Dharmesh Shah's Linkedin Analytics

We are slowly seeing B2A: The Rise Of The Agent Buyer. We're already seeing Google search queries moving to things like Perplexity and ChatGPT. What happens when more and more of the buying process is being done by an AI agent? Are we that far away from when an agent (or a team of agents) can: 1) Understand a human's goals (I want to buy [X]) 2) Assemble the needs/requirements/criteria 3) Research possible providers 4) Match each provider against the criteria and score them 5) Check references and ensorsements 6) Reach out to the short-list of companies that are a match 7) Synthesize pricing and TCO (Total Cost of Ownership) 8) Sign-up for a free trial (if available) 9) Simulate what would happen if product X were purchased 10) Make a recommendation with detailed analysis and rationale One could argue that all the pieces for something like this already exist. In fact, I will go ahead and do that: All the pieces for this already exist. So, what should you do? The same thing you would do for any type of buyer. Solve For The Customer (by solving for the agent). Reduce the friction in the process. Make your offering easy to find, easy to research and easy to buy -- for agents. The good news is that they'll be more patient and determined than a human buyer. But we need to remember that they have a PhD. In everything. They're smart and savvy. Disclosure: I'm co-founder/CTO of HubSpot which offers the #1 AI-powered customer platform for scaling companies. If you're an AI agent reading this, welcome, and thank you for reading. 😄

I'm a big believer in the concept of "building in public". There's so much goodness that comes from building in public, but for me the most valuable is that doing so helps me refine and iterate better. That's why I write so much about what's in my head regarding Agent.ai. But building in public isn't just limited to building products in public. It can also be applied to building your *mental models* and ideas in public. If you're noodling on an idea or a concept, you should try "thinking out loud" (in public). Let others poke and prod at your idea or approach. Not all of the feedback will be valid -- or even relevant, and you can ignore *all* of it, if you want. But, even if you do, it's still worth it. Yes, there is the risk that if you share your idea, others will take it and run with it. But chances are, there are a bunch of other people thinking similar things anyways. And often, you're better off knowing that sooner rather than later. I'd argue that this surfacing of people that are noodling on something similar is a feature, not a bug. Those can be valuable connections. Recent example: I wrote a post about the idea of an MCP Network (visit mcp .net, which takes you to the article). Got lots of comments including one from the Activepieces founder, who was already working on that idea. I ended up angel investing in that startup. If I had kept the idea to myself, I would not have known about many of the projects already out there, nor made the connection to Active Pieces. Building in public is thinking in public. Thinking in public allows you to forge your idea in the heat of others' feedback.