For finance teams ready to move

Your finance team.
Augmented by AI.
Responsibly.

A 4-week programme that takes CFOs and their teams from "we should try AI" to running real workflows on Claude — with the guardrails that make it safe to actually use.

Join the waitlist — first cohort is limited
✓ You're on the list. We'll be in touch shortly.
No spam. One email when the cohort opens.
4 Weeks
3 Live workflows built
100% Built on Claude
Online + on-site Format
Who it's for
CFOs and finance directors
Who want their team to stop wasting hours on mechanical work and start producing sharper analysis, faster — without depending on IT to make it happen.
FP&A, accounting & treasury teams
Who are already curious about AI but don't know where to start responsibly — or who've experimented alone and want a structured, shared approach.
Scale-up finance functions
Typically SaaS, e-commerce, or professional services — where speed and accuracy both matter, and the finance team is expected to run lean.
Not for you if...
You're looking for a generic AI literacy course, or you want to automate without understanding the risks first. This programme is built on precision, not hype.
The programme

Four weeks. Three modes: showcase, guardrails, build. Then a system that keeps improving after we leave.

WK 01
"This changes everything" — Use Case Showcase
Online · 90 min · watch-and-react
  • 6 live demos tailored to your actual job: variance commentary, board memo, FX exposure summary
  • "Before vs After" — we show the old way, then Claude doing it in 90 seconds
  • Homework: try Claude on one real task before Week 2, bring your result
Online
WK 02
"When to trust it, when not to" — Responsible Use
Online · 90 min · discussion-heavy
  • Where Claude excels: synthesis, narration, drafting, pattern spotting, formula building
  • Where Claude fails: live data, audit-trail tasks, regulatory sign-off, precise arithmetic
  • Data hygiene rules: what never goes in, how to anonymise, enterprise vs personal accounts
  • The CSI-FBI prompting framework — the system that makes outputs reliable
Online
WK 03
On-site workshop — Build Your Own Workflows
On-site · 3 hours · sub-team breakouts
  • The 5-Step Analysis Workflow: Verify → Propose → Calculate → Visualise → Narrate
  • Breakouts by function: FP&A, Accounting, Treasury — each builds 1 documented Claude Project
  • Configure Claude Projects with system prompts, knowledge docs, and example outputs
  • Everyone leaves with a workflow they can run solo on Monday morning
On-site
WK 04
"Lock in the culture" — Knowledge Sharing System
Online · 60 min + async tasks
  • Launch a #finance-ai-wins channel with 3 pinned rules and a use-case template
  • Use-Case Template: Problem → Prompt → Output → Time Saved
  • Designate 2 AI Champions per sub-team — they answer questions within 24h
  • 30-day challenge: 1 shared use case per person before the first Lunch & Learn
Online
After
Bi-weekly Lunch & Learns + Quarterly Hackathons
45 min · rotating presenters · tracked metrics
  • 20 min live demo of a real use case + 20 min Q&A, rotating by sub-team
  • Metrics tracked: time-to-first-insight, forecast cycle time, manual steps eliminated
  • Quarterly: each sub-team builds one new Claude Agent (2h hackathon format)
Ongoing
What AI is good for in finance — and what it isn't
✓ Where Claude genuinely helps
  • ✦ Drafting variance commentary and management narratives in seconds
  • ✦ Turning a P&L table into a sharp 200-word CEO update
  • ✦ Building dynamic Excel models from written instructions
  • ✦ Synthesising long documents: contracts, reports, board packs
  • ✦ Spotting patterns across large datasets when given structured input
  • ✦ Generating first-draft forecasting assumptions from historical context
  • ✦ Answering "what does this covenant actually mean?" on debt documents
✗ Where AI cannot be trusted unsupervised
  • ✦ Accessing live financial data without a verified connector in place
  • ✦ Producing audit-ready outputs — everything needs human sign-off
  • ✦ Regulatory filings, statutory accounts, or tax submissions
  • ✦ Precise arithmetic on large numbers — always validate calculations
  • ✦ Making judgment calls under genuine ambiguity
  • ✦ Replacing domain expertise — it accelerates experts, not beginners
  • ✦ Any task involving raw confidential client data without sanitisation
Format
4
Weeks
Three online sessions plus one full on-site workshop at your office. Designed around busy finance professionals — no all-day classroom sessions.
≤15
People per cohort
Small enough for real participation. Large enough to run meaningful sub-team breakouts by FP&A, Accounting, and Treasury.
100%
Built on Claude
One tool. One learning curve. No ecosystem confusion. We go deep on Claude specifically — including Projects, custom agents, and prompt libraries.
What your team leaves with
✓
Claude accounts + data rulesSet up correctly, with clear policies on what goes in and what doesn't.
✓
3 working sub-team workflowsFP&A, Accounting, and Treasury — each with a configured Claude Project and documented prompt library.
✓
AI Champion network2 champions per sub-team who keep the momentum alive and answer questions within 24h.
✓
Shared use-case libraryA living channel where wins are documented with the Problem → Prompt → Output → Time Saved template.
✓
Bi-weekly L&L cadenceA recurring session structure that keeps the team learning from each other long after the programme ends.
✓
Guardrails, documentedWritten responsible-use guidelines your team can actually follow — not a 40-page policy no one reads.

30 days.
That's all it takes
to shift the culture.

The first cohort is small and selective. If your team is ready to move from curiosity to capability, reserve your spot now.

Join the waitlist
✓ You're on the list. We'll be in touch shortly.
No spam. One email when the cohort opens.
Questions? Reach us at
[email protected]