AI tools and client data
Is CoCounsel safe for client data?
Short answer: Whether CoCounsel 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 CoCounsel Essentials and on CoCounsel Legal, Thomson Reuters says it does not train its models on the matter files you upload, by default. Every fact on this page is Thomson Reuters'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 Thomson Reuters’s own words
Thomson Reuters' legal AI: it reviews documents, drafts and answers research questions against Westlaw and Practical Law. A firm uploads the matter file to use it, so the question is not what it reads but what it keeps.
| What you are asking | CoCounsel Essentials | CoCounsel Legal |
|---|---|---|
| Trains the vendor's models by default | No, and Thomson Reuters extends it to the third-party models behind the product as well as its own.“we do not use your user content and user prompts to train or improve CoCounsel Essentials or any third-party GenAI large-language models…” source, read | No. The CoCounsel Legal page says it in one line and does not say more.“Data remains secure and private, and we never use it to train underlying models.” 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 | There is no published period. Retention is whatever your firm agreed with a Thomson Reuters client manager, which means somebody at your firm has to know what was agreed.“Retention of user content and prompts is customized to fit customer needs and set up at the organization account level…” source, read | Not found in the pages we read. |
| What an admin can see or control | Retention is set for the whole organisation rather than per user, so it is an account-level decision and not a user's.“We then delete user content and user prompts in accordance with the customer’s retention policy.” source, read | Not found in the pages we 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 | The United States, including the calls out to OpenAI and Google. Thomson Reuters publishes UK, Australian and Canadian hosting as alternatives.“We process all interactions with OpenAI GPT and Google Gemini in the U.S.” source, read | Not found in the pages we read. |
An empty cell is not a reassurance and not a criticism. "Not found in the pages we read." means the Thomson Reuters 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 Thomson Reuters 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 Thomson Reuters 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 Thomson Reuters’s own documentation. Each one is a change somebody can make this afternoon.
- Find out what retention your firm actually agreed to, because Thomson Reuters publishes no default. The answer is in your account, not on the website.“We then delete user content and user prompts in accordance with the customer’s retention policy.” source, read
- Tell staff to delete a matter's content when the matter closes. Thomson Reuters says they can do it themselves, which also means nobody else will.“Users also have the ability to delete their content and prompts in the product at any time.” source, read
Worth knowing
- The contractual bar on training reaches the model providers too, in Thomson Reuters’s own words.“Thomson Reuters GenAI third-party partners, such as OpenAI and Google, are contractually prohibited from using any customer data to train their models.” source, read
- Thomson Reuters says it has turned off the model providers' own abuse monitoring, which is the route by which a human reviewer would otherwise read a prompt.“we have established systemic controls to turn off third parties’ abuse monitoring solutions to ensure there is no human access to Thomson Reuters customers’ data…” source, read
- Outputs are yours, subject to Thomson Reuters' own terms — worth reading before a draft goes into a filing.“Subject to applicable law and Thomson Reuters terms and conditions, users own the outputs generated by CoCounsel Essentials.” source, read
Questions we get asked
Is CoCounsel safe for client data?
Whether CoCounsel 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 CoCounsel Essentials and on CoCounsel Legal, Thomson Reuters says it does not train its models on the matter files you upload, by default.
Every fact on this page is Thomson Reuters'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 Thomson Reuters train its models on what we put into CoCounsel?
On CoCounsel Essentials and on CoCounsel Legal, Thomson Reuters says it does not train its models on the matter files you upload, by default.
On CoCounsel Essentials: No, and Thomson Reuters extends it to the third-party models behind the product as well as its own.
Thomson Reuters's own words: "we do not use your user content and user prompts to train or improve CoCounsel Essentials or any third-party GenAI large-language models…" (read 2026-11-03).
On CoCounsel Legal: No. The CoCounsel Legal page says it in one line and does not say more. Thomson Reuters's own words: "Data remains secure and private, and we never use it to train underlying models." (read 2026-11-03).
How long does Thomson Reuters keep what we put into CoCounsel?
On CoCounsel Essentials: There is no published period. Retention is whatever your firm agreed with a Thomson Reuters client manager, which means somebody at your firm has to know what was agreed.
Thomson Reuters's own words: "Retention of user content and prompts is customized to fit customer needs and set up at the organization account level…" (read 2026-11-03). On CoCounsel Legal: Not found in the pages we read.
Can an administrator read what we put into CoCounsel?
On CoCounsel Essentials: Retention is set for the whole organisation rather than per user, so it is an account-level decision and not a user's.
Thomson Reuters's own words: "We then delete user content and user prompts in accordance with the customer’s retention policy." (read 2026-11-03). On CoCounsel Legal: Not found in the pages we read.
Will Thomson Reuters sign a Business Associate Agreement for CoCounsel?
We didn't find a BAA statement in the pages listed below; ask Thomson Reuters before any patient data goes in.
How do we find out which version of CoCounsel our staff are using?
Thomson Reuters sells this to practices rather than to individuals, so there is no personal account to find.
What decides the answers above is which plan your practice is on and what your own administrator has switched on, both of which are in your subscription and your admin console rather than on anybody's laptop.
Obsidian Ridge's free 30-day AI security assessment covers the tools nobody approved, which is the other half of the picture.
Sources
Thomson Reuters’s own pages, with the date each was read. Nothing on this page comes from anywhere else.
- legal.thomsonreuters.com/en/products/cocounsel-essentials, read
- legal.thomsonreuters.com/en/products/cocounsel-legal, read
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 Lexis+ with Protégé safe for client data?
Legal research · LexisNexis
Is ChatGPT safe for client data?
General assistant · OpenAI
Is Microsoft Copilot safe for client data?
General assistant · Microsoft