Skip to content
DT Nova
Document workflow automation

AI contract review: contracts checked against your playbook before a lawyer reads them

AI contract review reads incoming and outgoing contracts, extracts the clauses that matter to your company, compares each one with your own positions and shows the lawyer or contract manager exactly what deviates, what is missing and what the approved fallback wording would be. We build contract analysis AI for legal teams, procurement, sales and construction companies, with the review decision always made by a person, the key dates written into a contract register, and every step logged. Hosted in the EU, and your company owns the code.

  • You see the result before it goes live
  • Fixed price for the agreed scope, in writing before work starts
  • Source code, domain and accounts in your company's name

Request a quote

Contract review before and after automation

A supplier sends its standard terms with a purchase order. Today a lawyer or buyer reads all twenty pages, compares the liability cap, indemnities, payment terms and termination rights with what the company usually accepts, writes comments in the margin, and types the renewal date into a spreadsheet if someone remembers. A queue forms, and contracts are signed without review when the business cannot wait.

With contract review automation, the contract is read when it arrives. The system identifies the contract type and parties, extracts the relevant clauses and compares each one with your playbook. The reviewer opens a single screen: green for clauses within your positions, amber for acceptable deviations that need a note, red for points that need escalation or a change, and a suggested redline from your approved fallback clauses. The reviewer decides, edits and sends the markup. On signature, term, renewal and notice dates go into the register with reminders.

It is one of the document processes in our document workflow automation work, focused on the part where judgement is needed.

What we build

Contract analysis AI for different teams

Rules that shape contract review software

For contracts with consumers, the EU Unfair Terms Directive provides that a term which has not been individually negotiated is regarded as unfair if, contrary to the requirement of good faith, it causes a significant imbalance in the parties' rights and obligations to the detriment of the consumer, and that a term is always regarded as not individually negotiated where it was drafted in advance and the consumer could not influence it, particularly in a pre-formulated standard contract (Directive 93/13/EEC, Article 3(1) and (2), read on 12 November 2026). A playbook for consumer terms can therefore flag standard clauses that shift risk in this way before they are published.

Contracts contain personal data: names, signatures, contact details and sometimes employee or customer data in annexes. Where we process them for you, the GDPR requires a contract that sets out the subject matter, duration, nature and purpose of the processing, the type of personal data and the obligations of the controller (Regulation (EU) 2016/679, Article 28(3), read on 12 November 2026). We sign one, and the system keeps contracts in your own EU account.

Under the EU AI Act, the high-risk list for the administration of justice concerns AI systems used by a judicial authority, or on its behalf, to research and interpret facts and the law, or in a similar way in alternative dispute resolution (Regulation (EU) 2024/1689, Annex III, point 8(a), read on 12 November 2026). A company reviewing its own commercial contracts is outside that point; the provider and the company using the system must still take measures to support the AI literacy of the staff who use it (Article 4, as replaced by Regulation (EU) 2026/1744, read on 12 November 2026). For the wider policy, see our AI governance guide.

Accuracy and confidentiality

Contract language is varied, and a missed clause can matter more than a wrong one, so we measure both. Before anything is used on live contracts, a set of contracts your lawyers have already reviewed runs through the test version, and you see per clause type how often the clause was found, how often the system's judgement matched your lawyers' and where it missed. Clause types that do not reach the bar you set stay on manual review, and the results are rerun whenever the playbook or the model changes.

Confidentiality is designed in. Contracts stay in your account in an EU region, the models run under terms where your documents are not used for training, access follows teams and matters, and privileged documents can be excluded altogether.

Contract review software or a custom build?

There are established tools for AI contract review and contract lifecycle management, such as Harvey, Luminance, Kira from Litera, Ironclad, Spellbook and the contract features in Microsoft Copilot. For a legal team that wants a standard product and works mainly in Word, one of them may be the right choice, and we will say so. Custom work pays off when the review must follow your own playbook in detail, sit inside your procurement, sales or project systems, or keep contracts in your own EU account.

SituationContract review productCustom contract review
Legal team reviewing in WordUsually the quickest routeRarely worth building
Your own playbook and fallback clausesConfigurable within the productBuilt as your rules
Review inside procurement, CRM or project systemsConnectors of varying depthBuilt into those systems
Register with dates feeding ERP or CRMSeparate repositoryYour systems, one record
Data location and model choiceThe vendor's platformYour account, EU data region

How we build it

  1. Write the playbook. We go through recent contracts with your lawyers and contract owners and write down positions, thresholds and fallback wording for approval.
  2. Scope and fixed price in writing. Contract types, clause types, systems, review steps and what the first phase leaves out, with a fixed price for that agreed scope.
  3. A test version on your real contracts. Already reviewed contracts run through the system, and you see the results per clause type.
  4. Pilot and acceptance. One contract type is reviewed with the system in daily work, and the phase is accepted against the written scope.
  5. Roll out and measure. More contract types, with review time and queue length compared with the numbers recorded at the start.
  6. Maintenance. Playbook updates, model updates, test reruns and new clause types, on a monthly plan you can cancel with one month's notice.

As a standard term, the source code, playbook rules and test sets sit in a repository in your company's own account, and the hosting is in your name in an EU data region. Contract approval and e-signature flows are covered by our document workflow automation page, and model questions by generative AI development. See prices.

Questions we get

What is AI contract review?

Software that reads a contract, finds the clauses that matter to you, such as liability, indemnities, payment terms, termination, governing law or data protection, compares each with your own positions, and marks what deviates, what is missing and what is unusual. A lawyer or contract manager then reviews the flagged points instead of reading every page cold.

Does AI replace the lawyer?

No. The system prepares the review; a qualified person decides. It flags deviations, suggests wording from your approved fallback clauses and explains why a clause was flagged, with the passage highlighted. Accepting a risk, agreeing a change or signing stays with your lawyers and managers.

What is a contract playbook?

Your written positions per clause type: what you accept, what you accept with conditions, what needs escalation and what you never accept, with approved fallback wording. Many companies have it in heads and old emails. Writing it down is the first step of the project and is useful even without software.

How accurate is contract analysis AI?

It depends on your contracts and your playbook, so we measure it on them. A set of contracts your lawyers have already reviewed runs through the test version, and you see per clause type how often the system found the clause and judged it as your lawyers did. Clause types that do not reach your bar stay manual.

Is our contract data used to train AI models?

No. The models run in your account in an EU region under terms where your documents are not used for training, and access to contracts follows the permissions you set per team and matter.

Can it handle construction contracts and other industry forms?

Yes. Standard forms, such as construction or engineering contract families, are useful because the system can compare the special conditions with the base form and show exactly what the other side changed. The playbook then covers your positions on those changes.

Get a written quote

Tell us in a few lines what you need. We reply within one working day with a written quote: scope, timeline and a fixed price or an hour estimate. Asking is free and commits you to nothing.

See every price · Contact details

Request a quote