Glossary · The Truth Score
Truth Score™
The Truth Score™ is DataTruth’s score out of 100 for how far a company’s data can be trusted, built from six weighted pillars and placed in one of four bands, A to D.
59/ 100
Band CSignificant work needed
Weighted 75 · capped by Connectivity 44
- Accuracy22%88
- Completeness18%97
- Governance18%69
- Consistency16%88
- Connectivity14%44
- Lineage · limited signal12%50
- A 80+ · AI ready
- B 60+ · Minor issues
- C 40+ · Significant work needed
- D Below 40 · Foundation needed
In plain terms
The Truth Score™ answers one question for a non-technical reader: can this company’s data be trusted enough to build on? It is a number from 0 to 100, calculated from the exports of the systems behind a decision, such as the CRM, billing, ERP and finance, HR and payroll, and the spreadsheets that sit between them.
The score is built from six pillars, each weighted by how much it matters when AI starts acting on the data: Accuracy 22%, Completeness 18%, Governance 18%, Consistency 16%, Connectivity 14% and Lineage 12%.
The total places the company in a band: A (80 and above) AI ready, B (60 and above) minor issues, C (40 and above) significant work needed, and D (below 40) foundation needed.
Three pillars are critical: Accuracy, Completeness and Connectivity. A weighted average can hide a fatal gap, so if the weakest of the three is below 80, the overall score is capped at 0.55 × that pillar + 35, rounded. In our sample memo (real engine output on generated test data, not a client), the pillars average 75, but Connectivity is 44, so the cap is 59 and the Truth Score is 59: Band C, significant work needed.
It is rule-based, not AI-generated. The same data, scored on the same date, gives the same score, and every pillar shows what was measured and what wasn’t. Where a pillar has too little signal in the exports supplied, it is shown as limited signal or not assessed instead of being given a comfortable default. A score that flatters the data is worse than no score at all.
On its own, a score says how far the data can be trusted. It doesn’t say which AI ideas the data can run. That is the job of the Leverability read, which sits alongside the Truth Score in every Truth Score Read and AI Investment Review.
For the longer argument, read how the Truth Score measures AI readiness.
- See it in a real memoThe sample Truth Score ReadReal engine output on generated test data, not a client
- How we measure itThe method, including the capPillars, weights, bands and what is never guessed
- What happens to your exportsSecurity and data handlingExports only, NDA and data processing agreement first
Questions
Truth Score™, answered.
The questions people ask about Truth Score™, answered in plain English.
Book the callWhat is a good Truth Score?
Band A (80 and above) means the data is ready for AI to act on. Band B (60 and above) means minor issues to fix first. Band C (40 and above) means significant work is needed, and band D (below 40) means the foundation has to be built before AI is worth the money.
Is the Truth Score generated by AI?
No. It is rule-based: the same data, scored on the same date, gives the same score. Every red flag behind it is checked by hand, with sample records verified, before the memo is signed.
Can the score be compared between companies?
Same pillars, same rules, same caps, so a score is built the same way at each company. Where a pillar is not assessed, the memo says so, and the scope line says which systems were in each score, so like is compared with like. That is how an investor can read a portfolio: one company first, then every company on one method.
Why is the score lower than the pillar average?
Because of the cap. If Accuracy, Completeness or Connectivity is below 80, the score is capped at 0.55 × the weakest of them + 35. Strong pillars can’t make up for systems that disagree about who your customers are. The full rule is on our approach page.
Does a high score mean our AI project will work?
No. A high score means the data can be trusted. Whether a specific AI project can run on it is a separate question, answered by the Leverability read, which marks each AI idea deploy now, needs work or blocked.
Related terms
Keep reading.
- The Truth Score Leverability read The Leverability read says which AI ideas your data can use now, which need fixes first, and which it can’t support at all, with the reason in your own records. Read the definition
- AI readiness Data quality Data quality is how far data is fit for the job it is about to do: correct, filled in, consistent, owned, connected across systems and traceable to its source. Read the definition
- Method Cross-system reconciliation Cross-system reconciliation is matching records between systems, such as CRM and billing, or HR and the CRM, to find where they disagree. Every system can look clean on its own; the truth is in the gaps between them. Read the definition
- AI readiness AI readiness AI readiness is whether a company’s data can support the specific AI it is about to pay for: whether the data can be trusted, and whether it can run that use case. Read the definition
Definitions are the easy part. Your own number is the useful one.
Exports from two or more systems, matched between them, every red flag checked by hand. A signed memo with your Truth Score and a Leverability read, within 5 working days of complete exports.
Starts with a 30-minute call. Commits you to nothing.