Convictional shuts down, returns half its $49M after an AI-era pivot fails

Convictional is closing after a major shift in direction did not find the market it expected. The Kitchener-Waterloo company had built a profitable e-commerce dropship business with 3,000 customers and about $2 million in annual recurring revenue. In early 2025, it divested that business to pursue an AI-native alternative to Slack. Now it is shutting down and returning roughly half of the approximately $49 million USD it raised, including capital from YC’s growth fund, according to BetaKit’s report.

Founder Roger Kirkness put the problem plainly: the product they wanted to build was not what customers wanted, while the tools customers did want were being built internally. That is the business lesson here. AI can make a new product direction look obvious from the outside, especially when the market is moving quickly. But a working business with paying customers, repeatable revenue, and known demand is worth more than a compelling story about what might be next.

For owners, this does not mean ignore AI. It means test it against a real operating problem before betting the company on it. Start with a narrow workflow, a clear buyer, and evidence that people will pay or change behaviour. This is the kind of thing we build into an AI command centre, helping teams see where AI is producing measurable value rather than chasing a broad rebuild.