The five commitments
What every consultant agrees to.
Each commitment is enforceable. Each one is the floor — not a ceiling — for how a Verified consultant operates.
01
Honesty
Represent capabilities, experience, and timelines truthfully.
A Verified consultant will not exaggerate qualifications, claim experience they do not have, or promise outcomes they cannot deliver. Proposals, portfolios, references, and credentials must be accurate at the time of submission and kept current.
In practice: if a consultant has worked on three projects in an area, they will not present themselves as having done ten. If they cannot meet a timeline, they will say so before accepting the work.
02
Confidentiality
Protect client data and business information per signed NDA.
Client information — data, documents, business strategy, code, model details — is treated as confidential. A Verified consultant will not share, resell, or reuse client information beyond the scope of the engagement, and will protect it using reasonable technical and organisational safeguards.
In practice: data shared for one client’s engagement is not used to train models for, or inform work done for, a different client. NDAs are honoured both during and after the engagement.
03
Competence
Decline work outside demonstrated expertise; refer when appropriate.
A Verified consultant will not take on work they are not qualified to do well. Where a client’s need is outside the consultant’s competence, the consultant will say so and, where possible, refer the client to a more suitable Verified consultant on the marketplace.
In practice: a generalist AI consultant will not silently take on a regulated-industry audit they are unequipped to perform. A referral is preferred to a stretched engagement.
04
Client Interest
Recommend the right solution even when it costs the consultant a sale.
A Verified consultant’s recommendation must reflect what is best for the client, not what is most profitable for the consultant. This includes recommending no engagement at all when that is the honest answer, recommending a smaller scope than the client asked for, or recommending a different provider.
In practice: if a client’s real need is a free open-source tool rather than a custom build, that is what gets recommended — even if it means the consultant walks away without revenue.
05
Ethical AI
Refuse to build systems intended to deceive, manipulate, or violate human rights.
A Verified consultant will not knowingly build, advise on, or operate AI systems whose purpose is to deceive users, exploit cognitive vulnerabilities, manipulate vulnerable populations, or violate human rights as recognised under international law.
In practice: covers AI for unlawful surveillance, deceptive impersonation at scale, systems designed to defraud users, and tools whose primary use is to circumvent rights protections. Lawful, transparent use cases — including ones the consultant personally disagrees with — are not in scope.