Claude Code in Production · AI Enablement for Technical Teams

Your team has Claude Code. Turn access into compounding value.

A one-day, hands-on workshop for technical teams in regulated or data-sensitive sectors who already have Claude Code and aren't yet getting compounding value from it. Not an introduction to AI tooling: governance, cost discipline, and patterns that compound, built from your own codebase on the day.

Get in touch See what the day covers

Who it's for

Teams of 5–30, already using Claude Code.

Technical teams in regulated or data-sensitive sectors (financial services, healthtech, legal, insurance, public sector, or similar) who already have access to Claude Code and are not yet getting compounding value from it. Not a session for teams evaluating whether to adopt AI tooling at all.

In scope

Enablement, only

Teaching your team to configure, govern, and get compounding value from Claude Code you already have.

Out of scope

Not implementation, not retainer

No build or deployment work, and no ongoing retainer support: this is a fixed-price training day, not a step in a longer engagement.

What the day covers

09:30–17:30. Around six hours 45 minutes of session time.

Built from your own pre-session diagnostic answers (who's using Claude Code, for what, what's configured, what's currently ungoverned), not a generic deck.

09:30 · 45 min

Where you actually are

The team maps its own current usage on the wall: who uses it, for what tasks, what's configured, where data is going. Everything after this is specific to what's on that wall.

10:15 · 45 min

The delegation principle

Use the model for semantic reasoning over unstructured text, and delegate arithmetic, verification and decision-gating to deterministic code you can test.

11:00 · 15 min

Coffee break

11:15 · 75 min

Build your CLAUDE.md

Hands-on, not a template handout. The team writes theirs: coding standards, approved libraries, security rules, project context.

12:30 · lunch

Lunch

13:15 · 60 min

Permissions, secrets and data classes

What Claude may touch, what must never leave the building, and review-before-apply workflows. The person accountable for security needs to be in the room.

14:15 · 45 min

Cost and model routing

Why token spend escalates quietly, what session discipline looks like, and routing strategy. The team leaves with a written usage policy.

15:00 · 15 min

Afternoon break

15:15 · 60 min

Patterns that compound

Skills, subagents, and structuring work so month two is better than month one rather than the same.

16:15 · 45 min

Capstone: one real task, end to end

The team picks a genuine item from its own backlog and runs it through everything from the day.

17:00 · 30 min

The 30-day plan

Pilot group named, task categories agreed, what gets measured, and when you review.

What you walk away with

Artefacts, not a memory of a good day.

HONEST SCOPE

No enablement client has been through this workshop yet.

No case studies, testimonials, or client logos exist for this service line, and none are implied here. What's real:

  • PGCE and secondary teaching background, plus higher education teaching experience
  • Nine years of computational research at the University of Southampton
  • Fourteen people taken through the Anthropic Claude Partner Network Learning Path
  • A public GitHub portfolio of production Claude architectures at github.com/TechIfat
  • Anthropic Partner Network membership

Pricing

Fixed price. No Tier label.

A separate, self-contained product, not a step in the ComplyAI 0–3 engagement ladder.

FormatFee
One day, up to 12 people, remote (base offering)£3,500
One day, up to 12 people, remote, plus 30-day follow-up session£4,500
On-site: as above, plus travel at costas above + travel
13+ people: delivered as two cohorts, up to the 30-person team size the workshop is designed forpriced per cohort

No other pricing variant exists for this product.

Common questions

Straight answers on how this actually works.

Teams of 5–30 people in regulated or data-sensitive sectors (financial services, healthtech, legal, insurance, public sector) who already have Claude Code and aren't yet seeing compounding value from it. Not a session for teams deciding whether to adopt AI tooling; it's for teams that already have, and now need it governed and consistent without slowing down. Each cohort runs with up to 12 people in the room; see the next question for how larger teams split.
A full day, 09:30–17:30, with a coffee break, lunch, and an afternoon break built in. Around six hours 45 minutes of session time, built from your own pre-session diagnostic answers, not a generic deck: mapping where your team actually stands today; the delegation principle that separates what the model reasons about from what deterministic code verifies; writing your own CLAUDE.md live; setting permissions, data classes and secret-handling rules with your security owner in the room; a cost and model-routing policy; the patterns that make month two better than month one; and a capstone where you run one real backlog item through everything covered.
Identical content: the £4,500 option adds a 30-day follow-up to check the CLAUDE.md, permission rules and usage policy against four weeks of real use, and correct course before habits set wrong. On-site is the same again plus travel at cost. Teams above 12 people, up to the 30-person upper bound the workshop is designed for, run as two cohorts, each priced separately at the applicable rate.
You need to already be using Claude Code; this isn't an introduction to AI tooling. Beyond that, no assumed maturity; the day is built from your diagnostic answers, not a fixed curriculum.
Answered straight, not dodged: no enablement client has been through this workshop yet, and no case study or testimonial exists for it. What's real: a teaching background (PGCE, secondary and higher education), nine years of computational research at the University of Southampton, fourteen people taken through the Anthropic Claude Partner Network Learning Path, a public GitHub portfolio of production Claude architectures, and Anthropic Partner Network membership.
Your own CLAUDE.md, written data-class and permission rules, a usage and cost policy, a playbook of patterns proven on your own code that day, a 30-day plan with named owners, and one real backlog task shipped end to end during the session.

Start here

Get your team from access
to compounding value.

Get in touch to check dates and confirm scope for your team.

Get in touch One day. £3,500 up to 12 people, remote. hello@shiftaiconsulting.co.uk