Transformation Performance Index (TPI)
Measure AI transformation with evidence, not opinion.
TPI gives leadership one comparable view of AI transformation across readiness, execution, value and governance. Every score is grounded in evidence, so you can see what changed over time and where attention is needed.
One index. Four pillars of transformation. One comparable view every quarter.
TPI in brief
- What is TPI?
- The Transformation Performance Index is a 0 to 100 score for how an organization's AI transformation is progressing, published by Transformics.
- What does it measure?
- Four pillars: Readiness, Execution, Value and Governance, weighted 30%, 30%, 20% and 20%.
- What evidence does it use?
- Survey answers, documents and read-only telemetry, reconciled so that stronger evidence counts for more and disagreements are flagged.
- Why is it comparable over time?
- The same instrument runs every quarter, and every snapshot records the methodology version it was measured with.
- How does it support decisions?
- It shows leadership what changed, where attention is needed and which evidence stands behind each score, so the next move starts from facts.
Four pillars
Four questions every leadership team has to answer.
Readiness · RDX · 30%
Are we ready?
Strategy, data, skills and culture in place for AI.
Execution · EXI · 30%
Are we executing?
AI systems shipped and adopted in real work.
Value · VLI · 20%
Is AI creating value?
AI landing in business outcomes.
Governance · GVI · 20%
Is it governed?
Oversight, risk and compliance a board can stand behind.
How TPI works
From evidence to one comparable score.
Three kinds of evidence go in. They are compared before anything is scored. What comes out is a score leadership can trace.
1 · Inputs
Evidence from three sources
- Survey: how people across the organization see it
- Documents: strategies, policies, registers
- Telemetry: read-only signals from delivery tools
2 · Reconciliation
Sources compared, not averaged blindly
- Stronger evidence counts for more
- Disagreement between sources is flagged
- Coverage and uncertainty are recorded
3 · TPI
One index, four pillars
- A score from 0 to 100
- Readiness · Execution · Value · Governance
- Rolled up with published weights
4 · Output
What leadership sees
- The TPI and each pillar
- The evidence behind every score
- What changed and where attention is needed
Evidence model
Not a questionnaire. An evidence model.
Surveys say how people see the transformation. Documents and read-only telemetry show what is actually in place. TPI reconciles all three, so a confident self-assessment cannot carry a score that the evidence does not support.
- Stronger evidence counts for more. Telemetry outweighs documents, and documents outweigh survey answers.
- Disagreement is shown, not hidden. A gap of 1.5 points or more between sources is flagged for leadership.
- High-impact claims are checked. Claims read from documents are verified against the source text, and the strongest wait for a person's approval before they count.
Published weights, adjusted by the confidence of each evidence item.
- Telemetry1.0
Read-only signals from delivery tools
- Documents0.8
Policies, registers, plans, reports
- Survey0.5
Perception across roles and teams
Governance · human oversight, scored 0 to 5
- Survey says
- 4.1
- Documents show
- 2.4
- Gap
- 1.7
Flagged: a gap of 1.5 or more is shown, not averaged away.
What leadership gets
A score, and the story behind it.
Each quarter, TPI feeds an executive output: where the organization stands, what changed, which evidence supports it, where attention is needed and what to do next.
How Executive Intelligence worksExecutive output · Q4
Transformation Performance Index
TPI
64+5
- Readiness70 +4
- Execution65 +7
- Value61 +4
- Governance56 +3
- What changed
- TPI +5 since Q3 on the same instrument. Execution moved most (+7); two AI systems reached production.
- Evidence
- Survey, documents and telemetry agree on Execution. On Governance they disagree, and the gap is shown.SurveyDocumentsTelemetry
- Where attention is neededNeeds attention
- Governance: oversight evidence is thin for two production AI systems.
- What to do next
- A suggested initiative, awaiting an owner's approval: document human oversight before scaling further.
An observed change on a comparable instrument, never a claim of cause.
Beyond the score
A gap in the TPI becomes the next move.
TPI is where Transformics starts. A gap becomes a priority, an owned initiative, a tracked change with evidence, and a prepared next move. Change is observed and recorded, never claimed as cause.
TPI
A gap is measured
Governance trails the other pillars.
Executive Intelligence
It becomes a priority
Named, explained and ranked for leadership.
Steer
An owned initiative
An owner, a due date and a status.
Prove
A tracked change, with evidence
Next quarter, on the same instrument.
Orchestrate
The next move
Prepared from what was learned, approved by a person.
Executive Intelligence How Orchestrate prepares the next move
Change over time
Progress you can prove, quarter after quarter.
The same measurement logic is maintained over time, so change remains interpretable. Each snapshot is stored unchanged with the methodology version it used, and a comparison across major methodology versions is flagged rather than blended.
That is how Transformics proves progress: not with a one-off report, but with a comparable record of what moved, when, and on what evidence.
How longitudinal comparison works- 1Q2: Execution begins a sustained rise that continues through Q4.
- 2Q4: Governance is still the lowest pillar, and sources disagree on oversight evidence.
Pillar movement, Q1 to Q4
- Readiness60 → 70 (+10)
- Execution48 → 65 (+17)
- Value52 → 61 (+9)
- Governance46 → 56 (+10)
Why trust the score
A number you can trace back to its evidence.
Comparable
The same instrument every quarter, with the methodology version pinned to each snapshot.
Evidence-grounded
Survey answers are weighed against documents and telemetry, and disagreements are flagged.
Transparent
Every score traces to its sources and weights, with coverage and uncertainty beside the number.
Audit-ready
Snapshots are stored unchanged, and every AI-assisted step is logged.
TPI measures the organization's transformation, not individual people, and is not intended for individual performance evaluation.
Frequently asked questions about TPI
How is the TPI calculated?
Each pillar is scored from 0 to 100 from the evidence beneath it. The TPI is the weighted sum of the four pillar scores: Readiness 30%, Execution 30%, Value 20% and Governance 20%. A pillar is only scored when enough evidence sits beneath it; without that coverage it is shown as not yet scored, never estimated. The scoring math is deterministic: no AI model sets or adjusts the score.
How does evidence triangulation work?
Survey answers, documents and telemetry are reconciled before anything is rolled up. Each source class has a published weight (survey 0.5, document 0.8, telemetry 1.0), adjusted by the confidence of each evidence item, and the published score is the weighted average. When sources disagree by 1.5 points or more on the 0 to 5 scale, the disagreement is flagged and shown rather than averaged away.
Is TPI just a survey?
No. A survey captures how people across the organization see the transformation, and it is weighted below documents and telemetry. Where AI helps read documents, extracted claims are checked against the source text, and high-impact claims wait for a person's approval before they count.
Can I compare TPI from one quarter to the next?
Yes, that is what it is built for. The same instrument runs every quarter, every snapshot is stored unchanged with the methodology version it was measured with, and a comparison across different major methodology versions is flagged instead of presented as like for like.
Is TPI used to evaluate individual employees?
No. TPI measures the organization's transformation, not people. Telemetry is read-only and used at team level, and TPI reports are not intended for individual performance evaluation.
How do I know the score is reliable?
Every score shows its coverage and an uncertainty range, and traces back to the evidence and weights behind it. The Validity Center reports the statistical quality of the instrument itself as responses accumulate, and says so when a sample is too small to judge.
How do we get a first TPI?
Start a 30-day Paid Pilot to run the assessment with your own teams and evidence and get a first TPI. For a quick directional read first, the free 5-minute transformation check gives a self-reported indication; it is not a TPI.
Every formula, weight and version-history entry is in the methodology overview.
Get your first TPI
Start with a 30-day Paid Pilot.
Full Business functionality with pilot-scale usage limits.
$499one-time
Credited 1:1 toward your first annual subscription if you upgrade within 30 days of pilot expiry.