Answer capsule
When an employer pays for coaching, the client and sponsor need an explicit information boundary before the first session—not an improvised update after trust is tested.
What the source establishes
- The 2025 ICF Code of Ethics took effect on April 1, 2025.
- The Code defines a sponsor as the entity or person paying for, arranging, or defining the coaching services.
- Before coaching begins, ICF professionals are expected to explain the nature of coaching and co-create an agreement on roles, responsibilities, confidentiality, and financial arrangements.
- The Code separately addresses confidentiality, information exchange, records, technology, artificial intelligence, and multiple relationships.
Distinguish client, sponsor, coach, and platform
Employer-funded coaching can involve at least four parties with different interests. The executive is the coaching client; a manager, HR team, or enterprise may sponsor the work; a practitioner delivers it; and a platform may store, analyze, or summarize activity. Name each party and its role before services begin. Clarify who selects the coach, who can end the engagement, who pays, who controls records, and whether anyone occupies more than one role. This prevents the sponsor relationship from being treated as invisible administration when it materially shapes expectations and power.
The accountable team should translate this point into a named workflow, affected population, source data, human owner, approval right, exception path, retained evidence, and review date. That translation is what separates an interesting AI development from a decision that can be governed and evaluated.
Specify what the sponsor will receive
A promise of confidentiality is incomplete unless it identifies the reporting boundary. The agreement can distinguish scheduling, attendance, participation status, client-defined goals, aggregate program measures, progress statements, session content, assessments, transcripts, notes, and safety-related exceptions. For every category, state whether it is collected, who receives it, for what purpose, at what level of detail, and with whose authorization. Avoid labels such as engagement score or outcome insight when the sponsor cannot tell what participant behavior or conversation data generated them.
The accountable team should translate this point into a named workflow, affected population, source data, human owner, approval right, exception path, retained evidence, and review date. That translation is what separates an interesting AI development from a decision that can be governed and evaluated.
Re-contract when the purpose changes
Sponsor requests often expand after coaching begins: a manager asks for an update, HR wants themes across a cohort, or a platform offers new analytics. A pre-coaching agreement should not be stretched to cover a materially different use. Pause and explain the proposed change, its beneficiaries, data, consequences, and alternatives to the client and sponsor. Obtain the authorization and other basis the context requires, or decline the request. Record the revised boundary without pressuring the executive to trade session privacy for continued access to a company-funded benefit.
The accountable team should translate this point into a named workflow, affected population, source data, human owner, approval right, exception path, retained evidence, and review date. That translation is what separates an interesting AI development from a decision that can be governed and evaluated.
Evaluate the reporting path, not the promise
Diligence should follow information from session to sponsor. Inspect practitioner notes, platform dashboards, automated summaries, alerts, exports, support access, retention, deletion, subprocessors, and offboarding. Test whether a sponsor administrator can infer sensitive content from metadata or drill-down views even when transcripts are hidden. Ask how conflicts, legal demands, and imminent-risk situations are handled and communicated. The ICF Code is a professional ethics baseline for ICF professionals; it does not by itself establish legal privilege, universal confidentiality, platform security, coaching quality, or the correct response in every jurisdiction.
The accountable team should translate this point into a named workflow, affected population, source data, human owner, approval right, exception path, retained evidence, and review date. That translation is what separates an interesting AI development from a decision that can be governed and evaluated.
Decision test
Ask whether the source changes the decision itself, the evidence required, the implementation sequence, or only the language used to describe an existing capability. Record which claims are directly supported, which are provider statements, which require an independent test, and which remain unknown. A source-linked review should make uncertainty easier to see, not bury it inside a blended score.
Questions to take into review
The publication supports research and executive decision preparation. It does not provide legal, financial, accounting, employment, clinical, cybersecurity, investment, procurement, or implementation advice.