Every founder call this quarter has included the same line. “We want to do something with AI.” We never say no. We just ask three questions first.
Question one: is your CRM accurate enough to trust a hiring decision?
If the answer is no, AI forecasting will not fix it. It will hallucinate against dirty data, produce a confident number, and you will hire against it. The cost will show up nine months later when the pipeline does not land and the hires are carrying nothing.
Question two: do your leaders trust the numbers without rebuilding them?
Most leadership teams we meet rebuild the numbers in a spreadsheet before every board meeting. That is a trust problem, not a tooling problem. AI dashboards will not earn that trust back. If anything they will widen the gap between what the system says and what leaders believe.
Question three: are your sales stages evidence-based or opinion-based?
If stages move on feel, the AI will learn the feel. It will optimise the feel. Your sales motion will get better at producing opinion-led forecasts. Faster.
AI does not fix broken foundations. It accelerates them. If the data is unclear, AI scales confusion.
What actually works
Fix the foundations first. Clean the CRM. Define the stages. Install the reporting rhythm. Get leaders trusting a single source of truth. Once that layer is live and the people running it trust it, the AI layer has something real to optimise against. That is when the investment compounds.
We are not anti-AI. We are pro-sequence. Foundations first. Automation second. The businesses that do it the other way round are the ones still rebuilding their forecasts manually in two years’ time, with an AI tool running in the background producing confident, wrong answers.
AI is a multiplier. Multipliers work both ways. Make sure the thing you are multiplying is worth multiplying.
