Robert Ta argues AI agents could replace the $80 billion market research industry. A startup reproduced a multi-month consulting study overnight with better results. The other speaker warns that AI simulations often reduce people to stereotypes instead of capturing how customers actually feel.
If simulated agents can replace focus groups, what happens to the $80 billion market research industry? I saw this other startup where they replicated one of the big consulting firms' market research studies over a 6 to 8 month horizon, if I recall correctly, overnight. And they produced even better results than the original market research. This is a really lucrative industry. Like if you ever work in enterprise software and you went to go Google some shit about like a market, then you're going to see reports worth thousands of dollars. And people pay that, business pay for that because it's signal from market of the territory and how the territory is changing to capture a snapshot of a map of like what the territory is to bring that back to the people on the signal to ultimately impact decision-making towards the bottom line. I think that $80 billion market research industry was just proven to be disruptible. It makes a lot of sense. Why not run at-scale simulations that are domain-specific, tied to a particular problem for your own business and ecosystem that you would care about, versus paying for some more generic market research conducted by people out of your company. Why not just bring that capability in-house with tools like this?
And one thing I'll say is that a lot of the examples you give, you're kind of looking at a group of people in the abstract and you're creating stereotypes or like stick figure sort of representations of who who people are. And I think what similarly, what the language that they use about their simulations appeal to me is that you can put yourself as a human in their environment and like, what is this customer feeling when they're like given the adherence protocols for a prescription, for instance? And I think that the ability to like put ourselves in the shoes of one another and to like take not an abstract representation of a human, but a phenomenological and felt like relationship with the customer is where I hope that AI brings us to, because there's a lot of risk about just like continue to look at populations of people in the abstract as these stereotypes. And I think what it has the power to do is bring us closer to one another and closer to the felt experience of being a human, being a consumer, making decision in a certain situation.
