core to Anduril’s business.
“[W]e’re truly in an unprecedented era”
Pedro Franceschi, Co-founder and Co-CEO at Brex
AI has already made an enormous impact on the ways that people across a range of industries do their jobs every day. The same is true for the finance sector, an industry where massive volumes of unstructured data, complex internal processes and external regulatory requirements, and other historically manual workflows present perfect opportunities for AI assistance and automation.
Given that, AI and automation feature heavily in how we think about our platform and how we serve our customers. While AI has always been core to Brex’s product, we’ve been investing more heavily in the technology this past year. In fact, in September 2023, we released Brex AI and Brex Assistant to transform spend management and enable unprecedented levels of efficiency, compliance, and accuracy.
Ultimately, the future of finance will be shaped by financial leaders’ willingness to embrace and adapt this powerful technology – and by understanding where and how to apply it. We’re focused on enabling these leaders to use AI in a way that amplifies their human effort and boosts productivity.
There are, of course, those who are skeptical about the tech, and we’ve had folks ask us, “Do the benefits outweigh the AI risks, including data privacy and intellectual property concerns?” This is something that is top of mind for us as well, and since the technology is still emerging, there is room for error and risk if the right approach and tools are not chosen. It’s why we wrote this guide to educate decision-makers in companies who may be reviewing AI tools for internal use from a procurement, risk management, legal, and or data/privacy perspective.
On a personal note, I’ve been really excited about all of the energy and enthusiasm around AI. AI companies were a bright spot in a sluggish fundraising and deal-making environment in 2023 (plus, they seemed to breathe some major energy into tech hubs like San Francisco), and I have a feeling that 2024 will bring about even more transformation and opportunity. I recently had the chance to speak with early-stage founders at a YC event, and to no surprise, many of the awesome entrepreneurs there were building applications and tools around AI. They made me reflect on two things: 1) that we’re truly in an unprecedented era, and 2) that human ingenuity will always be beautiful and irreplaceable.
“I do think it’s dangerous and unwise to just implement AI for the sake of getting in on the hype cycle”
Immad Akhund, CEO and Co-founder at Mercury
The first most important thing is to just be aware of everything going on. If you’re an entrepreneur – especially a software entrepreneur – you have to be up to date on the latest developments, whether that’s by trying out consumer AI or reading articles about it. That’s just the bare minimum, so you can at least understand what’s possible. For most startups, AI is going to be an opportunity, not a challenge. One of the benefits of being a startup is that you can easily adapt to change and be more nimble.
That said, I do think it’s dangerous and unwise to just implement AI for the sake of getting in on the hype cycle. If you have your roadmap built out, and AI doesn’t fit into it naturally, or there’s no compelling reason to change course, you should prioritize listening to customers and proceeding with your normal product development.