New AI SDLC workshop
From Assisted Coding to Governed Delivery.
A half-day working session for leaders and senior practitioners who need to turn scattered AI-tool use into a shared delivery-system view, one chosen constraint, and a risk-bounded pilot.
The problem
Individual assistance is useful. It is not yet a team delivery model.
When code generation becomes cheaper, queues often move to intent, context, architecture, review, tests, release controls, and operational ownership.
The workshop gives engineering, product, platform, architecture, and security stakeholders one place to examine those constraints together. It starts with a concrete workflow, not a generic debate about AI maturity.
What the group produces
A decision-ready current-state view
- Current and chosen target working-mode map
- Capability heatmap across six organisational dimensions
- Evidence gaps and disagreements made visible
- One constraint blocking the next useful transition
A governed next move
- One bounded pilot and explicit non-goals
- Allowed tools, data, and permissions
- Human verdicts, checks, escalation, and stop conditions
- Owners, measures, and 30-day actions
Four-hour flow
1. Map
Locate selected workflows as Gated, Assisted, Parallel, Supervised autonomy, or AI-native using observable evidence.
2. Diagnose
Assess the capabilities needed to sustain the chosen mode and identify the principal constraint.
3. Design
Create and red-team one governed pilot, then assign owners, evidence, and a review date.
Who should be in the room
The people who own the outcome, workflow, platform, and risk.
Typical participants include the CTO or VP Engineering, engineering and product leaders, platform or DevEx, architecture, security or risk, and senior engineers who understand the current work. The standard group is 6–16 people.
The workshop is suitable when a leadership team needs alignment or a pilot decision. It is not generic prompt training, a vendor selection exercise, or a maturity certification.
Choose the next useful mode, not maximum autonomy.
Start with the trigger, one candidate workflow, and the decision leadership needs. I will tailor the session to the organisation’s products, risks, tools, and available evidence.