Software EngineeringAI-Native SDLC Transformation
Cohort AI-Native Engineering Programme
An engineering team that plans, builds, tests and reviews with AI agents as standard practice, with agreed working rules and guardrails, and a before-and-after measure of delivery performance taken from the client's own tooling.
Cohort AI-Native Engineering Programme
$50K–$120K
- Duration: 8–12 weeks
- Tier: Transform
- Format: Cohort / programme
You leave withA team operating as orchestrators of AI, with measured DORA improvement
Book a Free SessionSound familiar?
We bought AI coding licences for everyone. A few engineers swear by them, most just use autocomplete, and our delivery speed hasn't moved. Code review is now the bottleneck, and nobody's sure whether the AI-written code is making our quality worse.
What you leave with
Baseline report
Format: Document
How we know it’s acceptableSigned off by the engineering lead in week 2
Working agreements
Format: In client repo or wiki
How we know it’s acceptableAdopted by the team and referenced in their review checklist
Tooling configuration
Format: In client repo
How we know it’s acceptableWorks for every participant's setup
Course materials and exercises
Format: Shared drive and repo
How we know it’s acceptableChampions can run a session from them without us
Participant assessments
Format: Document
How we know it’s acceptableEach participant assessed against the agreed approach
Outcome report
Format: Document
How we know it’s acceptableBefore and after figures taken from the client's own tooling
Also included
- Review and testing guardrails for AI-written code
- Internal champions trained to keep the practice going
How it runs
8–12 weeks
Weeks 1 and 2
Baseline
Delivery metrics, team survey, tooling setup, draft working agreements.
OutBaseline report.
Weeks 3 to 6
Core modules, one per stage of delivery
Specifying and planning with AI, agent-led coding, AI-assisted testing, reviewing AI-written code.
OutExercises completed on the real codebase, working agreements version 1.
Weeks 7 to 10
Applied on live sprints with coaching.
OutThe practice used on real work.
Weeks 11 and 12
Final measurement, assessments, champion handover.
OutOutcome report, assessments, champions plan.
The 8-week version shortens the live-sprint phase.
Is this right for you?
This is for you if
- Buyerthe person who owns engineering delivery: a CTO, VP of Engineering or head of engineering.
- Also has to say yessecurity, because code and context go to AI tools; the engineering managers whose teams give up time for it; Finance, for tool licences; sometimes the learning and development budget holder.
This is the wrong answer if
- There is nothing to measure against. No CI/CD, no reliable deployment history. Fix the basics first, or the DORA claim cannot be made.
- No AI coding tools are approved. The policy decision comes first. Start with the SDLC AI Maturity Audit or advisory work.
- The team is in a crunch. If a major release lands in the next eight weeks, sessions will be the first thing cancelled.
- They want a one-day workshop. This is a twelve-week change programme. A short awareness session is a different, smaller thing.
- The real goal is headcount reduction presented as training. Engineers will see through it, and the programme will fail on trust.
What’s included, and what isn’t
In scope
- one cohort (proposed up to 15 engineers)
- sessions
- exercises on the real codebase
- tooling setup
- baseline and final measurement
- champion training
Out of scope
- tool licences
- delivering the client's backlog
- reorganising teams or processes outside engineering
- reporting individual scores to management
- coaching after the programme ends
Before we start
Before day one
- Approved AI tool licences and a security policy covering them.
- Access to CI/CD and repository data.
- A named engineering lead as sponsor.
- Session time blocked in the team's calendars.
- A codebase the exercises can run against.
What we’ll need from your team
- Each participant: four to six hours a week.
- The engineering lead: two hours a week.
- A platform or DevOps engineer: about four hours in total for metrics access.
- Champions: an extra two hours a week in the last four weeks.
Pricing
What we need to give you a quote
- Team size and roles.
- Languages and tools.
- Which AI tools are approved.
- Access to CI/CD and repository data for measurement.
- Calendar constraints and any release freezes.
After we hand over
Who owns it after we leave
The engineering lead owns the working agreements. The champions keep the practice going, and that takes about two hours a week each. Without named champions, the practice will fade within a quarter.
What we measure
- The four DORA measures (deployment frequency, lead time for changes, change failure rate, time to restore)
- Pull-request cycle time
- A developer survey
Before and after measured from the client's own tooling.
What comes next
- Further cohorts for other teams.
- The Claude Code Rollout Programmeif they standardise on Claude Code.
- An SDLC AI Maturity Auditfor other departments.
Questions buyers ask us
Most use autocomplete. We measure where you are in week 1, and the programme is about the rest of the cycle: planning, testing and review.
Fair question. Lead time and deployment frequency can. Change failure rate takes longer to show. We agree in week 1 which measures we'll report, and we report them honestly.
It can, without guardrails. Review and testing rules for AI-written code are part of what the team leaves with.
They don't come off it. The work is done on their real backlog, with about four to six hours a week of sessions.
No. It's about how the team works, and it holds whichever tools you use.