Fortbourne Request a call

Claims register

Every external claim, and what it does not establish

One row per external factual claim rendered anywhere on this site. If we assert something about the outside world, it is here with its source and its limits. A claim with no row is a defect -- tell us and we will fix it or withdraw the claim.

What this register covers. Every external factual claim rendered anywhere on this site, and nothing else. A client result, case study, testimonial, certification or partnership reaches a page only after its row is in this register.

How rows get approved. A row is approved only when the reviewer of record has re-read the source and signed it on a stated date. Until then the row states what is outstanding.

6 claims · 4 approved · 2 pending review

  1. 01

    Of twenty-five organisational attributes studied, redesigning workflows showed the strongest association with respondents' reported EBIT impact from generative AI.
    Where it appears
    / (homepage evidence card), /research/why-ai-cant-fix-a-process-nobody-owns/ and /ai-implementation/
    Source
    McKinsey -- The state of AI: How organizations are rewiring to capture value (March 12, 2025)
    Verified
    2026-09-10
    Limitations
    Observational and self-reported. It is a survey of respondents' own perceptions of EBIT impact, not a measured financial outcome. An association is not a cause. The respondent base is not the $15-40M US mid-market this firm sells to. It supports the sequencing argument; it does not prove it. The firm has no affiliation with McKinsey.
    Status
    pending review
    Reviewer
    Unsigned -- awaiting direct re-reading of the source.
    Why it is not approved
    The publisher's site returned no response when re-verification was attempted, so the source has not been re-read and the row stays unsigned.

  2. 02

    EOS gives a company a meeting rhythm and a scorecard.
    Where it appears
    /approach/ and /research/the-operating-baseline/
    Source
    EOS Worldwide -- The EOS Model
    Verified
    2026-09-11
    Limitations
    A plain description of a published methodology, used as competitive context. The firm does not copy its branded material, use its terminology as its own, or imply implementer status or any affiliation. Both pages position the firm as working underneath EOS rather than replacing it.
    Status
    approved
    Reviewer of record
    Fortbourne editorial review
    Signed
    2026-09-11

  3. 03

    A discounted or performance-linked engagement is a material connection that must be disclosed in a published endorsement or case study.
    Where it appears
    /firm/
    Source
    FTC -- Endorsement Guides: What People Are Asking
    Verified
    2026-09-11
    Limitations
    Paraphrased conservatively; the FTC guidance is broader than this sentence and is not reproduced. It is guidance, not legal advice and not a determination about any specific engagement. Applying it to a real case study requires legal review at the time of publication. The page commits the firm to the disclosure rather than interpreting the rule for anyone else.
    Status
    approved
    Reviewer of record
    Fortbourne editorial review
    Signed
    2026-09-11

  4. 04

    KPI tracking was also associated with reported AI value in the same survey.
    Where it appears
    /research/the-operating-baseline/
    Source
    McKinsey -- The state of AI: How organizations are rewiring to capture value (March 12, 2025)
    Verified
    2026-09-10
    Limitations
    Same as the workflow-redesign row: observational, self-reported, correlational, and not a study of the US mid-market. The guide states this inline and explicitly calls it consistent with the argument rather than proof of it.
    Status
    pending review
    Reviewer
    Unsigned -- awaiting direct re-reading of the source.
    Why it is not approved
    The publisher's site returned no response when re-verification was attempted, so the source has not been re-read and the row stays unsigned.

  5. 05

    Case identifier, activity and timestamp are the core requirements for event data usable in process analysis.
    Where it appears
    /research/the-operating-baseline/, /research/why-ai-cant-fix-a-process-nobody-owns/ (source list) and /operational-foundations/
    Source
    Microsoft Learn -- Process mining: prepare processes and data
    Verified
    2026-09-11
    Limitations
    This is vendor product documentation describing the input requirements of one product. It is the structural basis for the 'system events before screen recording' position; it is not an independent standard, and it does not establish that any particular implementation is adequate. Described as the standard requirement for process mining tooling, not as a formal standard.
    Status
    approved
    Reviewer of record
    Fortbourne editorial review
    Signed
    2026-09-11

  6. 06

    Referencing a public AI risk-management framework does not constitute certification by its author.
    Where it appears
    Site-wide -- the site footer carries this line on every page.
    Source
    NIST -- AI Risk Management Framework
    Verified
    2026-09-11
    Limitations
    The framework is voluntary guidance. NIST does not certify against it. This row exists to pin the negative claim: the firm is not NIST certified and does not say it is. It must never be softened into an implied endorsement, partnership or certification.
    Status
    approved
    Reviewer of record
    Fortbourne editorial review
    Signed
    2026-09-11

Rules this register is kept under