Client context
~40 people. Annual recurring revenue in the low eight figures. Founder-led, profitable, growing. Stated reason for the engagement: "We have invested in AI tools and haven't seen operating improvement. We don't know if the tools are wrong, the workflows are wrong, or we're wrong about what AI is good for."
What we mapped
- Three primary value streams: new business, delivery, and renewal.
- Twelve roles, fourteen tools, and the data flows connecting them.
- The decision cadences of the leadership team and the operators they depend on.
- The customer journey from first touch to expansion.
What we found
- The CRM and the delivery tool didn't share a customer record. Every handoff from sales to delivery produced a 30–60 minute re-keying tax. Two existing AI tools were generating notes that nobody was reading because they lived in the wrong system.
- Renewal conversations started at month 11. The signals were available at month 7 but nobody owned watching them. An "AI churn predictor" had been evaluated and abandoned because the underlying signal was not being surfaced anywhere.
- The leadership team made the same three decisions every quarter. Each time, the data pull took 4–6 hours and produced a slightly different conclusion. The decisions themselves were sound; the cost of asking them was disproportionate.
- Operator capacity was leaking into status reconciliation. An estimated 5–7 hours per operator per week spent answering "where is X?" across tools.
How we prioritized
We scored 23 opportunities on impact (operator capacity, revenue, retention, decision quality), effort (calendar weeks to value), and reversibility. The top-quartile recommendations were not the ones the team expected:
- Redesign the CRM-to-delivery handoff before automating it. Estimated ~80 operator hours back per month, plus an unlock for the existing AI tools.
- Build a renewal signal review at month 7, with a lightweight AI summary of usage, support, and account health pulled into one place.
- Standardize the three recurring leadership decisions with a templated input pack assembled automatically each quarter.
What the team executed first
The CRM-to-delivery redesign and the renewal signal review went into execution immediately. The leadership decision template was scheduled for the following quarter. The two pre-existing AI tools were repositioned inside the redesigned workflows; one was kept, one was retired.
What this Diagnostic was not
It was not a chatbot project. It was not a tool selection exercise. It was not a strategy deck. It was a working map of how the business runs, with a prioritized set of changes the team could start on Monday.