Can You Use AI on Sessions Submitted for an ICF Credential?

When you submit recorded sessions for an ICF credential, you are submitting evidence of one thing: that you, unaided, can coach at the level the credential names. The ICF credentialing requirements ask for recorded coaching sessions and a mentor-coaching record because an assessor needs to watch your competency, not a tool’s. Read that purpose plainly and most of the AI question answers itself - but not all of it, and the part that does not is exactly where coaches get anxious. This article works through it inside the guide to where AI belongs in coaching, where the credentialing answer is one branch of a larger decision.

So let us be precise about what the credentialing evidence is for, what ICF has actually documented, and where ICF has said nothing at all. An ICF credential submission on the ACC or PCC pathway contains recorded coaching sessions and a mentor-coaching record. The assessor evaluates the recording against the ICF Core Competencies markers. That is the function of the evidence: to demonstrate the candidate’s own coaching competency to a trained reviewer.

Name that purpose clearly, because the purpose is the rule. The credentialing question is not “is AI banned.” It is “does this use of AI put the candidate’s own competency in front of the assessor, or something else.” The recorded session exists to let an assessor watch you coach against the markers - that is its entire job. The mentor-coaching record is a different piece of evidence with a different job: showing that the candidate has reflected on and developed their coaching with a qualified mentor coach.

ICF publishes documented credentialing requirements - recorded-session specifications, mentor-coaching hour requirements - and this article works from those documented standards, saying so each time it cites them. The wider regulatory picture sits underneath this answer; what ICF’s AI Coaching Framework requires is the framework, and the credentialing question is one application of it.

Why the Recorded Session Is Where the Line Is Strictest

The recorded session submitted for assessment is the one place in a coaching practice where the augmentation-versus-substitution line is not a matter of personal judgment. It is a matter of what the credential certifies. Across an ordinary practice a coach weighs each AI use against their own discernment. On a credentialing recording the standard is fixed: the conversation has to be the candidate’s own coaching, because that is the thing being certified.

An AI tool that generates questions, reframes, or prompts inside the recorded conversation is putting its competency, not the candidate’s, in front of the assessor. ICF Core Competency 5 asks the coach to maintain presence, and Competency 7 asks the coach to evoke awareness through questioning and intervention. An in-session AI assistant that suggests the next question is performing the evocation work; an assistant the candidate consults mid-session is breaking presence. The assessor scoring those markers would be scoring the tool. That is substitution, and the broader cluster-wide treatment of where the augmentation vs. substitution line is stricter on credentialing sessions sets out why the same act that is a judgment call elsewhere becomes a hard line here.

A note-taker is a different case for the recording itself. Here it helps to be technically precise about what each category of tool actually does to the artifact. An in-session AI assistant alters the conversation - it changes what gets said in the room, because the coach acts on its suggestions. A note-taker does not. It reads the conversation after the fact, generating a transcript and a summary, and it does not change a word the coach said in the session. The recording remains the candidate’s unaided coaching. The candidate needs that distinction to know which tools touch the assessed artifact and which do not.

An in-session AI assistant changes the conversation the assessor is scoring. A note-taker only reads it afterward. One touches the evidence; the other does not.

The note-taker still raises two questions a candidate must not skip. The first is disclosure and confidentiality, which the ethics section below handles. The second is what gets submitted: the assessed evidence is the session recording, never the AI-generated summary. An assessor evaluates the actual coaching conversation, not a tool’s account of it.

Where is ICF silent here? ICF’s documented credentialing requirements specify the recorded-session format and the competency standard. They do not enumerate every AI tool, and as of this writing ICF has not published AI-specific credentialing guidance naming, for example, AI transcription. This article will not invent that ruling. The durable answer is the principle - the assessed artifact must be the candidate’s own coaching - and that principle holds whether or not ICF ever publishes a tool list. The list is not ICF’s to provide today, and not this article’s to fabricate.

AI Before and After the Session - A Different Question

The anxious candidate often fears that any AI contact with their practice contaminates the submission. It does not. Preparation before the recorded session, reflection after it, and the work of assembling a mentor-coaching write-up are ordinary practice activities. They follow ordinary practice rules, not the stricter credentialing-artifact rule. The caution in the previous section is real and worth taking seriously - which is exactly why this section grants the room around the submission honestly, so the caution reads as precision rather than blanket prohibition.

Preparation before a session is ordinary practice. Using AI to think through a client situation, review themes from prior sessions, or sketch a possible structure does not contaminate the recording. The one way prep becomes a problem is if the prepared material is read verbatim into the session - and that would be a presence problem visible on the recording itself, not a hidden contamination. The assessor would see a coach reciting rather than coaching. The prep is not the issue; scripted delivery is.

Reflection after a session is reflective practice, not credentialing evidence. A candidate using AI to review their own notes, surface patterns, or prompt their thinking is doing the ordinary discernment work the cluster teaches. It is governed by that discernment, not by a credentialing rule, because the post-session reflection is not the artifact an assessor evaluates.

The mentor-coaching record is its own submission, and it is the most overlooked half of this question. Candidates fixate on the recorded session and forget the mentor-coaching evidence has its own integrity requirement. The mentor-coaching relationship exists to develop the candidate’s coaching through reflection with a qualified mentor coach. AI cannot stand in for the mentor coach, and it cannot stand in for the candidate’s own reflection. The record must reflect a real developmental relationship. AI used to tidy the wording of a written reflection is a formatting aid and unremarkable. AI used to generate the reflection itself defeats the record’s purpose - there would be no developmental thinking for the record to document.

ICF’s documented mentor-coaching requirements specify the relationship structure and the hour requirements. They do not address AI tools in reflection write-ups. This article states that plainly and applies the principle - the record must be the candidate’s own developmental reflection - rather than inventing a rule ICF has not written.

Consider a candidate who used an AI tool to brainstorm approaches before a session they later decided to record and submit. They now worry the recording is disqualified by association. Walk the worry to its resolution: the assessor reviews the recording. The prep is invisible to the assessor and irrelevant to the assessment - unless it surfaced inside the session as scripted, read-aloud questions, which would be a presence problem the assessor could see directly. If the recording shows the candidate coaching, it stands. Prep before the session and competency on the recording are two different things, and the credentialing rule attaches only to the second.

The Ethics Layer the Credentialing Question Sits Inside

The credentialing question does not float free of ethics. It sits inside the same Code that governs every coaching session. A recording made for a credential submission still triggers the disclosure and confidentiality obligations of the ICF Code of Ethics - and a credentialing submission adds a wrinkle the candidate has to account for.

Standard 2.5 of the Code commits the coach to:

“Be aware of and set clear, appropriate, and culturally sensitive boundaries that govern interactions, physical or otherwise, that I may have with my clients or sponsors.”

That is the standard. A note-taker or any recording tool used on a session destined for a credential submission is an interaction the standard asks the coach to govern openly. The client must know what is being recorded, by what tool, where the recording goes, and - this is the credentialing-specific part - that it will be reviewed by an ICF assessor. The mechanics of writing that disclosure into a coaching agreement are their own subject; Standard 2.5 still applies when the client is also the evaluator covers the operational template, and a candidate preparing a submission should work from it rather than improvise the wording.

Confidentiality, addressed in Section 4 of the Code of Ethics, governs how the recording is handled. A credentialing submission means a third party - the ICF assessor - reviews the session. That is not a confidentiality breach. It is a disclosed, consented disclosure. But it only stays on the right side of the Code if it was actually disclosed and actually consented to before the session was recorded.

Here is the credentialing-specific wrinkle. When the candidate is recording a session specifically so it can be assessed, the client is consenting to more than coaching. They are consenting to their session becoming evidence in someone else’s credentialing process. Standard 2.5 disclosure still applies, and arguably applies with more care, because the client’s consent now has a second object. The article names this without over-interpreting it: the Code establishes the disclosure and confidentiality principles, and it does not script the exact credentialing-submission consent language. Where ICF is silent on the precise wording, the honest move is to apply the principle and verify current expectations, not to manufacture a clause and present it as ICF’s.

One line worth stating plainly near this content: this article reads documented ICF standards as professional education. It is not legal advice, and it is not a substitute for confirming current credentialing and ethics requirements directly with ICF or with an ICF-accredited training provider.

Before You Submit: The Questions to Answer First

The analysis above distils into a sequence a candidate can run on their own submission. Each question gates the next, and the first one can end the evaluation early. Pull the specific session you intend to submit, and answer these four questions about it in order.

1. Did any AI tool participate in the coaching conversation on the recording? This is the early disqualifier. If an AI assistant generated questions, reframes, or prompts inside the recorded session, that recording is not a clean demonstration of your competency - stop here, and do not submit that session. If no tool participated in the conversation, the recording is your own coaching and the evaluation continues.
2. Is the submitted artifact the recording itself, not an AI-generated summary? The assessor must review the actual session. An AI-generated transcript or summary is a convenience for you, never the assessed evidence. Submit the recording.
3. Was the client told - and did they consent to - the recording, the tool, and the ICF assessor review? Standard 2.5 disclosure applies. If the disclosure conversation did not happen before the session, it cannot be retrofitted onto a recording already made.
4. Does the mentor-coaching record reflect your own developmental reflection? AI can tidy the writing. It cannot be the reflection. If the record would not exist as your own thinking without the AI, it does not meet the record’s purpose.

A candidate who can answer all four cleanly can do something specific and valuable: explain, to an ICF assessor, exactly how AI did and did not touch their submission. That is a defensible position, and it is the point of running the sequence.

Consider a candidate running the four questions on a session they plan to submit. Question 1: no AI tool participated in the conversation - it passes. Question 2: they had been planning to submit the AI-generated transcript for the assessor’s convenience - corrected, they will submit the recording. Question 3: they disclosed the recording to the client but never mentioned that an ICF assessor would review it - a gap, the disclosure was incomplete. Question 4: the mentor-coaching record is their own reflection - it passes. The sequence produces a verdict: the session is salvageable, but the disclosure gap means this particular client must be re-consented or a different session submitted. The candidate now has a specific, time-bound action instead of a vague dread.

The sequence cannot do one thing, and the honesty about that is the trust signal. It cannot confirm a specific ICF policy that ICF has not published. ICF’s documented standards are the floor, not a complete AI rulebook, and ICF may issue AI-specific credentialing guidance in the future. Where a candidate needs a verdict the documented standards do not give, the honest move is to verify current requirements directly with ICF or with an ICF-accredited training provider - which is also the most reliable way to walk the pathway with the rules held clearly.

Before you upload anything for assessment, pull the specific session you plan to submit and answer those four questions about it. If you can answer all four cleanly, you can defend your submission to any ICF assessor. If you cannot, you now know exactly which session to re-record and which conversation to have - before the deadline, not after it.

If you would rather walk the credentialing pathway with these rules held clearly from the start, that is what an accredited program is for. Tandem prepares and assesses this exact submission - the recorded sessions, the mentor-coaching record - on Tandem’s ICF ACC program and Tandem’s ICF PCC program, where the AI question gets answered as you train rather than the night before you submit.

This article quotes the ICF Code of Ethics, Standard 2.5, and references the 2025 ICF Code of Ethics, the documented ICF credentialing requirements, and the ICF AI Coaching Framework (November 2024). It is professional education, not legal advice. Where ICF has not published guidance on a specific AI question, this article says so rather than inventing a ruling - confirm current credentialing requirements directly with ICF or with an ICF-accredited training provider before you submit.