Should I kill my 5-year-old agency and build an AI product instead?

#StartupQA - “Should I kill my 5-year-old agency and build an AI product instead?”
That was the question in one of my 1:1 calls last week. Founder of a five-year-old agency, having large German companies as their customers, technical background. Enjoys the work. But every feed tells him the real money is in AI products.
My answer: No. Don’t kill it. Rethink it.
Companies could do a lot with Claude and Co. already. They just don’t. Not because it’s hard - because nobody wants to own the outcome. Whom do they yell at if the AI gets it wrong? Themselves. So they pay someone else.
So maybe that’s an opening worth pursuing: continue to be the agency, get paid that people shift the responsibility to you. Own the outcome and get paid for that.
If you seriously consider getting into a new business, give it a test drive before you commit: fake it for two weeks. Cold call without your brand behind you. “Hi, we’ve never done this before, want to buy?” vs. “Hi, we did this for XYZ.” Then decide.
Want to ask your question? Grab a 1:1 slot - link in the comments 👇
More from Startup Q&A
Everyone gives me advice. How do I know which to follow?
How do you spot founder talent before they even have an idea?
We're building a social venture, we just need money and a room
Do I need ISO certifications before I can sell to enterprises?
What value does a business founder bring to a deeptech startup?
