AI tools and client data
Is Granola safe for client data?
Short answer: Whether Granola is safe for client data turns on which plan your staff are on, which account they signed in with, and what you have told them they may put in it — not on the product itself. On Individual and team plans, Granola may use what the meeting notes contain to train its models unless somebody turns that off; on Enterprise, Granola says it does not train its models on what the meeting notes contain, by default. Every fact on this page is Granola's own, quoted from its own pages and dated, and where we did not find a statement this page says so rather than filling the gap. Obsidian Ridge's free 30-day AI security assessment shows which AI tools your staff actually use.
See which AI tools your staff usePlan by plan, in Granola’s own words
A notetaker that listens on the laptop rather than joining the call as a participant, so it leaves no bot in the meeting and nothing for the other side to notice. Granola says the recording goes once the transcript exists; the transcript and the notes stay.
| What you are asking | Individual and team plans | Enterprise |
|---|---|---|
| Trains the vendor's models by default | Yes, on de-identified data, until somebody turns it off. Granola puts the switch in the account rather than asking an administrator.“We only use de-identified data to train AI models, which you can opt-out of” source, read | Off by default here, which is the opposite of the individual plans — and Granola notes an enterprise can still choose to opt in.“Enterprise users have model training turned off by default, but may opt-in through the Services.” source, read |
| How to turn training off | In your own settings, per person. Which means it is on for everyone who has not been told to look.“You can opt-out within your Granola account settings” source, read | Not found in the pages we read. |
| How long data is kept | As long as the account lives. No period of its own, and see the note below about what happens to anything already in a model.“We will retain your information as long as necessary to serve you, to maintain your account for as long as your account is active…” source, read | As long as the account lives. No period of its own, and see the note below about what happens to anything already in a model.“We will retain your information as long as necessary to serve you, to maintain your account for as long as your account is active…” source, read |
| What an admin can see or control | Not found in the pages we read. | Settings an administrator sets and a user cannot override, which is what makes the default above hold.“Enterprise Workspace Products have admin-enforced settings and are opted-out by default.” source, read |
| Single sign-on | Not found in the pages we read. | Not found in the pages we read. |
| Business Associate Agreement | Not found in the pages we read. | Granola will execute a BAA, on Enterprise. A practice on an individual or team plan is not offered one in the pages we read.“Enterprise Services Users that are Covered Entities or Business Associates under HIPAA may execute a Business Associate Agreement with Granola…” source, read |
| Where the data is processed | The United States, on AWS.“We store your Personal Data in Amazon Web Services ("AWS") servers located in the U.S.” source, read | The United States, on AWS.“We store your Personal Data in Amazon Web Services ("AWS") servers located in the U.S.” source, read |
An empty cell is not a reassurance and not a criticism. "Not found in the pages we read." means the Granola pages listed under Sources, read on the dates shown, do not answer it — so there is nothing for us to repeat, and something for you to ask Granola before client information goes near it.
What this means for
The duty does not change because the tool is new. What changes is where the client information ends up.
- Law firms. A client confidence pasted into a personal account has left the firm’s control. Our law firm page sets out the duties we build evidence for; the decision here is whether a personal account may touch a client matter at all.
- Dental and medical practices. Granola publishes a BAA position for this product on Enterprise, which is where that conversation starts. See dental practices.
- CPA and accounting firms. Your written information security plan has to describe the tools people actually use, not the ones you approved. See accounting firms.
- Financial advisers. Client account data in a personal assistant is the exposure; see financial advisers.
- Software teams. A coding assistant reads the repository and can run commands, so the question is what it can reach, not what it was asked. See what AI agents on work laptops can reach.
How to tell which version your staff use
The plan is the whole answer, and you cannot tell from the outside. Three ways to find out, in order of how much they prove.
1. Ask
Quick, free, and incomplete. People forget the tool they installed once and the account they signed in with.
2. Look at the billing and the admin console
This finds what the business pays for. It does not find a personal account, which is the one that matters here.
3. Look on the computers
A tool somebody installed themselves never appears in an IT list. Obsidian Ridge’s free 30-day AI security assessment takes a quiet inventory for 30 days and shows what each tool could reach, then a CISSP goes through it with you. See the assessment, or read a sample set of findings.
Settings to check today
From Granola’s own documentation. Each one is a change somebody can make this afternoon.
- Turn the training opt-out on, per person, on anything that is not an Enterprise workspace. Nobody else will do it for them.“You can opt-out within your Granola account settings” source, read
- Know that the audio does not survive the transcript. What you are governing afterwards is text, not a recording.“We do not retain or store such recordings once the transcription is created.” source, read
- Revoke the integration tokens when somebody leaves. Granola deletes them on revocation, which means revocation is the trigger and it is yours.“Access control features, such as license keys and authorization tokens, for third-party integrations or connections are retained only while the connection is active…” source, read
Worth knowing
- The most candid sentence any vendor on this site publishes. Anything already de-identified and used in a model may never come out.“De-identified and aggregated data, including when used in AI models, may be retained indefinitely, as it cannot reasonably be attributed to you individually.” source, read
- And Granola explains why, which applies equally to every other vendor here that trains on de-identified data.“removal may not be technically feasible without complete model retraining or dataset reconstruction.” source, read
- The model providers are barred separately, so the training above is Granola's own and not theirs.“We do not allow third parties such as OpenAI or Anthropic to use your Personal Data to train AI models.” source, read
Questions we get asked
Is Granola safe for client data?
Whether Granola is safe for client data turns on which plan your staff are on, which account they signed in with, and what you have told them they may put in it — not on the product itself.
On Individual and team plans, Granola may use what the meeting notes contain to train its models unless somebody turns that off; on Enterprise, Granola says it does not train its models on what the meeting notes contain, by default.
Every fact on this page is Granola's own, quoted from its own pages and dated, and where we did not find a statement this page says so rather than filling the gap.
Does Granola train its models on what we put into Granola?
On Individual and team plans, Granola may use what the meeting notes contain to train its models unless somebody turns that off; on Enterprise, Granola says it does not train its models on what the meeting notes contain, by default.
On Individual and team plans: Yes, on de-identified data, until somebody turns it off. Granola puts the switch in the account rather than asking an administrator.
Granola's own words: "We only use de-identified data to train AI models, which you can opt-out of" (read 2026-11-03).
On Enterprise: Off by default here, which is the opposite of the individual plans — and Granola notes an enterprise can still choose to opt in.
Granola's own words: "Enterprise users have model training turned off by default, but may opt-in through the Services." (read 2026-11-03).
How long does Granola keep what we put into Granola?
On Individual and team plans, Enterprise: As long as the account lives. No period of its own, and see the note below about what happens to anything already in a model.
Granola's own words: "We will retain your information as long as necessary to serve you, to maintain your account for as long as your account is active…" (read 2026-11-03).
Can an administrator read what we put into Granola?
On Individual and team plans: Not found in the pages we read. On Enterprise: Settings an administrator sets and a user cannot override, which is what makes the default above hold.
Granola's own words: "Enterprise Workspace Products have admin-enforced settings and are opted-out by default." (read 2026-11-03).
Will Granola sign a Business Associate Agreement for Granola?
On Enterprise: Granola will execute a BAA, on Enterprise. A practice on an individual or team plan is not offered one in the pages we read.
Granola's own words: "Enterprise Services Users that are Covered Entities or Business Associates under HIPAA may execute a Business Associate Agreement with Granola…" (read 2026-11-03).
A signed BAA is between your practice and Granola; we do not sign it for you, and we do not read a compliance badge as one.
How do we find out which version of Granola our staff are using?
Look on the computers. A tool somebody installed themselves, or signed into with a personal account, does not appear in any IT list. Obsidian Ridge's free 30-day AI security assessment takes that inventory for 30 days and shows what each tool could reach.
Sources
Granola’s own pages, with the date each was read. Nothing on this page comes from anywhere else.
All 35 tools, side by side, or run the AI tool risk checker over the ones you know about.
Other tools people ask about
Same question, different answers. The ones in the same category are the ones staff swap between.
Is Acrobat AI Assistant safe for client data?
Document assistant · Adobe
Is Zoom AI Companion safe for client data?
Meeting notetaker · Zoom
Is Otter.ai safe for client data?
Meeting notetaker · Otter.ai