What Intercom's transformation to Fin tells you about yours.

Not a rebrand in the cosmetic sense, but a full identity change. The help desk product still exists under the Intercom name, but the company is Fin, the domain is Fin, the email addresses are Fin. When they built their AI agent three years ago, they gave it a separate name deliberately. They needed the market to evaluate it on its own terms, not against the baggage of what people thought Intercom was.

The logic was straightforward: if you already have someone in a box in your head, it's hard to get them to see something genuinely new.

The thing Des said that stayed with me was about the nostalgia that gets thrown at founders when they make these decisions. People send him photos of their Intercom T-shirts. Former customers tell him they loved what it was. And his answer to that is essentially: you can't be the prisoner of other people's nostalgia. You have to have the conviction to move on, even when moving on means leaving something behind that people genuinely valued. Something I struggled with last year at SaaStock.

The performance numbers back the decision. Fin resolves over 70% of customer queries autonomously. They've tripled engineering productivity through Claude Code. They've reduced team sizes not by cutting people but by finding that individuals can now take on more, and that the ratio of engineers to output has fundamentally shifted. In areas where they used to run teams of eight to ten, including an engineering manager, a product manager, a designer and five engineers, those teams are now smaller and more capable.

On what makes something genuinely AI-native versus just AI-wrapped, Des had a test I haven't heard put better anywhere. Look at how the designers think. A real AI-native designer doesn't start with boxes, buttons, and drop-downs. They start by asking where the problem actually happens, and whether there's a way to connect to that system directly so that no one has to do the thing manually at all. The goal is to go from seven screens to no screens. Anything short of that mindset, he argues, is just cosmetic.

He also made a point about brand that's worth noting. In the current market, brand gets you into the consideration set. After that, it's performance that wins the contract. For anything significant, buyers are running proper bake-offs, giving three vendors a thousand real questions and scoring the outputs. Fin welcomes that, partly because the evaluation process tends to produce longer, more committed contracts once they win.

That shift in the sales motion is worth noting. Enterprise SaaS used to mean showing up with a deck and a feature checklist. Now it means accepting that the product has to prove itself under test conditions before anyone signs anything.

Next weeks newsletter will be part 2 based on my conversation with Des. If you can’t wait, I highly recommend you listen to or watch the podcast.

The AI Revolution Show is on Apple Podcasts and YouTube. The full conversation with Des is worth your time.