← Back to the Brail website

What is an approved claims library?

Part of: BuildLast updated 17 November 2026

An approved claims library is the set of statistics, customer quotes, results and statements an agent may use, each with its source, its owner and any rules about when and how it applies.

Also called: claims library, proof point library, approved proof points.

Why it matters.

Claims are where AI output does the most damage. A figure from an old deck, a customer quote used without permission or a result rounded up in a draft can end up in front of customers, regulators or the press. Agents fill gaps with plausible numbers when they have nothing approved to use.

A claims library gives agents something approved to reach for. Each claim carries its source, who owns it, whether it can be used publicly, which audiences it applies to and how it must be worded, such as 'customers report' rather than a bare figure. Work that uses a claim outside the library can be flagged before it ships.

Example.

BeforeA sales deck generated by AI quotes '3x faster onboarding', a figure nobody can trace. Legal finds it a week after the deck has gone to twenty prospects.
AfterThe agent cites only claims from the library, each with its attribution wording. A draft that introduces an unsourced number is flagged for a person before it goes out.

How Brail handles it.

Brail gives agents your approved proof points with the wording each one requires, and flags work that uses a number or quote you haven't approved. Your team sees the unsupported claim before a customer does.

Questions people ask.

What belongs in a claims library?
Statistics, customer results and quotes, awards and certifications, comparisons and any statement that needs evidence. Each needs its source, its owner and any rules about how and where it can be used.
How do we stop AI from making up numbers?
Give it approved numbers to use, tell it to leave a gap rather than invent one, and check every draft for figures that aren't in the library. See grounded output.
Who should own the claims library?
Usually product marketing, with legal or compliance approving anything regulated. See governed context.
  • Grounded output: Grounded output is AI-generated work in which every factual claim, figure and quote traces back to an approved source, rather than to whatever the model guessed or remembered.
  • Grounding rate: Grounding rate is the share of AI-generated work in which every claim, figure and quote traces back to an approved source, measured before the work ships.
  • Brand compliance: Brand compliance is checking output against a brand's rules (visual, verbal, audience, channel and asset) before it's published, and flagging or fixing what doesn't match.
  • Governed context: Governed context is context for AI that has an owner, an approval step and a version history, so every agent and tool can rely on it and reuse it.
  • Marketing ontology: A marketing ontology is a shared model of the things marketing talks about (products, audiences, segments, messages, proof points, channels) and how they relate, so every agent reads them the same way.