Case Studies | SaaS B2B

Reading every contract, in every territory

Automated Contract Analysis

About

A SaaS B2B business working with Cloud Combinator on AWS. The client is anonymised at their request.

Challenge

The client's requirements defined four focus areas, each with measurable acceptance benchmarks tested during user acceptance.

Accurate extraction from complex contracts

Amazon Textract has to reliably extract information from multi-page contract PDFs so that a Bedrock model can populate the required due-diligence fields, with a target of at least 90 percent of fields correctly populated, while gracefully handling Textract's asynchronous job quotas without dropping documents.

Finding the clauses that matter

The system must identify clauses that fall outside the client's parameters, such as termination and rent suspension, and surface the client's defined red flags, targeting at least 85 percent recall of material clauses and 90 percent of known red flags, validated against the legal team's manual review.

Multi-territory and multi-language support

Contracts span the UK, Australia, Spain, and Germany, with territory-specific field sets, and arrive in English, Spanish, and German. For Spanish and German contracts the output must be produced in both the source language and an English translation alongside it.

Secure, throttled, and auditable at scale

Uploads must be scoped to the authenticated user via short-lived pre-signed URLs with no cross-user access, the pipeline must stay within Textract and Bedrock quotas through queuing, and every processing step must be timestamped in DynamoDB for auditability, all while meeting a five-minute-per-contract processing target.

Solution

The work moved through two stages, proving the approach before hardening it into a production service deployed in the client's own AWS environment.

4

Territories supported: UK, Australia, Spain, Germany

90%

Due-diligence field extraction accuracy target

5 min

Per-contract processing SLA target

By the numbers:

  • 4 - Territories supported: UK, Australia, Spain, Germany
  • 90% - Due-diligence field extraction accuracy target
  • 5 min - Per-contract processing SLA target
Changes

Acceptance is defined against measurable benchmarks verified during UAT: at least 90 percent of due-diligence fields correctly populated, at least 85 percent recall of material clauses, at least 90 percent of known red flags surfaced, a per-contract SLA of five minutes and a fifty-contract batch within sixty minutes, ten concurrent users without degradation, and no cross-user document access. The figures below are the engagement's acceptance targets.

  • Grounded, structured AI — Amazon Bedrock with Claude produces defined JSON outputs for due-diligence fields, concerning clauses, red flags, summaries, and manual-check queries, each tied to the extracted contract content rather than free-form text.
  • Quota-safe by design — Throttled SQS queues between Textract and Bedrock keep the pipeline within service limits, so a large batch is processed reliably without dropping documents.
  • Territory-aware — Separate prompts per territory apply the right field set, and Spanish and German contracts are returned in both the source language and English.
  • Secure and auditable — Pre-signed uploads scoped to each user, Cognito login, Bedrock Guardrails, and end-to-end DynamoDB timestamping give both security and a complete audit trail.
  • Handover — The service was deployed into the client's AWS environment via CloudFormation with a walkthrough, documentation, and templates for the team.

With extraction, analysis, and reporting proven across four territories, the client has a pipeline that is no longer a proof of concept but a production base ready to extend to more territories, contract types, and downstream integrations as the portfolio grows.

AWS Stack

Amazon Textract

For reliable text and data extraction from multi-page contract PDFs.

Amazon Bedrock

With Claude for grounded due-diligence checks, clause and red-flag identification, and document summaries.

Amazon Bedrock Guardrails

For AI-misuse checks on incoming documents.

Amazon S3

For document and output storage, with cross-region replication for redundancy.

Amazon DynamoDB

For status tracking and timestamped auditability at every stage.

Amazon SQS

And Amazon SNS for throttled, quota-safe orchestration between services.

YOU MIGHT LIKE

Related success stories

View all case studies

Case Studies | Insights

Utilising Language Recognition, Speed, and Enhanced Security to Make Social Media a Force for Good

  • Here, we take a detailed look at how the Cloud Combinator team collaborated with another cutting-edge AI service provider that provides intelligent systems to “make social media more social” for brands and users alike.
  • Arwen AI is a UK-based startup specialising in AI solutions to manage and enhance brands’ social media interactions. Founded in 2020 by Matt McGrory, Dr. David Cole, and Joel Bailey, Arwen. AI focuses on using AI to automatically detect and remove spam, toxic comments, and other unwanted content from social media platforms.
  • The team at Arwen have three core products. ‘Moderate’ is focused on identifying and removing toxic content from social media channels. ‘Engage’ helps brands identify and engage with meaningful conversations on social media, and ‘Customize’ allows brands to apply bespoke algorithms to their channels - creating an even more effective moderation and engagement.
Read more
CONTACT US

Ready to turn AI into impact?

We'll help you spot the highest-value opportunities, reduce risk around your first AI initiative, and define a clear path to results from day one.

Why talk to us:

Outcome-driven recommendations

AWS-recognised delivery expertise

Risk-aware AI adoption

Clear next step, not a sales pitch

Start with a focused 20-minute conversation about your goals — no pressure, no commitment.

We use cookies

This website uses cookies to enhance user experience and to analyze performance and traffic on our website.

See our Privacy Policy for details.