Methodology
The Transformation Performance Index
TPI is a 0–100 composite score that captures how an organization is performing across the four things a board has to answer for. Same instrument every quarter, comparable across quarters. Triangulated across survey responses, customer-uploaded documents, and read-only telemetry, so a single source can't carry the number alone.
The four board pillars
“Are we set up to do AI well?”
“Are we shipping?”
“Is it working?”
“Are we in control?”
How the score is structured
One number for the board, built from a clear hierarchy you can trace down to the evidence. The TPI rolls up through 4 board pillars, 16 dimensions, and 48 evidence-anchored questions in the Core Assessment, each dimension weighted, each question answered from evidence rather than opinion alone.
Readiness, Execution, Value, Governance: the four questions a board answers for.
Each pillar breaks into weighted dimensions, so a score is never a black box.
Every dimension is answered by anchored questions, then triangulated across sources.
Beyond the Core Assessment, 26 optional diagnostic modules let a team go deeper on a specific topic, from data infrastructure to EU AI Act readiness, on demand.
What's in the full methodology
Customers (Paid Pilot and above) get the full methodology page inside the platform: per-pillar weight rationale, the source-class triangulation table (telemetry 1.0 · document 0.8 · survey 0.5 · …), the per-module dimension breakdowns, the evidence-value tier matrix, the cross-module rule library, the validity scoring framework, and the methodology version history.
- Triangulation formulas, how multiple sources reconcile to a single dimension score.
- Pillar weight rationale, why Readiness and Execution carry 30% and Value/Governance 20%.
- Cross-module rule library, 79 pattern rules across families A–K.
- Validity Center, five textbook validity dimensions per pillar, with honest insufficient-data states.
- Sector profiles, banking, public-sector, mid-market, early-stage weight variants.
- Hallucination guards, the four-layer narrator guard, audit trail, and EU LLM inference routing (AWS Bedrock, source eu-west-1).
See your own TPI in 30 days
Paid Pilot, $499. Full Business functionality for 30 days, with pilot-scale usage limits. Credited 1:1 toward your first annual subscription if you upgrade within 30 days of pilot expiry.
Run a Paid Pilot →Already a customer? View the full methodology.
Pattern-rule findings and validity scores are diagnostic signals, not legal verdicts and not employment-decision inputs. Investigate before acting.