AI tools and client data
Is Acrobat AI Assistant safe for client data?
Short answer: Whether Acrobat AI Assistant 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 Acrobat and Reader with the AI Assistant add-on and on Acrobat AI Assistant for enterprise, Adobe says it does not train its models on the documents staff open, by default. Every fact on this page is Adobe'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 Adobe’s own words
The assistant built into Acrobat and Reader: it summarises a PDF, answers questions about it and cites the passage. Nobody installs it — it appears in the PDF reader already on the machine, next to whatever document is open, which in a practice is a contract, a chart or a tax return.
| What you are asking | Acrobat and Reader with the AI Assistant add-on | Acrobat AI Assistant for enterprise |
|---|---|---|
| Trains the vendor's models by default | No, and Adobe says it of fine-tuning as well as training. Its model provider is barred by contract separately.“Adobe does not use customer data to train or fine-tune LLMs.” source, read | No, and Adobe says it of fine-tuning as well as training. Its model provider is barred by contract separately.“Adobe does not use customer data to train or fine-tune LLMs.” source, read |
| How to turn training off | Not found in the pages we read. | Not found in the pages we read. |
| How long data is kept | Twelve hours for the document, the prompt and the answer. Read it with the two exceptions in the notes below.“the uploaded document, prompts, and generated responses from a session are all automatically deleted from the Adobe cloud after 12 hours.” source, read | Twelve hours for the document, the prompt and the answer. Read it with the two exceptions in the notes below.“the uploaded document, prompts, and generated responses from a session are all automatically deleted from the Adobe cloud after 12 hours.” source, read |
| What an admin can see or control | Not found in the pages we read. | An administrator picks who gets it, by user and group, rather than it appearing for everybody with Acrobat.“including granular admin-level controls, such as selecting users and user groups for access and enabling just-in-time provisioning of the product.” 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. | Not found in the pages we read. |
| Where the data is processed | It depends which Acrobat. On the desktop and mobile apps the chat history is on the machine; in the browser it is in Adobe's cloud.“On the desktop and mobile versions of Acrobat, chat history is stored on the user’s device, whereas in the web version…” source, read | It depends which Acrobat. On the desktop and mobile apps the chat history is on the machine; in the browser it is in Adobe's cloud.“On the desktop and mobile versions of Acrobat, chat history is stored on the user’s device, whereas in the web version…” source, read |
An empty cell is not a reassurance and not a criticism. "Not found in the pages we read." means the Adobe 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 Adobe 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. We didn't find a BAA statement in the pages listed below; ask Adobe before any patient data goes in. 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 Adobe’s own documentation. Each one is a change somebody can make this afternoon.
- Turn off cloud chat history for the organisation. It is the one setting that moves the retained part of the conversation off Adobe's cloud.“Enterprise admins can also use the Adobe Admin Console to turn off cloud-based chat history storage for their organization.” source, read
- Know the scope. The answer comes from the open document, which means the exposure is the document somebody chose to open and no more.“By default, AI Assistant grounds its responses in the content of the document you open or explicitly select.” source, read
- And know what it does not reach, which is worth telling staff who assume the worst: not the file server, not the mailbox, not the web.“Neither does it crawl your company’s file systems, email, or other repositories, nor does it search the public web for answers.” source, read
Worth knowing
- The 12 hours has two exceptions, and both are in Adobe's own sentence: the chat history it keeps, and anything a user reports as a problem.“Except for retained chat history and any content a user reports…” source, read
- Adobe names the model behind it and the contractual bar on both training and human review, which is the pair of commitments to look for.“we currently use Microsoft Azure OpenAI Service, which is contractually prohibited from manually reviewing or training its LLM on Adobe customer data.” source, read
- On a shared document, one person's questions are not visible to the others on it.“Even when interacting with shared documents on the web, responses remain private to each user and are not shared with others.” source, read
Questions we get asked
Is Acrobat AI Assistant safe for client data?
Whether Acrobat AI Assistant 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 Acrobat and Reader with the AI Assistant add-on and on Acrobat AI Assistant for enterprise, Adobe says it does not train its models on the documents staff open, by default.
Every fact on this page is Adobe'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 Adobe train its models on what we put into Acrobat AI Assistant?
On Acrobat and Reader with the AI Assistant add-on and on Acrobat AI Assistant for enterprise, Adobe says it does not train its models on the documents staff open, by default.
On Acrobat and Reader with the AI Assistant add-on, Acrobat AI Assistant for enterprise: No, and Adobe says it of fine-tuning as well as training. Its model provider is barred by contract separately.
Adobe's own words: "Adobe does not use customer data to train or fine-tune LLMs." (read 2026-11-03).
How long does Adobe keep what we put into Acrobat AI Assistant?
On Acrobat and Reader with the AI Assistant add-on, Acrobat AI Assistant for enterprise: Twelve hours for the document, the prompt and the answer. Read it with the two exceptions in the notes below.
Adobe's own words: "the uploaded document, prompts, and generated responses from a session are all automatically deleted from the Adobe cloud after 12 hours." (read 2026-11-03).
Can an administrator read what we put into Acrobat AI Assistant?
On Acrobat and Reader with the AI Assistant add-on: Not found in the pages we read. On Acrobat AI Assistant for enterprise: An administrator picks who gets it, by user and group, rather than it appearing for everybody with Acrobat.
Adobe's own words: "including granular admin-level controls, such as selecting users and user groups for access and enabling just-in-time provisioning of the product." (read 2026-11-03).
Will Adobe sign a Business Associate Agreement for Acrobat AI Assistant?
We didn't find a BAA statement in the pages listed below; ask Adobe before any patient data goes in.
How do we find out which version of Acrobat AI Assistant 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
Adobe’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 Slack AI safe for client data?
Workspace assistant · Slack
Is ChatGPT safe for client data?
General assistant · OpenAI
Is Microsoft Copilot safe for client data?
General assistant · Microsoft