Why we consult differently
The opening move of a typical AI consulting engagement is the consultant telling the client what’s possible. Our opening move is the consultant watching the client’s team do the work. Before we recommend a single skill, connector, or workflow, we’ve spent real time — not interview time, shadow time — observing how the job actually runs.
Consulting based on observation is slower than consulting based on frameworks. It also produces recommendations that survive contact with reality. If what you need is a 40-slide strategy deck, we’re not your firm.
Three engagement shapes
Discovery Sprint (2 weeks)
For teams 4–8 weeks out from deployment who need answers built from evidence.
- Shadow sessions across 3–5 functions
- System landscape audit
- Workflow Map per function
- Skill backlog prioritized by observed pain
- Go/no-go recommendation vs. ChatGPT Enterprise, Copilot
Pricing: low-to-mid four figures fixed fee.
Advisory Retainer
For teams deploying Claude Cowork internally who want a specialist on call.
- Weekly 1-hour strategy session
- Unlimited async questions
- MCP architecture reviews
- Quarterly workflow-observation day
- Quarterly adoption review
Pricing: monthly, starting low four figures.
Full Implementation
For teams who’d rather we run it.
Four-phase fixed-fee engagement: Discovery, Foundation, Connect, Adopt. 60 days of post-launch support.
What makes us Claude Cowork specialists
- Discovery-led. Every engagement starts with observation. Our recommendations are evidence, not hypotheses.
- Product-specific. We consult on Claude Cowork, not “AI adoption” in the abstract.
- MCP-native. Our consultants have built 20+ MCP connectors.
- Mid-market focused. Right pricing, right timelines, right team size.
- Shipped, not pitched. Every consultant has delivered at least one full Claude Cowork implementation.
Who we work with
Fit: ~50–2,500 employee companies in the US and UK on Claude Team or Claude Enterprise (or actively evaluating).
Verticals we’ve worked in: Legal and professional services, Financial services / FinTech, B2B SaaS, Management consulting, Real estate, Media and publishing.
Not a fit: Consumer AI product companies, pre-seed startups, enterprises >10,000 seats.
What a Workflow Map actually looks like
Every Discovery Sprint produces one Workflow Map per function shadowed — the single most valuable artifact we hand over. A Workflow Map is a one-page visual that shows:
- The workflow as observed (not as documented)
- Every system the worker touches and the order of touches
- Decision points and the judgment rules behind them
- Bottlenecks, timed
- Pain/gain moments
- Opportunities ranked by impact × feasibility
- Skills we recommend building, in priority order
Your internal ops team will use these for things that have nothing to do with Claude.
FAQ
Not sure which engagement fits?
Book a 30-minute call. We'll listen, ask questions, and recommend the engagement shape — even if it's not us.