Memory and privacy

What About the People I Talk to Who Never Agreed to This?

> ANSWER[AEO]

Signing up for a conversation-memory tool means you agreed to it; it does not mean the person on the other end of your calls agreed to anything, and they are the one whose words become structured, persistent, AI-queryable data. Account-level consent and the consent data protection law actually asks for are two different things, and the gap between them sits with you, not with the software. The practical answer is a short, clear disclosure at the start of each recorded conversation — Closer does not make that disclosure for you.

Isn't the account holder's agreement enough, since they're the one using the app?

No — and this is a structural feature of how personal data protection works, not a technicality. These frameworks distinguish between the party who controls a tool and the individual whose personal data is actually being processed.

When a salesperson records a call, the client's voice, statements and any personal details they share are that client's personal data. The client is not a party to the agreement between the salesperson (or their employer) and the software provider. An independent basis for collecting that data still has to exist. Regulators and legal commentators are consistent on the point: where consent is required, it has to come from the data subject, and it cannot be inferred from a business relationship with a third-party vendor.

Doesn't the fact that they said it to a real person, out loud, mean it was already given?

Not in the way that matters here. There is a well-recognised difference between disclosing something in a specific conversational context and having that disclosure extracted, structured and persisted somewhere it can be retrieved indefinitely and cross-referenced with other facts about you.

Research on AI and privacy treats this recontextualisation — data given for one purpose being captured, structured and reused in ways the person never anticipated — as one of the core sources of legitimate privacy concern with AI systems, not a fringe worry. Someone answering a rep's question honestly on a phone call has not thereby agreed to become a permanent, linked node in that rep's client-memory system.

So what actually addresses this?

The defensible pattern is that consent is sought at the point of the conversation, not assumed from the account holder's sign-up:

What does Closer actually do about this — and what does it not do?

The honest version, checked against what is built rather than what a mitigation section would ideally say:

None of that substitutes for telling the person. It makes the promise you make them one you can keep.

Key takeaways

  • > Your agreement with a software vendor is not the client's agreement to be recorded and remembered.
  • > Saying something out loud in one context is not consent to have it extracted, structured and stored indefinitely.
  • > Consent belongs at the point of recording, in that conversation, every time — not once at sign-up.
  • > Closer has no consent prompt: making the disclosure is the user's job, not the product's.
  • > Closer keeps no audio, and a client's memory scope can be exported and erased in full.

Frequently asked questions

If I sign up for this app, does that mean the people I talk to have also agreed to be remembered by it?

No. Your agreement to the app's terms covers your own participation — it does not extend to the other person in the conversation, and Closer does not collect their consent for you.

  • They are a third party to your relationship with the software, not a party to it.
  • Their personal data still needs its own lawful basis for collection.
  • In most jurisdictions that means their informed consent to the recording, independent of your account-level agreement with the product.
Does it matter that the other person's information was already 'their data to share' since they said it out loud to me?

It matters less than it might seem. Saying something to one person in a private conversation is different, both intuitively and under most data protection frameworks, from having that statement made permanently retrievable.

  • Extraction, structuring, linkage to an identity and retrieval across every future conversation are new uses, not the original one.
  • Voluntarily disclosing information in one context does not automatically consent to it being repurposed and persisted in another.
  • What that structuring looks like in practice: what a client knowledge graph is.
What should a responsible product actually do about this, given that the salesperson — not the client — is the one signing up?

The research on this points toward disclosure and consent happening at the point of recording, not just at the point of software sign-up. Closer does not do this for you — the disclosure is yours to make, in that conversation, every time.

  • The person on the call needs to be told, in that specific conversation, that it may be recorded and why.
  • Being informed once in a terms-of-service document the account holder agreed to on their behalf does not cover them.
  • Why the imbalance persists even when the rules are followed: memory power asymmetry.
Does Closer ask the client for consent before recording?

No. Closer has no in-app consent prompt or automatic recording announcement today — asking the person in front of you, and deciding whether recording is appropriate at all, is entirely your responsibility.

  • Nothing in the product will prompt you, and nothing will make the disclosure for you.
  • If your process depends on a recorded disclosure, you have to make that disclosure yourself at the start of the conversation.
  • That obligation returns every time you record, not once at sign-up.

Sources