(Spoiler: it’s not the tools.)
For 18 years I led marketing inside companies. In 2026 I’m running B2Better as a solo consultancy. Same depth of work. A fraction of the headcount.
The question I get asked the most: how?
The honest answer: most of what gets pitched as “AI scaling for solopreneurs” doesn’t.
Here’s what does.
What ACTUALLY scales
- A locked Brand Operating Spec for every client
- Three AI agents max (research, drafting, reporting)
- One orchestrator (a Claude Project with the spec preloaded)
- A reporting layer that feeds back into the spec
- A 5-retainer ceiling I refuse to break
That’s it. Five things.
No 47-tool stack. No army of AI clones. No “vibes-based” prompt engineering.
What DOESN’T scale (even though everyone says it does)
- Adding tools when output stalls
- Onboarding a new client without a spec doc
- Generic prompt libraries written by someone who doesn’t know your client
- “AI does the strategy” — no, AI accelerates the strategy YOU bring
- Posting more without a feedback loop to what converted last month
Every one of these is a trap I’ve hit. Each one cost me at least a week of recovery before I noticed I was running backwards.
The 5-retainer ceiling
I cap at 5 retainer clients. Not 50. Not “scale to a 30-person agency.”
Five.
The math: 5 retainers at ~$7K average = $420K. Plus sprints and audits = $500K target. One human. No employees. No managers. No burnout.
This is the architecture I keep coming back to: trust over volume. Five clients who get the depth a $250K full-time CMO would have given them, at a fraction of the cost, with the AI stack doing the leverage work I used to do with junior coordinators.
Could I run 8? Probably.
Would the work hold? No.
- More clients = thinner specs
- Thinner specs = drifting AI output
- Drifting AI output = the engine becomes the chaos I built it to replace
The ceiling is the discipline.
What 2026 actually changed
The stat that made this real for me: 47% of solopreneurs are using generative AI to reclaim time from repetitive admin work, per Metaintro’s 2026 solopreneur tools breakdown. The monthly tool spend most are running? $100-$500. Compared to a $250K full-time hire.
The economics finally work for ONE person to compete with a small agency. Not because the AI is magic. Because the AI is reliable enough to remove the bottleneck that used to require a second human.
B2Better recently delivered a full content calendar, four long-form drafts, a keyword gap analysis, and a LinkedIn post series for a client in a single week, without a single subcontractor or coordinator. In 2023, that scope would have required at least one part-time contractor and a two-week turnaround. The stack did not replace judgment. It removed the bottleneck between judgment and output.
What I’d tell anyone starting
If you’re starting a productized B2B service in 2026:
- Lock the spec FIRST. Not the website. Not the tool stack. The spec.
- Pick THREE agents. Not seven. Not ten. Three.
- Build the orchestrator before you take client #2.
- Set the ceiling. Decide your number now, before the FOMO sets in.
- Build your reporting loop before you need it. Not after.
Most of the people I see struggling aren’t struggling because they lack tools. They’re struggling because they have ten tools and zero system.
The AI didn’t give us superpowers.
It gave us LEVERAGE.
What you do with the leverage is still up to you.
(For more on the systems side of B2B consulting, see my recent piece on digital marketing consulting.)
Question for the founders and fractional CMOs reading this: what’s the ONE part of your work that AI hasn’t actually scaled yet, no matter how many tools you’ve added?