What a Chief Workforce Officer must be able to do
The competency framework, the code of ethics, and how the Institute will recognise both. Published so that organisations can use them before anyone is certified against them.
Six domains. Eighteen competencies. Three levels of scope.
Each competency is written as a result the holder achieves, with the behaviour an assessor expects at each level. The unit of evidence is a decision, not an hour.
Workforce composition and strategy
The CWO decides, rather than discovers, what share of the organisation’s work is done by people, by agents, and by the two together.
A1Maps the workforce as one system: human roles, agents in production, and the tasks that pass between them
Produces an accurate map for one function, naming each agent’s owner and the human tasks it changed
Maintains the organisation-wide map as a governed record, reconciled with HR and IT systems, and uses it in planning
Has published a mapping method or taxonomy adopted beyond their own organisation
A2Sets a deliberate target composition with scenarios and reversibility
Can state the current mix for their area and the assumptions behind the planned mix
Presents composition scenarios to the executive team with costs, verification load, and the conditions under which each would be reversed
Advises boards on composition strategy and can show outcomes of past decisions against what was forecast
A3Aligns composition with strategy, risk appetite and the organisation’s capacity to check its own work
Identifies where a proposed deployment exceeds the area’s capacity to supervise it
Owns the composition decision at executive level alongside the CFO and CIO, and can defend it to the board in those terms
Shapes how regulators, investors or sector bodies assess workforce composition
Allocation of work by verification cost
Work goes to an agent only when the cost of checking its output is known, budgeted, and lower than the value produced.
B1Estimates verification cost before deployment, including the expertise needed to check and its scarcity
Applies a verification-cost estimate to a proposed deployment in their area and records it
Requires a verification estimate in every business case for an agent and has rejected or re-scoped deployments on that basis
Has published or validated a verification-cost method used by other organisations
B2Budgets verification capacity as an explicit resource, in hours and in people
Tracks checking hours for agents in their area and reports when they exceed plan
Holds a verification budget at organisation level and reallocates it as agents are added or retired
Advises on verification capacity as a strategic constraint at sector or policy level
B3Declines, defers or reverses deployments where no one competent can check the work
Escalates a deployment that lacks a competent checker, with evidence
Has personally declined or reversed at least one deployment on verification grounds and documented the reasoning
Sets the standard others cite for when an agent should not be deployed
Agent lifecycle and ownership
Every agent has a named human owner with the expertise to check it and the authority to stop it, and moves through a governed lifecycle in a vocabulary that does not pretend it is a person.
C1Ensures each agent in production has one accountable owner who can check its work and stop it
Can name the owner, purpose and review date of every agent in their area, and the gaps
Maintains an organisation-wide register of agents and owners; no agent runs without one; ownership is re-assigned when people move
Has defined ownership standards adopted by other organisations or by a vendor’s governance product
C2Governs selection, evaluation, update and retirement, with criteria set before deployment
Evaluates an agent against pre-set criteria and recommends continue, update or retire
Runs the lifecycle for the whole portfolio, with periodic evaluation and a record of retirements and why
Contributes evaluation methods or retirement criteria to the public body of knowledge
C3Sets autonomy and oversight thresholds and permissions jointly with technology and security leaders
Documents the autonomy level and oversight arrangement for agents in their area
Co-owns the autonomy and permissions policy with the CIO and CISO and has arbitrated a disagreement about it
Advises on autonomy standards at regulatory or industry level
Human role design and protection
Roles are redesigned around agents so that accountability stays with people who are able to carry it, and the entry layer that produces the organisation’s future experts is protected.
D1Redesigns roles so that the human who signs off has the time, expertise and authority to do so
Rewrites at least one role description around an agent with the checking responsibility and time made explicit
Has redesigned a family of roles across the organisation and measured role clarity and accountability after the change
Has published role-design patterns for hybrid work that others use
D2Protects the expertise pipeline: entry-level work, apprenticeship paths, and the route from checker to expert
Identifies which entry-level tasks in their area an agent would remove and what would replace them as learning
Maintains an explicit pipeline plan with headcount, tasks reserved for development, and a ten-year view of where experts will come from
Contributes to sector or public policy on the expertise pipeline
D3Manages transitions fairly: consultation, reskilling, redeployment, and honest communication
Delivers a transition for their own team in line with law and the organisation’s commitments
Designs the organisation’s transition policy and has led at least one transition with documented outcomes for the people affected
Shapes good practice on transitions through research, standards or policy work
Capacity measurement and evidence
The whole capacity of the workforce is measured, including verification hours and rework, and claimed gains are tested against realised outcomes.
E1Measures total capacity, not agent output: hours saved, hours spent checking, rework, and net effect
Reports net capacity for agents in their area, with checking time included
Owns the organisation’s capacity measure and reconciles it with finance; vendor-reported savings are not accepted without it
Has published or validated a capacity measurement method
E2Distinguishes pilot and experimental gains from realised, sustained outcomes
Can explain why a pilot result may not generalise and what would confirm it
Requires a post-deployment evaluation at a set interval for every agent and acts on the results
Contributes evaluations or data to the public evidence base
E3Reports to leadership with uncertainty stated plainly and assumptions visible
Presents results with the assumptions and limits named
Briefs the executive team and board in those terms and has changed a decision by doing so
Sets reporting standards used by others, or teaches them
Ethics, law and accountability
Accountability is held, not shared; the law is met and anticipated; the holder says no when the evidence requires it and shows their reasoning.
F1Holds accountability undivided and resists its diffusion to agents, vendors or committees
Can state who is accountable for each outcome in their area and has corrected a case where no one was
Has designed the organisation’s accountability structure for hybrid work and defended it under pressure
Has shaped how accountability for agent work is understood in law, standards or professional practice
F2Meets and anticipates legal duties: AI literacy of staff, employment and consultation law, data protection, sector regulation
Ensures staff in their area meet the AI-literacy duty appropriate to their roles and keeps records
Owns the organisation’s compliance with workforce-related AI obligations and has briefed counsel and the board on emerging ones
Contributes to the development of law, standards or guidance in this area
F3Exercises professional courage: declines, documents dissent, and learns in public
Has recorded a dissent from a decision in their area, with reasons
Has declined an executive request on evidence or ethics grounds and documented it; shares lessons from failures inside the organisation
Publishes failures and lessons for the field; mentors others in doing so
Five behaviours that run through every domain
Scored in the professional discussion rather than in the portfolio.
Asks what was measured and how before accepting a claim, from a vendor, a consultant or a colleague.
Holds the question open long enough to hear the people affected, and separates what is known from what is assumed.
Treats the workforce whose composition they decide as a constituency with standing, not as a cost line.
Describes agents, decisions and risks in words a board, a regulator and a frontline employee would all understand in the same way.
Says no when the evidence requires it, and changes position in public when the evidence changes.
What a holder owes, and to whom
Seven principles, twenty standards, a pledge, and enforcement by a panel the Institute does not control. The duties run to the workforce affected by the holder’s decisions, not only to the employer who pays.
- Undivided accountability
The holder accepts personal accountability for the composition of the workforce they govern and for the consequences of each allocation decision. Accountability is not transferred to an agent, a vendor, a model provider, a committee or a dashboard.
- Duty to the people affected
The holder owes a duty of care to the employees, contractors and candidates whose work is changed by agents, including those who leave, that is independent of their duty to the organisation.
- Duty to those who rely on the work
Customers, patients, students, citizens and colleagues who rely on output produced or checked with agents are entitled to have that output verified by someone competent to do so.
- Truthfulness about capability
The holder represents what agents and people can do accurately, including to their own board, and neither inflates productivity claims nor conceals verification cost or failure.
- Rigour without anthropomorphism
The holder describes agents in the language of tools and systems, so that moral and legal responsibility is never rhetorically shifted onto them.
- Stewardship of expertise
The holder protects the organisation’s capacity to produce its own experts, and treats the removal of entry-level work as a decision with a decade-long cost, not a saving.
- Independence of judgement
The holder’s advice and decisions are free from undisclosed financial interest in any vendor, platform or consultancy whose products they deploy or recommend.
Accountability and ownership
- 2.1Maintain, or ensure the organisation maintains, a register naming one accountable human owner for every agent in production, with that owner’s competence to check the work and authority to stop it confirmed in writing.
- 2.2Do not deploy, approve or recommend an agent for work that no available person is competent to verify, and document the reasoning whenever a deployment is declined or reversed on these grounds.
- 2.3Do not sign off on, or allow others to sign off on, agent output they have not had the time, expertise and authority to check.
Honesty and evidence
- 2.4State verification cost, rework and failure rates alongside any productivity claim made to leadership, a board, a regulator or the public.
- 2.5Do not present pilot or vendor-reported results as realised organisational outcomes.
- 2.6Correct publicly any material misstatement about workforce or agent performance for which they were responsible, as soon as it is known.
Duties to the workforce
- 2.7Consult the people whose roles will change before the decision is final, in substance and not only in form, and comply with all applicable consultation, employment and anti-discrimination law.
- 2.8Design redesigned roles so that the accountable human has the time and authority to carry the accountability; do not create roles that are accountable in name only.
- 2.9Maintain and report a plan for the expertise pipeline whenever entry-level work is reduced.
- 2.10Ensure that staff who operate or rely on agents receive AI literacy appropriate to their role and context, as the EU AI Act and comparable laws require, and keep records.
Language and representation
- 2.11Use the vocabulary of selection, deployment, evaluation, update and retirement for agents in formal documents, policies and public statements; do not describe agents as employees, colleagues or team members in any document that allocates responsibility.
- 2.12Do not misrepresent the credential, its level, or the Institute’s endorsement; use a designation only as awarded and only while it is current.
Conflicts and independence
- 2.13Disclose in writing to the employer or client any financial interest, past or present, in a vendor, platform or consultancy relevant to a decision, and recuse from the decision where the interest is material.
- 2.14Accept no payment, equity or benefit from a vendor in connection with a deployment decision.
- 2.15Assessors and board members: do not assess, vote on or influence the application of anyone they have taught, employed, coached or advised in the preceding two years, or with whom they have a financial relationship.
Confidentiality and data
- 2.16Protect workforce data, including agent performance data that can identify the people who supervised it, to the standard of applicable data-protection law and of the organisation’s own commitments to its staff.
- 2.17In portfolios, essays and research submitted to the Institute, anonymise organisations and individuals unless written consent to name them is held.
Professional conduct
- 2.18Maintain the competencies in this framework through continuing development and peer supervision, and do not practise at a level a credential does not cover.
- 2.19Report to the Institute a credible breach of this code by another holder, and cooperate truthfully with any investigation.
- 2.20Do not bring the credential into disrepute through conduct, in or out of professional life, that a reasonable member of the public would regard as incompatible with the principles above.
I accept personal accountability for the workforce decisions I make and advise on. I will not allocate work to an agent that no competent person can check. I will tell the truth about what people and agents can do. I will protect the people whose work I change and the expertise my organisation will need. I will hold to this code and accept the Institute’s process if I fall short of it.
Complaints are heard by an Ethics and Appeals Panel with a majority of members from outside the Institute. No one who taught, employed or assessed a holder sits on any stage of their case. Suspensions and revocations are published; other outcomes are published anonymised, with an annual summary. The Institute’s founder, staff and faculty are subject to the code like any holder.
Not yet. Deliberately.
The Institute will credential people against this framework only after an independent credentialing board is seated and a first cohort has been assessed, free of charge, entirely by external assessors, with the results published. That is expected in 2028.
Until then there is no designation to buy. There is a framework to use, a code to adopt, and a pilot cohort to join if you already hold the mandate.