Start with cost, not capability
The instinct in most AI adoption conversations is to start with what the technology can do. That's backwards. Start with where your operating cost is actually going — payroll hours on repetitive tasks, response-time lag that's costing you leads, manual reporting that eats a day a week from someone valuable.
Once you have that list, AI capability becomes a filter, not a starting point. You're no longer asking 'what could AI do here?' in the abstract — you're asking 'does an AI tool solve this specific, already-identified cost?'
Rank by effort-to-payoff, not novelty
It's tempting to start with the most impressive-sounding automation. We rank opportunities instead by a simple ratio: implementation effort versus measurable payoff within 90 days. The unglamorous automations — intake forms, lead routing, report generation — usually win that ranking, and winning early builds the internal trust to tackle bigger initiatives later.
Assign a number to every recommendation
Every item on an opportunity audit should have an estimated hours-saved or cost-reduced figure attached, even if it's a range. Vague recommendations don't survive budget conversations. Specific ones do.