Ground Truth License

The License is for teams that already know they want the Match Rule Truth Engine year-round and are comfortable running the scoring themselves. It suits an SI practice that wants the same generation engine standardized across every project without opening a new engagement each time, or a governance team that wants to regenerate a domain's ground-truth dataset the moment a rule changes, on its own schedule. It's billed annually, and nobody has to get on a call to buy it.

This page covers what the License includes and how you buy it. It does not cover Match Rule Validation scope, which lives on that offer's own page.

You get a year of Match Rule Truth Engine access: generate a domain-shaped Labeled Ground-Truth Dataset with its Known-Duplicate Pair File whenever you need one. Regenerate a fresh truth set the moment your data model changes, or the moment a new project needs its own domain-shaped duplicates. It's the same underlying engine that builds the truth set inside Match Rule Validation, sold here on its own, without the engagement wrapped around it.

The License hands you the generation half of the validation loop and leaves the scoring half with your team. The loop runs like this. Generate a domain-shaped Labeled Ground-Truth Dataset with its Known-Duplicate Pair File. Run your current match rules against that dataset in whatever platform they live in. Compare your rule output to the answer key and bucket every candidate pair into true positives, false positives, and false negatives. From those three counts your own team computes precision, recall, and F1 for the rule set as it stands today, then regenerates and repeats after the next tuning change.

That loop is the same one Match Rule Validation runs, with one difference that matters for scoping: in Match Rule Validation the scoring, the ranked failure list, and the report are delivered work, and under the License they are your analysts' work.

What the License doesn't include is the doing. Nobody scores your match rules against the truth set for you, ranks the failures, runs a scorecard call, or writes a Gate-Review Certainty Report. You get the generation engine; the analysis is on your team. If what you actually want is a scored, done-with-you outcome, that's Match Rule Validation, not this.