How AI Changes the Math on Solo Consulting
Solo consulting has always been capped by one person's hours. AI doesn't lift that cap with motivation or hustle. It removes the production work that ate your week, so a practitioner who could run two engagements at a time can often run three. Same expertise, same rates, roughly 40 to 50% more revenue. The judgment stays yours. The grunt work doesn't.
Why the old solo math had a hard ceiling
For decades the model was simple: you brought expertise, the client brought their mess, and you sold the place where the two met. Your value was what you knew. Your ceiling was how many problems you could personally touch in a year.
The wall wasn't your skill. It was the production tax: research that ate three days, a 40-page diagnostic written mostly from scratch, decks rebuilt for every engagement, hundreds of pages of documents read by hand. None of that was where you earned your fee. It was just the price of getting to the part where you did.
I used to think this ceiling was permanent, something you only escaped by hiring or quietly raising rates until clients flinched. I was wrong. The ceiling was never your knowledge. It was the manual labor stapled to it.
Is AI really an advantage if you're 45+ and already busy?
This matters more after 40, not less. A 31-year-old using Claude gets a competent assistant and not much else, because they don't yet have the pattern library to direct it well. You do. Twenty years of seeing how deals die in committee, which recommendations land, where a client is quietly resisting: that's exactly the judgment AI can't fake and can't replace.
So the gain scales with your depth. AI handles the drafting and the synthesis; you supply the discernment that tells good output from plausible-but-wrong output. The more you know, the more that division of labor pays. Your experience isn't a nice-to-have here. It's the moat.
Where AI actually moves the needle in a practice
Let's be specific, because vague promises are useless.
Research and synthesis. A policy consultant building a landscape analysis used to lose three to five days reading and organizing source material. Pointed at the same sources, PDFs, web pages, internal memos, Claude (especially in a project with its own memory) compresses the assembly to an afternoon. What to include, what to discount, what it means for this client: still entirely your call. The hauling is not.
Deliverable production. That 40-page diagnostic? The structure, the standard analysis sections, the framing of findings get drafted fast. Your hours go to interpretation and recommendations, the part that actually needs you in the room.
Client communication. Proposals, engagement summaries, status notes, meeting recaps. Necessary, rarely your highest-value work. Claude drafts, you edit. The writing usually gets sharper, because editing under no deadline beats writing under one.
Analysis at scale. An HR consultant running thematic analysis across dozens of interview transcripts, or a supply-chain advisor flagging risk provisions across a stack of vendor contracts: Claude reads all of it at once. The bottleneck stops being "can I process this in time" and becomes "what does this tell me."
What should solo consultants actually do with the time AI frees up?
When production compresses, you get a choice. Most solo consultants pick one of three, and plenty blend them.
| Path | What you do with the time | Effect on the practice |
|---|---|---|
| More clients | Take a third concurrent engagement | ~50% revenue lift, no new staff, no rate change |
| Faster delivery | Run the same engagement in half the calendar time | Charge a speed premium; win time-sensitive work |
| Better margins | Same volume, less time per job | Effective hourly rate climbs; fixed-price work gets far more profitable |
None of these requires raising your price or thinning your quality. They're just what happens when 35 to 40% of your production time comes back.
What does that look like in dollars?
Take a typical 52-year-old strategy consultant running three-month engagements at $30,000 each. Before Claude, she could realistically juggle two at once: $60,000 a quarter, roughly $240,000 a year before expenses. Hard ceiling. She was already saying "no" to good work because there were no more evenings left to steal.
Integrate Claude into research, synthesis, and deliverable production and the effective time per engagement drops about 35%. Now three run in parallel. Revenue goes from $240,000 to $360,000, a $120,000 swing, with the same clients, the same prices, the same standard of work. The catch is real but small: she now spends maybe two hours a week editing AI drafts that occasionally wander off, which is a rounding error against what she got back.
Nothing heroic about those numbers. That's just the arithmetic of getting a third of your production time back.
The 35% Audit: a simple first pass to find your AI-ready work
Don't reinvent your practice. Run one test first. For a single week, log where every working hour actually goes. Most consultants find 30 to 50% of their time sitting in work that's necessary but not distinctive: research, drafting, formatting, organizing, synthesizing. Call it your 35%. As a rule of thumb: if you wouldn't put that task on the cover slide of your case study, it goes in the 35% pile, those are the jobs you hand to Claude first.
Then attack the biggest block, not all of them. A quick filter: if you'd happily hand a task to a smart but green associate, you can probably hand its first draft to Claude. If research dominates, build one structured research-and-synthesis workflow with Claude and run it until it's boring. If it's report writing, start with the structural sections of your deliverables. If it's proposals, draft from your rough notes. One workflow at a time, biggest payoff first. You're not rebuilding the practice. You're reclaiming the 35%.
The part AI still can't touch
Be honest about the limits, because overselling this is its own kind of amateur hour. Claude doesn't replace the judgment that comes from having seen a hundred situations like your client's. It doesn't read the room when a client team is dug in and needs a different approach. It doesn't know which recommendation will survive the committee and which one dies on slide nine.
That's the work you were always actually paid for. AI clears the lower-order labor that kept stealing time from it. Which is the whole point: the value you bring is precisely the part that can't be automated, and now you get to spend more of your week there.
If everyone uses AI, doesn't the edge disappear?
It's the most reasonable objection, and it's backwards. Your edge was never in producing documents. Anyone could produce documents, slowly. The edge was always the depth of what you know, the quality of your calls, and the trust clients put in you. AI compresses production for the whole field at once. Expertise and judgment stay scarce. So the experienced practitioner's relative advantage doesn't shrink. It widens.
Is this hard to learn if you're not technical?
Lower than you'd expect. Working with Claude is closer to briefing a sharp associate than writing code: plain language, back and forth, no syntax. The actual skill is framing the problem clearly and saying what "good" looks like. That's something you've done for twenty years every time you handed off work to someone junior. You already have the rare half.
Start Monday. Pick the single biggest time sink from your week, hand its first draft to Claude, and edit the result. Do that one task for two weeks before you add a second. The practice that quietly doubles its capacity isn't the one that overhauls everything in a weekend. It's the one that moves one workflow at a time and never looks back.