Metric Monday: Vendor Interactions Remain Steady Despite LLM Adoption

saraboostani

Happy Metric Monday!

In our buyer research this year, nearly every buyer (94%) reported using an LLM at some point during their buying journey, with usage peaking in the middle. At the same time, the number of interactions buyers had with vendors—both human and digital—did not change compared to prior years. Statistically, it’s the same as 2023 and one fewer than 2024.

Given how quickly LLMs have proliferated, it's an interesting dynamic to watch! The idea that “LLMs make us faster” surely holds in some areas. In my own work, they’ve taken a number of manual tasks off my plate. But they still require processing, validation, and follow-up. How much faster they make us—and how much work they truly replace versus simply support—remains a blurry line.

That’s also what the buyer research suggests. While buyers are using LLMs, they’re still doing roughly the same amount of work as before. They continue to engage vendors and vendor content at about the same rate. Based on the use cases buyers report, LLMs are largely helping them process and synthesize information, rather than replacing major parts of the buying process.

A quick pulse check: are you at the point where you’d ask an LLM what to buy—and from whom—without any fact-checking or additional work? That still feels a long way off to me.

We’ll continue tracking this in the years ahead. Who knows—maybe we’re closer than it seems. What do you think?

@Kerry

Kerry

A pretty long way off, I think. I've been having this conversation in person with people a lot lately. Think about what happens when you have a chat with your LLM of choice on a topic where you are knowledgeable. It can be incredibly gratifying to engage on a topic you are passionate about, because the LLMs are really good at meeting you where you are. But I bet you also notice some very odd gaps in their knowledge or understanding, or just execution of a response, and it leaves you scratching your head sometimes. Like they go from the calculus to messing up 2+2 on a moment's notice.

CarlyShideler

I don't think we're close to fully trusting LLM recs and just going wild with it haha. I def see myself as an consumer, and many others, using LLM to compare and see what it says. However, we are all still doing additional research and fact-checking

Kerry

I totally agree, Carly. And if you've ever had the experience of talking with someone about a topic they learned about entirely from ChatGPT (or similar), that seals your understanding that LLMs are still flawed information sources. They're improving rapidly, but still.