Reskilling

Your team doesn't need replacing. It needs re-skilling.

The engineers who know your systems, your customers and your history are the hardest thing to hire. Agentic tooling changes how they work — not whether you need them. We take the team you already have and move it up the maturity curve, in the room, on your codebase.

The maturity model

Where the work actually moves

Where most teams sit, and where the intensive takes them.

LevelYour roleWhat the AI does
1 · AssistedWrite code with suggestionsAutocomplete, chat
2 · AutomatedReview and approveWrites, tests, commits
3 · OrchestratedCoordinate agentsMultiple agents, in parallel
4 · AutonomousMorning review onlyWorks overnight, merges
5 · EvolvingProvide visionSelf-improves from failures

Most teams plateau at Level 1. The intensive gets a team to Level 3–4.

What is it worth for your team?

Your numbers, Anthropic's data

Pick the role, put your own numbers in, and move the sliders until it reflects your world. The speed-up figures come from Anthropic's Economic Index — their data, not our claim.

— estimated value per year, this team
Anthropic's observed speed-up
—
Hours back, per person per week
—
Cost of a 1-day intensive
—
Pays for itself in
—
Cumulative net position · 52 weeks —
— estimated value per year, this team
Hours back, per person per week
—
Cohorts needed (max 30 each)
—
Cost, 1 day per cohort
—
Pays for itself in
—

Where the speed-up numbers come from

The Anthropic Economic Index is a public dataset measuring how AI is actually being used across the economy. Anthropic samples real Claude conversations, maps them to the tasks that define each occupation in the US Department of Labor's O*NET / SOC classification, and estimates how long each task takes with and without AI assistance.

  • All 717 occupations the release publishes usable times for are searchable above — nothing is filtered to flatter us
  • Release 2026-06-26, covering May 2026
  • Published per occupation: estimated task time unaided, task time with AI, and the split between automation and augmentation
  • Figures are copied from the release, not re-rounded; where it published nothing we show nothing rather than a zero
  • Independent of us — we did not commission it and cannot tune it

How to read this. Two honest caveats, because a calculator that flatters itself is worth nothing. Anthropic's times are model estimates, not stopwatch measurements of your team. And they describe the tasks people already choose to bring to an AI — which are, by definition, the tasks it is good at. A speed-up on those tasks is not a speed-up on the whole job, which is why you set the share of the week yourself, and why dragging the speed-up below Anthropic's figure is a reasonable thing to do. Treat the output as an order of magnitude for a conversation, not a forecast.

What the team leaves with

01

Not slides. The intensive is built around artifacts that exist by the end of the day and keep working afterwards.

  • A testing pipeline the team trusts enough to let agents commit against
  • Pre-commit hooks and semantic security review wired into CI
  • A spec-driven workflow, with tests generated from requirements
  • An orchestration pattern for several agents working in parallel
  • A memory system so the agents improve rather than repeat themselves

Does it pay for itself?

02

We won't quote you a productivity percentage — nobody credible has one. Here is the threshold instead, so you can apply your own judgement.

  • A 1-day intensive for a team of ten breaks even at about fifteen hours saved per engineer — once
  • Across a year that is roughly eighteen minutes a week, per person
  • Everything beyond that is return

Break-even shown at an assumed fully-loaded cost of $100 per engineer-hour. Substitute your own figure — the arithmetic is yours to check, which is the point.

Shapes it comes in

03

The same material, sized to what you can take out of the calendar.

  • Enterprise Intensive — three hours, executive-friendly, straight to production patterns
  • 1-Day Intensive — the full workshop, hands-on, one team
  • 3-Day Track — depth, with time to apply it to your own repositories

Move the team you have.

Tell us the team, the codebase and the dates. We'll come back with a quote and a plan for the room.

Source: MIT Media Lab, State of AI in Business 2025 — reported by Fortune, Aug 2025