Use cases
Contract review. Against your playbook, every time.
An envoy reads the contract, checks it against your positions and hands back redlines in Word, with tracked changes.
Limitation of liability
Each party’s total liability under this Agreement shall not exceed two times the fees paid in total.
Term and renewal
This Agreement renews for successive three-year terms unless either party gives ninety days’ written notice.
Confidentiality
The obligations in this clause survive for one year after this Agreement ends.
Governing law
This Agreement is governed by the laws of the State of New York.
Notices
Any notice under this Agreement must be in writing and sent to the address set out above.
Every contract, against your playbook. Redlined in Word, ready to send.
In Claren
Ask in plain words. Get redlines back.
The envoy opens the contract on its own computer, checks every clause against your playbook, and hands back the Word file with tracked changes and a note on each.
Welcome to software you own
Claren Apps
01
Ask for it.
An app is a tool an envoy builds for your team: a tracker, a dashboard, a review screen. You describe it in plain words. No template, no project, no vendor.
02
Built on your data.
It sits on your team’s datastore, so it opens with your records already in it, and stays current as envoys file new ones.
03
Yours to change.
Use it beside your work, or in its own window. Ask for a new column, a new view or a new rule, and it changes. It belongs to your team.
What comes back. In the file itself.
- 01Tracked changes, in Word
- 02A comment on every change
- 03Your fallback positions, applied
- 04A summary of what's off-playbook
What AI contract review is
Software that reads a contract, compares it with what you will accept, and tells you what to change.
AI contract review uses a language model to read a contract the way a lawyer would: clause by clause, against a standard. The standard is the important part. A model on its own knows what contracts usually say. It doesn't know what your company will sign.
So good contract review AI works from a playbook: your standard positions, the fallbacks you will accept, and the points that go to someone more senior. Given that, it can do what a reviewer does. It finds the clauses that matter, compares each with your position, and proposes the change.
For an in-house legal team, the value is in the repetition. The same few contract types arrive every week, from sales, procurement, HR and partners. The positions barely change. The time to apply them does. AI for contract review takes the first pass, and often the whole review, off the team.
There are two kinds of output to tell apart. Some AI contract review tools produce a report: a list of risks and suggestions for a lawyer to work through. Others produce the redline: the contract itself, marked up with tracked changes, ready to send. The first helps you do the review. The second does it.
How AI contract review works
Six steps, from the request to the redline.
- The contract arrives. A request comes in from the business, with a Word document or a PDF attached. In most tools a lawyer uploads it. In Claren, an envoy picks it up from the Slack channel it was posted in.
- The contract is read and sorted. The model identifies what kind of contract it is, who the parties are, which side you are on and which clauses carry the risk.
- Each clause is checked against the playbook. Limitation of liability, indemnities, termination, governing law, data protection: each one is compared with your standard position and your fallbacks.
- Changes are drafted. Where a clause falls short, the change is written in your preferred wording, or your first acceptable fallback is applied.
- Each change is explained. A comment sits beside every edit, saying what changed and why, so the counterparty and your own team can follow it.
- The result goes back. A tracked-change redline in Word, and a short summary of what is off-playbook and needs a lawyer's decision. Nothing is sent outside the team without sign-off.
What good AI contract review gets right
Five things separate a review you can send from a review you have to redo.
- It works from your playbook. Not from a generic view of market terms. Your positions, your fallbacks and your red lines, written once and followed on every contract. In Claren, the playbook goes into the envoys' memory, and every envoy on the team follows it.
- It is consistent. The same contract should get the same answer on Monday and on Friday, whoever asked. A team-wide playbook does that. A playbook pasted into one person's chat does not.
- It makes the redlines a lawyer would make. Tracked changes in the Word document itself, in the house style, with fallbacks applied in the right order. Not a list of suggestions to retype.
- It explains every change. A comment on each edit, so a reviewer can check the reasoning in seconds and the other side understands the ask. Claren adds one to every change it makes.
- It knows when to stop. Some points are for a lawyer, not a model: an uncapped indemnity, an unusual governing law, a change to a red line. Good review flags these clearly. Claren lists them in an off-playbook summary, and asks before anything leaves the team.
The contracts in-house teams review with AI
The same handful of types fill most weeks. Each needs its own checks.
| Contract | What gets checked | What comes back |
|---|---|---|
| NDAs | Mutual or one-way, the definition of confidential information, permitted disclosures, term, non-solicitation, governing law. | A redline to your standard, often ready to sign without a lawyer reading it again. |
| Supplier and SaaS agreements | Liability caps, indemnities, service levels, data protection, security terms, auto-renewal and termination for convenience. | Tracked changes with your fallbacks applied, and the points that need procurement or IT flagged. |
| Customer contracts on their paper | Liability, IP ownership, warranties, payment terms, audit rights and anything that departs from your own terms. | A redline sales can send back, and a summary of what to escalate before the deal closes. |
| Data processing agreements | Roles, sub-processors, international transfers, breach notice periods, audit and deletion. | Changes to your DPA position, with each one explained for the privacy lead. |
| Order forms and statements of work | Consistency with the master agreement, scope, fees, acceptance and term. | Conflicts with the master agreement called out, and the order form corrected. |
| Renewals and amendments | What has changed since the last version, notice periods and new obligations. | The differences that matter, and the dates filed where the team tracks them. |
AI contract review tools compared
Three ways to put AI on a contract: a general assistant, a Word add-in, or an AI employee.
Reviews against your playbook
- Claren
- ✓Yes
- A general assistant
- If someone pastes it in
- A Word add-in
- ✓Yes
Tracked-change redlines in Word
- Claren
- ✓Yes
- A general assistant
- Copied in by hand
- A Word add-in
- ✓Yes
A comment explaining every change
- Claren
- ✓Yes
- A general assistant
- If asked
- A Word add-in
- Depends on the product
Fallbacks applied in order
- Claren
- ✓Yes
- A general assistant
- If asked
- A Word add-in
- If set up in the playbook
A summary of what's off-playbook
- Claren
- ✓Yes
- A general assistant
- If asked
- A Word add-in
- Depends on the product
Who starts the review
- Claren
- The envoy, when a contract lands in Slack
- A general assistant
- A lawyer, in a chat
- A Word add-in
- A lawyer, in the document
One playbook for the whole team
- Claren
- ✓Yes
- A general assistant
- One per person's chat
- A Word add-in
- ✓Yes
Every reviewed contract filed with its terms
- Claren
- Datastores, kept by envoys
- A general assistant
- —No
- A Word add-in
- Not in the add-in
Work beyond contract review
- Claren
- Diligence, obligations and the team's other requests
- A general assistant
- Anything, unchecked
- A Word add-in
- Drafting in Word
| Claren | A general assistant | A Word add-in | |
|---|---|---|---|
| Reviews against your playbook | ✓Yes | If someone pastes it in | ✓Yes |
| Tracked-change redlines in Word | ✓Yes | Copied in by hand | ✓Yes |
| A comment explaining every change | ✓Yes | If asked | Depends on the product |
| Fallbacks applied in order | ✓Yes | If asked | If set up in the playbook |
| A summary of what's off-playbook | ✓Yes | If asked | Depends on the product |
| Who starts the review | The envoy, when a contract lands in Slack | A lawyer, in a chat | A lawyer, in the document |
| One playbook for the whole team | ✓Yes | One per person's chat | ✓Yes |
| Every reviewed contract filed with its terms | Datastores, kept by envoys | —No | Not in the add-in |
| Work beyond contract review | Diligence, obligations and the team's other requests | Anything, unchecked | Drafting in Word |
— Not part of it.
Assistant, add-in or AI employee
Each is suited to a different shape of work.
A general assistant such as ChatGPT or Claude is good for a quick question about a clause, or a first read of an unfamiliar contract. It is not tailored to legal workflows. It has no team playbook unless someone pastes one in, no record of what was agreed last time, and its edits have to be carried into Word by hand. Best for the lawyer who wants a second opinion on one document, and has time to check it.
A Word add-in such as Spellbook or Microsoft's Legal Agent in Word sits beside the document and suggests redlines, against a playbook if you have set one up. It is suited to drafting help for the lawyer already in the file. It still waits for that lawyer to open the contract, work through the suggestions and send the result. Best for individual lawyers who want help inside Word, and nothing more.
An AI employee works the other way round. Claren gives an in-house legal team envoys: AI employees, each with its own computer and memory. An envoy takes the contract from the app or straight from Slack, reviews it against the team's playbook, makes the tracked changes in Word with a comment on each, and hands back the redline and an off-playbook summary. It files the contract and its terms in the team's datastore as it goes. The lawyer's job becomes the decision, not the markup.
That is the difference between a tool that helps a lawyer review a contract and a teammate that reviews it.
Where AI contract review goes wrong
Three failures, and how Claren is built to avoid each one.
Generic answers. A model with no playbook falls back on what contracts usually say. It will call a liability cap reasonable that your board would never accept, and miss the clause your team always strikes. The fix is not a better prompt. It is your positions, held where every review can reach them. In Claren, the playbook lives in the envoys' memory, and every contract is checked against it, clause by clause.
Invented authority. General models can produce case names, statutes and market statistics that do not exist. Courts in New York and in London have dealt with lawyers who filed citations of exactly this kind, and regulators on both sides of the Atlantic expect lawyers to check AI output before relying on it. Contract review should rest on your playbook and the document in front of it, not on authority the model supplies. Claren's envoys explain each change by reference to your position, in a comment you can check against the clause.
No record. A review done in a chat window disappears into one person's history. Nobody can see what was conceded, which fallback was used or when the contract renews. Claren's envoys file every contract they review, with its key terms, in the team's datastore, and every user action is kept in an audit trail. The next review starts from what the team already agreed.
How to choose AI contract review software
Seven questions for an in-house team. The answers sort the options quickly.
- Does it do the review, or help you do it? A list of suggestions still needs a lawyer to make every change. A redline you can send is the work done. Ask to see the output on one of your own contracts.
- Does it use your playbook, for everyone? One playbook, set once and followed by the whole team, or a profile each lawyer keeps for themselves. Only the first gives the business the same answer whoever it asks.
- Is the output in Word, with tracked changes? Counterparties expect a marked-up Word document. Anything else means retyping, and retyping means mistakes.
- Can you see why each change was made? A comment on every edit is what lets a lawyer check a review in minutes rather than redo it.
- How do contracts reach it? If requests come from the business in Slack, the software should pick them up there. A tool that waits for a lawyer to upload each file has only moved the queue.
- Where do the results go? Reviewed contracts should end up somewhere the team can find them, with their dates and terms, not in a chat history.
- Is it safe for your contracts? Look for SOC 2 Type II and ISO/IEC 27001, a choice of data region, no training on your data, and each organisation's data kept apart. Claren meets all four.
Free AI contract review tools, and their limits
A free tool will read a contract. It won't review it the way your team does.
The free AI contract review most people try first is a general assistant: upload the contract to the free tier of ChatGPT, Claude or Gemini and ask what is wrong with it. You will get a fluent answer. For a first read of an unfamiliar document, that has its place.
The limits show up quickly. A free tool doesn't know your positions, so its view of what is acceptable is the market's, not yours. It returns prose, not tracked changes, so someone carries every edit into Word by hand. It keeps no record the team can use. And consumer plans are not built for confidential material: depending on the provider and your settings, what you upload may be used to improve the model, and putting privileged documents into a public tool risks privilege.
Free online contract checkers and browser extensions have the same shape, with a narrower model behind them. They suit a quick look at a low-value contract. They are not suited to a legal team's standard contracts, which need the team's own playbook applied the same way every time.
Claren has a free plan too, for one person: a private envoy with its own computer, scheduled routines and app connections. It is the simplest way to see AI contract review done against your own positions, on your own contracts, before bringing in the team.
Getting started with Claren
From your first contract to the team's whole queue.
- Pick the contract type that fills your week. For most in-house teams it is NDAs or supplier agreements. Start there, where the positions are settled and the volume is highest.
- Give the envoy your playbook. Paste it in, upload it, or let the envoy draft one from a few contracts you have already negotiated. It goes into memory once.
- Review a few contracts you have already done. Compare the envoy's redlines with your own. Correct it where it differs, and the correction goes into its notes.
- Link a Slack channel. Point a routine at the channel where the business sends contracts, and the envoy picks up each one as it lands, then posts the redline back in the thread with your sign-off.
- Build the record. Every reviewed contract is filed in the team's datastore. Ask an envoy for a tracker or a review screen on top, and you have your own view of every contract in flight.
- Add the next contract type. Customer paper, DPAs, order forms. Each one is another playbook section, and more work off the team.
Read next
See Claren in action.
Contract review questions
Does it use our playbook?
Yes. Your positions go into the envoy's memory once, and every contract is checked against them.
What format are the redlines in?
Tracked changes in the Word document, as a lawyer would make them.
Will we know why each change was made?
Yes. Every change comes with a comment, and a summary lists what's off-playbook.
Is privileged material safe?
Everything is treated with privilege in mind. Nothing crosses between matters or organisations, and every user action is kept in an audit trail.
Is our data used to train AI models?
No. Neither Claren nor our model providers train AI models on your data.
What is AI contract review?
Software that reads a contract, compares each clause with your team's positions and proposes the changes. The best AI contract review returns the contract itself, redlined with tracked changes, rather than a list of risks for a lawyer to work through.
What is the best AI contract review software for in-house legal teams?
The one that does the review, not just helps with it: your playbook applied to every contract, tracked-change redlines in Word, a reason for every change, intake from where the business already asks, and a record of every contract reviewed. Claren is built for exactly that, as AI employees for in-house teams.
Is there free AI contract review?
Yes. General assistants will read a contract on their free tiers, though without your playbook, without tracked changes and on terms not built for confidential material. Claren has a free plan for one person, with a private envoy that reviews against your own positions.
Can ChatGPT review contracts?
It can read one and comment on it. It doesn't know your playbook unless you paste it in, it doesn't make tracked changes in Word, and it keeps no record for the team. It suits a first read. For a team's standard contracts, an envoy that reviews against the team's playbook is the better fit.
How accurate is AI contract review?
As accurate as the standard it reviews against. With a clear playbook, AI applies the same positions to every contract, which is more consistent than a busy team can be. Claren explains every change in a comment and lists what's off-playbook, so a lawyer can check the review in minutes.
Does AI contract review replace a lawyer?
No. It replaces the markup. The lawyer sets the playbook and makes the calls that need judgement. Claren's envoys do the rest, flag what needs a decision and ask before anything leaves the team.
What contracts can AI review?
Any contract you have positions for. In-house teams usually start with NDAs, supplier and SaaS agreements, customer contracts on the other side's paper, DPAs, order forms and renewals. Claren's envoys read Word documents and PDFs.
How long does it take to set up AI contract review?
With Claren, as long as it takes to share your playbook. If you don't have one written down, an envoy can draft it from contracts you have already negotiated. Link a Slack channel, and contracts are picked up from there.
Related




