From intelligence to implementation

1. 1. Institutional question definition

   Agreeing the decision actually being made, and who makes it. Most engagements fail here, by starting from a technology rather than a decision.

2. 2. Evidence and systems audit

   What the institution already holds, what it can measure, and what it cannot. Including the data it believes it has and does not.

3. 3. Country or sector decision workshop

   Working through the evidence with the people who will own the outcome, using this instrument as the shared record.

4. 4. Responsible AI opportunity assessment

   Where a system would help, where it would not, and where it would create harm or exclusion that outweighs the gain.

5. 5. Pilot design

   A bounded test with a stated hypothesis, a defined population, disaggregated measurement, and a pre-agreed definition of failure.

6. 6. Implementation

   Building and running it, with the delivery states this platform uses applied to the work itself.

7. 7. Measurement and review

   Against the pre-agreed measures, disaggregated by who was reached and who was not.

8. 8. Scale, correction or withdrawal

   All three are real outcomes. An engagement that ends in withdrawal because the evidence did not support continuing is a success of the method, not a failure of it.

What Orvantis Intelligence is

An AI transformation and implementation partner. It built this platform and this instrument. It is directed by Hannah Kwakye, AI Transformation Consultant.

No client, funder, partner, endorsement, testimonial, project result or usage figure is claimed anywhere on this platform, because none has been evidenced here. The absence is deliberate: an instrument that refuses unsourced claims about countries cannot make unsourced claims about itself.

Expression of interest

If an institution wants to discuss a question this instrument raises, the honest first step is the first stage above — defining the decision. Enquiries can be directed through Hannah Kwakye.