OTTRA METHOD

Method for
Humans and AI.Five stages that keep AI‑assisted engineering aligned.

Bring people, AI and the engineering platform into one disciplined way of working. Clear intent, controlled access and accountable decisions, from the first brief to the next improvement.

Human-led. AI-assisted.

A human in a pink #Human T-shirt and a transparent assistant in an orange #AI T-shirt review an OTTRA-marked design at a workbench. — Illustrative human and AI collaboration Pause hologram

Illustrative human and AI collaboration Pause hologram

Meet a practitioner
Neil Douek
Socio Tech Lead
Human intent, AI alignment and engineering judgement.
Meet Neil

Connect what you want AI to do with the boundaries, evidence and human decisions that make the work useful.

Talk through your engineering approach

Five stages.
Human accountability.

Every engineering task makes a promise to someone. These five stages keep that promise connected to the people it serves, the controls it needs and the evidence behind each decision.

Define the promise.

Start with the people who need the result. Agree what must improve, what must stay protected and how you will know the work has met its purpose.

Humans: Set the purpose, constraints and acceptance criteria.
AI: Helps explore options and surface unanswered questions.

Purpose, people and acceptance criteria come together in one agreed brief. — The first control is clarity.

What your team getsA shared brief the team can test.

Set the boundaries.

Give people and AI only the access and authority the task needs. Agree what can be done, what requires approval and when to stop and ask.

Humans: Agree permissions, data limits and escalation.
AI: Works within the authorised tools, data and task scope.

People and AI work inside an agreed boundary, with access to tools and data controlled. — Access follows the task.

What your team getsClear limits before anyone acts.

Build the evidence.

Keep tests, security findings and operational signals connected to the work. Show what was checked, which revision was checked and what still needs attention.

Humans: Choose the checks and challenge gaps in the results.
AI: Helps prepare tests, investigate findings and explain its work.

Tests, security checks and operational signals connect to a reviewable body of evidence. — Make confidence inspectable.

What your team getsA reviewable record of what was checked.

Make the decision.

A passing check is one input. A person judges whether the work meets the intent, whether the remaining risk is acceptable and how to stop or recover if needed.

Humans: Approve, reject or request more work. Own release and recovery.
AI: Presents its contribution, assumptions and uncertainties for review.

Evidence reaches a human reviewer, who decides whether to accept the work or ask for changes. — Judgement stays with people.

What your team getsA named owner and a recorded decision.

Close the loop.

Follow the result into use. Look at user feedback, failures, rework and effort, then improve the next brief, the workflow or the controls.

Humans: Decide what needs to change and what to try next.
AI: Helps find patterns and suggest improvements for review.

Operational feedback returns through a human review to improve the next brief. — Delivery starts the next conversation.

OTTRA alongside your team

Put the method
to work.

We help build the platform, controls and skills that make the Method part of everyday engineering.

Engineer the environment. Platforms, integrations and permissions shaped around your work.
Operate it well. Clear responsibilities, security checks and practical support.
Build capability. GitLab and Claude enablement, training and engineering.

Talk about your team