A 30-YEAR CPA'S FRAMEWORK FOR FIGURING OUT WHICH 10-15 HOURS OF YOUR WEEK AI SHOULD BE DOING

Twelve years in Big Six public accounting teaches you exactly what billable time looks like — by the hour, by the task, by category. Every six minutes of the day has a code, and every code has a value. The same logic applies to everything else: the emails that need crafting but not expertise, the research that needs assembling but not judging, the first drafts that need structure but not the specific competence only decades of practice produce. For years, both categories — expert work and infrastructure work — traded at the same effective rate. Not because anyone didn't know better. Because there was no better option.

There is now.

The Distinction That Determines Everything

AI doesn't replace expertise. It handles tasks that require competence but not the specific, earned competence built over thirty years. The man who learns to make that distinction clearly — the way a CPA distinguishes what the client is paying for from what the firm has to do to produce it — will recover ten to fifteen hours per week currently consumed by tasks that require his presence but not his expertise. Most professionals over fifty who run the audit below are surprised by that number. They shouldn't be. The billable-time logic was always pointing here.

The Automation Audit

List your fifteen to twenty most time-consuming recurring tasks — professional, administrative, communications, research, planning. Be specific and honest about what actually consumes your days, not what's supposed to. Then run this prompt:

"Here is my list: [everything — email drafting, report prep, research synthesis, scheduling, document review, financial monitoring, presentation building, meeting prep, follow-up tracking]. For each task, categorize: Category A — AI can handle this substantially or completely, I review before use. Category B — AI handles 70-80 percent, my judgment provides the final 20-30. Category C — my specific expertise or relationship is irreplaceable, AI is supporting only. Estimate hours per week each currently consumes."

The Category A list is consistently longer than expected. Most professionals who complete this audit find thirty to forty percent of their weekly hours sitting there — first-pass research, template communications, presentation structure, meeting agendas, background prep. None of it requires what you specifically know. It requires time and competence, and AI provides competence reliably, at scale, in seconds.

Three Entry Points to Redeploying the Recovered Time

Content publishing — expertise translated into guides, books, or frameworks that don't require your time present in every transaction; AI handles research assembly and structural drafting, your judgment provides what makes it credible. Expert service packaging — AI prepares client-specific analysis before engagements and drafts deliverables after, so the ratio of billable output to total hours invested improves materially without compressing the expert work itself. Advisory and fractional consulting — AI compresses the time required to stay current and prepare across multiple simultaneous engagements, the kind of relationships where the premium is paid for pattern recognition, not hourly presence.

The man who has been pricing both categories of work at the same effective rate has been undercharging for the expert component by the full cost of the non-expert work surrounding it.

The Bottom Line

None of this is passive in the sense of requiring no work. What changes is the ratio of expert time to non-expert time within the work you're already doing — and that ratio is worth a specific, calculable number of hours per week once you actually audit it instead of assuming it.