Journal
Field notes on running a consulting practice with AI in it. No hype, no benchmark recaps — just what holds up on client work.
New here? Start with these
-
The three categories of consulting work — and which ones AI should never touch
Where the line sits between work AI should touch and work it shouldn't.
-
Stop prompting, start briefing: delegating whole deliverables to Claude
The habit change that fixes most bad AI output: brief it, don't prompt it.
-
Stop pasting into chat. Point Claude at a folder.
The setup post — where your files live and how Claude gets to them.
What's left to review when the AI checks its own work
Newer models check their own drafts before you see them. That raises the floor and makes the polish more convincing — here's the review that stays yours.
The effort dial: how much thinking does a task deserve?
Claude Opus 5 lets you dial how hard it thinks. Which consulting tasks earn maximum effort, which are actively worse for it, and how to set defaults.
Which Claude model you actually get now depends on your plan
Fable 5 is included on Max, not on Pro. Opus 5 is the new default. What the split actually means for a consulting practice — and when to pay for...
What a paid template kit has to beat
We sell a paid Claude template kit, so read this skeptically: what a paid kit has to beat, the checklist to hold any kit to, and who shouldn't buy one.
The client update email template (steal this one)
A copy-paste client update email template: five sections, under 200 words, with a filled-in example — and the extraction rules for what stays out.
Claude Cowork left the desktop. Here's what that buys a consultant.
Claude Cowork now runs on the web and your phone — sessions that keep working with your laptop shut. What that changes for consultants, honestly.
From transcript to deliverable: the AI-assisted research-report workflow
The four-stage workflow for turning stakeholder transcripts into a client-ready deliverable — where Claude helps, where it doesn't, and the synthesis trap.
The 30-minute monthly pipeline review that doesn't get postponed
The four-step monthly pipeline review that takes 30 minutes, catches deals dying of neglect, and ends with one specific action — no retro overhead.
The one-pager you should write before every AI-assisted task
The 5-section brief consultants write in 90 seconds before any AI task. Why it turns generic output into specific, and the format that compounds.
Stop pasting into chat. Point Claude at a folder.
Claude can work in your folders now — read raw notes, write finished deliverables. What that changes for consultants, and the folder anatomy that works.
The scope-creep conversation: four scripts for the conversations you've been avoiding
The four scope-creep patterns most solo consultants hit, with the specific response language for each — without sounding defensive or apologetic.
How to keep a client's context across Claude sessions (without re-pasting)
The folder structure, the per-engagement Project setup, and the three context files that make Claude operate like it actually remembers your client.
Every article, by topic
The complete Journal index. Bookmark this page — new pieces land here on publish.
Setting Up Claude
- Stop pasting into chat. Point Claude at a folder.
- How to keep a client's context across Claude sessions (without re-pasting)
- Claude Projects vs. skills vs. raw prompts — when to use which
- Claude skills for solo consultants: what they are, and the first three to build
- The actual Claude workflows of a solo consultant, six months in
Working With AI
- What's left to review when the AI checks its own work
- The effort dial: how much thinking does a task deserve?
- From transcript to deliverable: the AI-assisted research-report workflow
- The one-pager you should write before every AI-assisted task
- The context-window mistake most consultants make with Claude
- Stop prompting, start briefing: delegating whole deliverables to Claude
- Why your AI prompts feel generic — and how solo consultants fix it
Winning Work
- The case-study harvest: one wrap, three marketing pieces, thirty minutes
- An 8-step proposal walkthrough on a real (fake) engagement
- The non-numeric outcome — what stops case studies from reading like marketing fiction
- Why most consulting proposals lose — and the one shape that closes
- The 90-second discovery dossier (when you don't have an SDR)
Running Client Work
- The client update email template (steal this one)
- The scope-creep conversation: four scripts for the conversations you've been avoiding
- The five-section Friday update that keeps clients quiet
- The first hour after a deal closes: what to ship a new client
- The client-meeting workflow that turns transcripts into next actions
Practice Systems
Model Watch
- Which Claude model you actually get now depends on your plan
- Claude Cowork left the desktop. Here's what that buys a consultant.
- Don't build your practice on one AI model
- Fable 5, Opus, or Sonnet: which Claude for client work
- Claude Fable 5 for solo consultants: what actually changes
- I picked Claude — here's the one task ChatGPT actually wins