Case study · Legal & procurement
Every contract checked against your playbook
Legal and procurement teams review similar contracts again and again, looking for the same risky clauses. This is how we used AI to extract key terms, compare them with the client's negotiation playbook and flag deviations, while lawyers make every decision.
Project overview
AI-assisted contract review.
- Industry
- Legal & procurement
- Client
- Confidential
- Solution
- Contract review
- Platform
- Azure OpenAI
- Timeline
- 10–14 weeks
- Team
- 3 people
- Documents
- NDAs, MSAs, supplier contracts
- Output
- Clause summary and deviations
- Decision
- Always by a lawyer
Expected business outcomes
What the client set out to achieve
Measured against lawyer-reviewed samples.
-
40–60%
Faster first-pass review
Lawyers start from a clause summary and flagged deviations.
Target range; measured on timed reviews
-
90%+
Clause extraction accuracy
Key clauses checked against a lawyer-labelled sample.
Launch threshold per clause type
-
Searchable
Contract register
Terms such as renewal dates and liability caps searchable across contracts.
For contracts processed by the system
Business challenges
Where the client started
Review is thorough but repetitive.
Volume
Standard contracts queue for lawyer review.
Playbook
Negotiation positions in a document few consult.
Register
Key terms such as renewal dates tracked in spreadsheets, if at all.
Risk
Unfavourable clauses sometimes slip through under time pressure.
Business
Sales and procurement wait for legal sign-off.
Proposed solution
What the build covers
Extraction, comparison and a register, with lawyers in control.
-
Clause extraction
Parties, term, renewal, liability, indemnity and other key clauses extracted.
-
Playbook checks
Clauses compared with preferred and fallback positions.
-
Review workspace
Lawyers see flags alongside the original text.
-
Contract register
Searchable key terms and dates across all contracts.
-
Reminders
Renewal and notice dates generate reminders.
-
Confidentiality
Processing in the client's tenant with access controls.
Project requirements
What the client asked us for
The request was faster review. The design keeps lawyers in charge: the AI highlights, the lawyer decides.
The playbook is turned into clear rules with the legal team before any comparison is built.
What the project had to deliver
- 01Clause summaries
- 02Deviations from playbook
- 03A searchable register
- 04Lawyers decide
Key challenges & considerations
What made it harder, and what we did
-
Contracts on counterparties' paper
What we did
The test set included third-party paper, and accuracy was measured per clause type.
-
Scanned or poorly formatted documents
What we did
Scanned contracts went through OCR with confidence checks, and poor ones were flagged.
-
Playbooks never written down
What we did
The playbook was written down with the lawyers before the first build.
-
Confidentiality requirements
What we did
Documents stayed inside the client's tenant, with access limited by matter.
What we watched for
- Write the playbook down first
- Measure per clause type
- Keep lawyers deciding
- Protect confidentiality
Development process
Playbook first
The rules are agreed before the AI is built.
-
Playbook
Preferred and fallback positions written as rules.
-
Sample
Contracts labelled by lawyers for testing.
-
Extraction and checks
Built and measured against the sample.
-
Workspace and register
Review workspace and register delivered.
-
Rollout
Used on live contracts with ongoing quality checks.
Technology stack
Language models for legal text
Private models with structured outputs and evaluation.
AI
- Azure OpenAI
- Retrieval-augmented generation
- Prompt and output guardrails
Data
- Azure AI Search
- Document pipelines
- Vector index
Application
- Python / .NET services
- Web or Teams interface
Evaluation
- Test question sets
- Quality and safety metrics
- Usage monitoring
Quality assurance & deployment
Measured against a test set before launch
AI quality was measured, not judged from a demo. A test set of real examples with agreed answers decided whether the system was ready, and the same set was re-run after every change.
- A test set of real examples with agreed answers
- Accuracy thresholds agreed with the business before launch
- Safety checks: refusals, permissions and prompt-injection tests
- A person reviewing low-confidence output
- Cost and response time measured at real volumes
Deployment
- 01Workspace and registerReview workspace and register delivered.
- 02RolloutUsed on live contracts with ongoing quality checks.
- 03Pilot, then staged rolloutA small group of real users first, with results reviewed weekly before widening access.
- 04Monitoring and warrantyAccuracy, drift, usage and cost tracked after launch, and six months of free bug fixing.
Final deliverables
Review focused on what matters
Lawyers start each review with a clause summary and flagged deviations, and spend their time on negotiation rather than reading for standard terms.
Every processed contract adds to a searchable register with renewal reminders.
- Faster first-pass review
- Playbook deviations flagged
- Contract register
- Renewal reminders
Handed over to the client
- 01Production AI serviceDeployed in the client's own cloud tenant
- 02Source code and modelsIn a repository the client can access
- 03Test set and resultsThe evidence behind the launch decision
- 04MonitoringAccuracy, usage and cost
- 05Six-month warrantyFree bug fixing after delivery
Success metrics
How success is measured
| Measure | Outcome | How it is measured |
|---|---|---|
| Faster first-pass review | 40–60% | Target range; measured on timed reviews |
| Clause extraction accuracy | 90%+ | Launch threshold per clause type |
| Contract register | Searchable | For contracts processed by the system |
Verified reviews
What clients say about our custom software work
Verified reviews from our clients on the same kind of work, published on Clutch.
-
Custom Software Development
Web-Based Coupon App Development
The work produced by the team was of exceptional quality.
Jon Thies
CTO, Model Rocket Feb 2022 – Mar 2023
-
Custom Software Development
Mobile & Web Platform Development & Design
Silver Scintilla was responsive, respectful, and enjoyable to work with, quickly addressing any questions or requests.
DeShawn Brown
CEO & Founder, Lithios Jun 2021 – Nov 2022
-
Custom Software Development
Web-Based HRM System Development
Silver Scintilla provided the best solution with scalable functionality, personalized features, and robust security.
Suheb Khan
Director, Collaborative Insight Technologies Jul 2020 – Jun 2022
Free discovery call
Have an idea? Let's turn it into AI-powered software.
Book a free discovery call with our experts. Share your idea and we will help you shape the scope, timeline and budget, under NDA.
- Free consultation
- NDA before we talk
- Transparent estimate