Responsible AI
Last changed 15 September 2026
In short
AI is a small, optional helper here, never the coach. CoachYantra uses it in three places: to suggest the coaching competencies a coach’s own journal entry touches, to suggest more reflection questions, and to answer questions about the product and coaching practice.
The coaching stays human. AI never listens to, transcribes, summarises or analyses a session. It never reads a coach’s private notes, anything a coachee writes, messages, agreements, reports, invoices or payments, and it never decides anything about a person.
People decide. Everything AI offers is labelled as a suggestion, and nothing it suggests is kept unless the person keeps it.
What we hold ourselves to
The coach decides. AI proposes; a person accepts, changes or ignores what it proposes. Competency suggestions are off until the coach switches them on, and the coach can switch them off again at any time.
The session is not material for AI. What happens in a coaching conversation belongs to the coach and the person they coach. No AI feature in CoachYantra is given it, and none is built to infer it.
Send as little as the job needs. Each feature sends the least it can to do its work, and two of the three send nothing anyone has written about themselves or a client.
No judgement of people. AI here does not rate, rank, diagnose or assess a coachee or a coach, and it will not say whether something counts towards a credential. Those are human judgements.
Written by people. Reports to a company are written by the coach and approved by the coachee. Agreements are templates the coach chooses and signs. Neither is generated by AI.
Said plainly. Every place AI appears says so, says what it sends, and marks what it offers as a suggestion.
Where AI is used
Competency suggestions on a coach’s journal
A coach’s journal is for reflecting on their own practice. If the coach switches suggestions on (Setup, Your profile), the words of a journal entry are sent as they write it (apart from answers about how a client showed up), together with the names and definitions of the ICF or EMCC competencies. What comes back is up to five competencies the entry may touch, each tied to a phrase in the entry, shown under “Suggested, for your reflection”. A suggestion is saved only if the coach keeps it, and then only the competency, never the phrase. If an entry reads as being about a client rather than the coach’s own practice, CoachYantra says so before it is saved. Private notes on a session are a separate place and are never sent.
More reflection questions
A coach or a coachee who wants more questions to reflect on can press “Suggest more questions”. Only the angles and competencies they picked, and the questions already offered, are sent; never anything they have written. The questions that come back are labelled as suggested, and one is kept only if it is answered.
The help box
Coaches can ask about the product or about coaching practice. The question is sent with CoachYantra’s own product guide and ICF and EMCC reference material, and nothing from the practice: the help box cannot see clients, notes, journals or figures. Questions that ask for a verdict about a person, or for clinical or legal advice, are refused before anything is sent, as far as CoachYantra can recognise them. What a coach types is sent as written, so it should not include names or anything about a client.
Where AI is never used
Sessions, in any form. A coach’s private notes. Anything a coachee writes, including their journal and their comments on a report. Messages between a coach and a coachee. Reports to a company. Agreements and their clauses. Invoices, tax, payments and payment reminders. Credential hours. Reading a coach’s calendar, which follows fixed rules. Any decision about an account, a price or access to the product.
Who provides the AI
The three features use Claude, from Anthropic, called from our servers. CoachYantra does not store what is sent, keeps only the suggestions a person chooses to keep, and writes neither to its logs. We do not allow what we send to be used to train models: To confirm: to confirm against the signed Anthropic terms before this page is published If the AI is unavailable, each feature carries on without it. Anthropic is listed with our other providers in the privacy notice.
How we keep to this
These promises are checked in the code, not only written here. Our test suite, which must pass before any change is released, fails if AI is called from anywhere new, if the help box or the question writer can read practice data, if private notes could reach AI, or if a suggestion could be saved with a phrase from the entry; the database itself refuses a stored phrase. The coach’s choice about suggestions is checked on our server, not only on the screen.
Changes
Before AI is used anywhere new, this page will say so and the date above will change. A new use that would send what someone has written will be off until the person chooses it.
Questions or concerns about how CoachYantra uses AI: our privacy officer, Glady Cruz, at glady@mikecruzllp.com, or coachyantra@mikecruzllp.com.
This page describes what the software actually does; every factual claim in it is checked by a test or a database constraint in the product. It is not legal advice. We keep it current and revise it as our business and legal requirements evolve.
CoachYantra™
© 2026 Mike Cruz LLP. All rights reserved. CoachYantra™ and the CoachYantra mark are trade marks of Mike Cruz LLP.