AI Enneagram Assessment: Our Ethics and Transparency

01 Our approach to ethics

EnneagramTest is a professional tool built for therapists, coaches and counsellors. That means the people using it (both the practitioners who send the link and the clients who take the assessment) deserve a clear account of what it does, how it works, and what responsibilities we accept in running it.

We take those responsibilities seriously: to clients, whose inner lives are the subject of every session; to practitioners, whose professional reputation is bound up in the tools they recommend; and to the Enneagram itself, a rich framework that is too often reduced to a social media quiz. Our aim is to work at the level the framework deserves.

This page sets out our commitments plainly. If anything here is unclear, or if you have concerns we haven't addressed, please write to us at [email protected].

02 What these assessments are

An EnneagramTest assessment is a guided personal discovery conversation conducted through our AI, exploring personality patterns, motivations and instinctual drives through the lens of the Enneagram. The conversation typically runs for 20–30 minutes and is conversational in nature: the client is asked questions, invited to reflect, and gently followed up on where more depth is possible.

At the end of the session, the AI generates a detailed, personalised report that covers the client's dominant type and confidence level, subtype profile, core motivations, strengths, blind spots, growth areas, and more. The client is immediately emailed their Enneagram type and subtype. The practitioner then reviews the full report and, once they are satisfied it is ready, releases it to the client via a secure link.

The purpose of the assessment is to provide a rich, thoughtful starting point for deeper work with a practitioner. It gives both client and practitioner a shared language and a substantive base to build from. The conversation itself can be valuable (many clients report that the process of articulating their patterns out loud is clarifying in its own right), but the report is designed to be a resource, not a conclusion.

03 What these assessments are not

It matters as much to be clear about what EnneagramTest is not as about what it is.

The Enneagram is an interpretive framework with a rich tradition of practical application, but it is not a validated psychometric instrument in the clinical sense. The reports we generate reflect what a client shared during their session and should be understood as an aid to self-recognition, not an objective measurement of personality. A client who disagrees with their result may well be right. Disagreement itself is often productive ground for a practitioner conversation.

04 AI transparency

The EnneagramTest assessment is conducted by an AI (specifically, Claude, developed by Anthropic). The AI works to our methodology, follows our assessment structure, and generates the final report based on the full context of the conversation. The practitioner does not conduct the conversation directly: they send the assessment link to their client, and the AI handles the session on their behalf.

We made a deliberate decision to be transparent about this rather than to present the AI as a human guide or to obscure its nature. Clients are told clearly at the start of the session that they are speaking with an AI guide, and that the AI is representing their practitioner. This is a firm commitment, not a preference. No client of EnneagramTest should ever be misled about whether they are speaking to a human.

Why transparency matters: There is a debate in the AI industry about whether disclosing an AI's nature changes the quality of conversation; some research suggests people share differently when they know they're talking to a machine. Our view is that this consideration does not override the basic ethical obligation of honesty. Clients deserve to know what they are engaging with. We have designed the AI guide to be warm, thoughtful and skilled within the constraints of full transparency, and client engagement has been consistently positive.

Practitioners are not permitted to represent the AI as a human, or to suggest to clients that the assessment is being conducted by them personally in real time. If you have questions about how to position the assessment to clients, the Practitioner Manual includes suggested language.

05 Data and privacy

When a client takes an EnneagramTest assessment, their conversation with the AI is processed in real time by Anthropic's Claude API to generate the report. This means the content of the conversation passes through Anthropic's infrastructure as part of that generation process, subject to Anthropic's data handling and privacy policies.

What we store

The conversation transcript and the generated report (PDF and HTML, held in Cloudflare R2 under EU jurisdiction) are automatically and permanently deleted 30 days after the session by a scheduled daily job. After that point, the platform retains only a lightweight record: the client's name, email address, and Enneagram type. These records are not shared with third parties.

At the end of the session, the client is immediately emailed their Enneagram type and subtype. The practitioner then reviews the full report and, once they are satisfied it is ready, releases it to the client via a secure link.

Your rights

Clients have the right to request access to, or deletion of, any data we hold about them. These requests can be made via [email protected]. Full details of our data practices are set out in the Privacy Policy.

06 Security

We take the security of client data seriously and have designed the platform with appropriate safeguards at each layer.

• Reports are emailed over encrypted connections: all email delivery uses TLS encryption in transit.
• Primary client data is stored in Cloudflare D1, with generated reports held in Cloudflare R2 under EU data residency.
• Practitioner dashboards are protected by Cloudflare Access: access to the practitioner-facing admin interface requires email OTP verification via Cloudflare. There are no stored passwords; each login is separately authenticated.
• Assessment links are single-use per client: links are tied to a specific client record and cannot be reused or shared to access another client's session.

If you identify a security concern or vulnerability in the platform, please contact us immediately at [email protected].

07 Your freedom in the conversation

Participation in an EnneagramTest assessment is always voluntary. Clients may skip any question they are not comfortable answering; the AI will acknowledge this without pressure and move on. No part of the conversation is compulsory.

Clients may also end the session at any point. If a session is ended early, no report is generated and nothing is recorded beyond the fact that the session was not completed.

The quality and personalisation of the report does depend on the depth of engagement during the conversation. A client who engages openly and reflectively will receive a richer, more individually tailored report than one who gives minimal responses. The AI is designed to work thoughtfully with whatever a client offers, but we are transparent that more engagement produces a better outcome. This is offered as information, not as pressure.

08 Wellbeing and consent

EnneagramTest assessments are restricted to adults aged 18 and over. The platform is not designed or validated for use with younger people, and practitioners should not send assessment links to clients under 18.

Before a session begins, clients must read and actively consent to the session terms. These include a plain-language explanation of what the assessment involves, how data is used, and the nature of the AI guide. A session cannot begin without this consent step being completed.

Emotional depth

The Enneagram works at the level of core motivation: the fears and desires that sit beneath much of how people behave and relate. This depth is what makes it valuable in a practitioner context, but it also means the conversation can occasionally land somewhere personal and significant for a client. Type descriptions touch on childhood patterns, emotional defences, and core wounds. This is by design; it is also a reason for care.

The assessment is not appropriate for people who are currently in acute mental health crisis, or who are at significant risk of harm. If a client raises concerns during the session that suggest they need immediate support, the AI is designed to pause and provide appropriate signposting.

UK mental health support

If a client is in distress or mental health crisis, they should contact their practitioner directly, or reach out to a dedicated support service:

• Samaritans — 116 123 (free, 24 hours a day, 7 days a week)
• Mind — mind.org.uk/need-urgent-help
• Crisis text line (Shout): Text SHOUT to 85258 (free, 24/7)