Over the past few months, investors have been asking founders and CEOs of their portfolio companies how they can integrate AI to improve their product. It’s a valid question, but it’s less interesting and revealing than the question I’ve been asking myself and fellow founders when discussing the rise of AI: How is AI an existential risk to your company?
4 Ways To Survive in the Age of AI
- Look past what is already here and identify edge cases.
- Look for paradigm shifts in your industry.
- Tend to your infrastructure.
- Identify opportunities that AI creates.
By asking the question now, you can start to think about the enduring value your company can contribute to a market in which AI is not just incipient but flourishing. In the long run, it’s much more productive to ask, “What will AI force my company to become in the next ten years?” rather than “How can I hop on this trend train for the next six to 12 months?”
Just as being too slow to recognize foundational shifts from AI will endanger companies, so too will making knee-jerk reactions to your product that turn out to be more predictive than factual. Here’s what we are doing at my company, Stytch, to prepare for the AI onslaught.
Identify Edge Cases
Most of the discourse around AI today revolves around ChatGPT’s current capabilities, but very few companies are talking seriously about additional features and products that are not quite proven out yet.
Based on current AI capabilities, authentication and fraud and bot protections like ours only become more relevant (and we’ve already done some thinking about the implications of current AI technology for our industry). The edge case we’re really thinking about are things like AutoGPT, an extension of ChatGPT which introduces the concept of autonomous AI agents.
Rather than needing explicit direction from the user like ChatGPT does, AI agents simply need general directions like “budget my finances,” “pay my bills,” “make a dinner reservation,” etc., and uses self-prompting to figure out the steps required to accomplish your objective.
Look for Paradigm Shifts
It’s easy to look at immediate ways AI is accelerating current trends in labor, programming, and data processing. But the companies to look at are the ones who are doing something you’ve never heard of, or never thought of before. Those are the technologies that could easily take you by surprise in a decade — or sooner.
At Stytch, we foresee big paradigm shifts in cybersecurity and authentication. One is in the realm of identity and access management. The entire foundation of identity and access management is based on building one specific profile for roles and access permissions per person or machine.
But in a world where a person has bot agents acting on their behalf, it’s likely they’ll want to set a different, more restrictive set of permissions for that AI agent than they would set for themselves. Current role-based access control infrastructure isn’t designed to handle this kind of scenario, even in a B2B context that has organizations and members with variable permissions.
Another is in authentication. Authentication flows today are predicated on optimizing the UI for a human being: people need arrows, instructions, buttons and real-language error messages to guide them through a login process. In a world where bots are communicating directly with access management systems, the entire foundation of our user interfaces would come undone.
Don’t Neglect Infrastructure
A lot of the buzz around AI today, especially in larger news outlets, revolves around art generation or ChatGPT’s ability to write a passable college essay. But AI’s deepest implications will likely be seen on the backend as they challenge or render obsolete current infrastructural systems.
Per the examples described above, both scenarios deriving from AI agents involve fundamental shifts in the premise of our business: what a user is, how they are identified and what they need are all totally different in a world where AI navigates the web on behalf of individuals. And what they require of our business are much more fundamental shifts in how we set up databases, define terms and permissions. Instead of accelerating processes, the AI agent threatens to overhaul them entirely.
The abiding optimism of Silicon Valley can verge on a cliché, but the whole point of considering the death of your company is to embrace the possibility of a rebirth. That means thinking unflinchingly through the edge scenarios you’ve identified and applying the same innovative mindset that helped you build your current company to planning what you could become.
In this future of AI agents, Stytch’s focus would need to shift toward creating AI-friendly authentication interfaces and anticipating how to manage more complex access controls as apps begin to provide different entrances for different types of users: real human-beings and their permissioned AI agents. The rise of autonomous AI agents in our digital lives could remove a ton of tedium for humans.
The whole point of considering the death of your company is to embrace the possibility of a rebirth.
If we return to Stytch’s mission, which is to eliminate friction from the internet, these advances fall in lock-step with our purpose. In light of that, our goal is to be ready to support these processes by anticipating those infrastructural changes now, so we avoid the future in which other startups are better architected for this shift than us.
While the analysis of Stytch above sounds rather clinical, I’ll be the first to admit AI’s rapid development and potential impacts on my company definitely scare me. I also have found that Stytch will have a much brighter future if my co-founder and I think these scenarios through now instead of waiting to see what happens.
As CEOs and founders, we must ask ourselves the uncomfortable questions. I hope this blueprint of how we’re thinking through our own death at Stytch helps you look at the future a bit more creatively too.