The TAATIQ Method
Four outcomes
Two linked questions: should we proceed, and if so, how should technology participate? Not AI and Fix First answer the first; Assist and Automate answer the second. Stops are shown with the same prominence as go outcomes, and the method's credibility rests on Not AI being as useful as Automate.
Not AI
Do not pursue AI for this problem. A process, people, policy, existing software or simpler intervention may be the better path. Never means "no technology."
Fix First
Do not test the intervention yet. The foundation is not ready. Carries a reason: Evidence (count it), Process (write it down) or Data (start the records). Re-score in 60 to 90 days.
Assist
Proceed with human control. Technology materially increases what people can accomplish while a person retains the relevant decision or control. The people affected are told.
Automate
Proceed with defined autonomous execution. Technology performs defined workflow steps or the full workflow within agreed controls, with defined exception handling and human escalation where required.
Trust disposition
What the Trust reading permits, so a sensitive, low-reversibility case can never fall through to a live pilot. The method does not block experimentation; it controls deployment.
Manageable
May proceed, subject to the normal controls in the charter.
Requires review
May proceed to a controlled test or prototype. Live implementation requires the specified Trust conditions first.
Not yet established
No live implementation until the required controls have been demonstrated. Prototyping to design them is encouraged.
The AI Fit test, at Align
AI is considered where all five hold. If not, the method does not force an AI recommendation. A test, not a score.
- The inputs are variable, unstructured or language-based.
- A deterministic, rules-based solution cannot adequately solve the problem.
- The required data or context exists or can reasonably be created.
- Errors can be detected, bounded or appropriately controlled.
- AI creates materially more value than the credible non-AI alternatives.
Adoption posture, how comfortable you are with technology handling the task, shapes the change plan. It never decides whether automation is appropriate and never overrides Trust.