Answer capsule
ICF’s 2026 Coaching Platform Standards say a platform offering evaluation or ROI assessment should document the validity of its method and explain what is measured. A buyer should review the construct, population, data access, and decision use before treating usage or a sponsor dashboard as coaching impact.
What the source establishes
- ICF Coaching Platform Standards v1.1 is dated March 9, 2026 and is written as guidance for platform developers, providers, sponsoring organizations, coaches, clients, procurement, and program administrators.
- The standards say a platform must provide documentation demonstrating the validity of an evaluation or ROI assessment methodology and should be transparent about what is measured, how, and why the constructs matter.
- The standards distinguish platform availability and usage from coaching effectiveness and call for transparency about indicators of coach quality and effectiveness.
- The standards call for transparency about what coaching data is shared, who sees it, why it is shared, and what a data-processing agreement covers, including collection, storage, use, deletion, access, and ownership.
Name the construct before accepting the metric
The direct systems decision is what the platform claims to measure: access, engagement, satisfaction, goal progress, behavior change, manager observation, performance, retention, business result, or return. These constructs answer different questions. A high login rate or completed session count can show use, while an ROI percentage requires an outcome definition, comparison, cost boundary, time period, and attribution method.
Ask for the exact construct, theory of change, population, instrument, scoring rule, baseline, comparison, observation window, exclusions, missing-data treatment, and decision threshold. A dashboard label such as impact, growth, readiness, or return should remain a provider claim until the documentation shows what evidence and inference produce it.
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.
Inspect validity in the configured use
Method documentation should explain development, evidence, limitations, and the conditions in which the assessment is intended to be used. The buyer should compare those conditions with the actual participants, language, culture, roles, delivery modes, coach mix, AI features, sponsor goals, and organizational context. A published method can be valid for one purpose and inappropriate for another.
Test whether the platform preserves the denominator, attrition, low participation, contradictory indicators, negative outcomes, and uncertainty. Provider-selected case stories and aggregate improvement among completers should not be generalized to everyone offered coaching. Independent evidence, where available, should remain distinct from provider research and the buyer's own production measures.
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.
Reconcile measurement with coaching-data access
Impact analysis may draw on goals, sessions, notes, surveys, assessments, feedback, HR records, performance data, or manager reports. The platform should show what the client, coach, sponsoring organization, provider, administrator, model, support team, and subprocessor can see and why. An aggregate chart can still expose a person when the cohort is small or the combination of attributes is distinctive.
The buyer should inspect minimum group sizes, suppression, aggregation, purpose, retention, model-improvement use, correction, deletion, export, and access controls for every input and output. The sponsor's desire to prove value does not erase client confidentiality or turn intimate coaching content into an unrestricted analytics source. Contract language and the configured interface should agree.
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.
Keep platform, coaching, and business conclusions separate
A reliable platform can deliver weak coaching; effective coaching can occur on a platform with limited analytics; and a sound coaching outcome may not produce an attributable business return. The approval record should keep service quality, method validity, coach quality, participant experience, behavior evidence, organizational outcome, and economic result in separate evidence classes with named owners.
ICF's standards are professional guidance, not a platform certification, independent validation, privacy audit, outcome study, or finding that one metric is appropriate. Buyers should verify the current version, actual product configuration, contracts, method documentation, data flows, coach and AI roles, and production evidence. Qualified measurement, coaching, privacy, security, HR, procurement, and legal review may be required.
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.