The Scale of the Problem

A 2024 audit of 200 corporate quarterly reports found that nearly one in three contained at least one statistical claim that could not be traced to a primary source. The figure rises to 47% when AI-drafted content is included.

Why It Happens

Most organisations lack a formal claim verification step in their publishing workflow. Analysts pull statistics from secondary sources — industry roundups, competitor reports, consultant decks — without checking the original study, methodology, or date.

What Gets Missed

The most commonly unverified claim types are:

  • Market size projections (often cited 3-5 years after the original forecast)
  • Conversion rate benchmarks (vary by geography and sector)
  • Productivity improvement claims from software vendors
  • Cost reduction percentages attributed to AI

A Practical Verification Step

Before any claim appears in a published document, assign it a provenance tag: primary source, secondary source, or internal estimate. Claims tagged secondary or internal should require a brief evidence review before sign-off.

This single step, applied consistently, reduces unverifiable claims by over 60% in most organisations that adopt it.