Case Studies | FinTech

Screening documents in any language

AI Document Screening

About

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

Challenge

The challenge had four focus areas.

Reading documents in any language

Verification documents arrive in a wide range of languages. The service had to detect the language of each utility bill and provide a full British English translation, highlighting which parts had and had not been translated.

Extracting the facts that decide a check

Beyond translation, the pipeline had to judge whether a document even looks like an utility bill, match the name against the client, confirm the issue date is within the last six months, identify the issuing entity, compare the country on the bill with the client's country and return the address.

Safety against misuse

Because the pipeline exposes a large language model to uploaded files, every document is first put through an AI-misuse check designed to catch attempts to hijack or manipulate the model, returning a clear rationale when a document is rejected.

Security, throughput and data residency

Documents had to be uploaded securely, processed within Amazon Bedrock's service limits and kept in-region. The solution was deployed in eu-central-1 (the client's site) with queueing to respect model rate limits.

Solution

An user requests a pre-signed URL through Amazon API Gateway and an AWS Lambda function, then uploads the document to a dedicated folder in Amazon S3. That upload triggers a Lambda function which places the event on an Amazon SQS queue, throttled so the pipeline never exceeds Amazon Bedrock's rate limit of one hundred calls per minute, and records the document, uploader and timestamps to CloudWatch.

5,000

Documents per month in the sizing basis

200

Files uploaded at once in throughput testing

Eu-central-1

Deployed in-region for data residency (the client's site)

By the numbers:

  • 5,000 - Documents per month in the sizing basis
  • 200 - Files uploaded at once in throughput testing
  • Eu-central-1 - Deployed in-region for data residency (the client's site)
Changes

The engagement delivered against its success criteria: documents can be uploaded securely to AWS, and a LLM on Amazon Bedrock screens each one for misuse, translates it from a wide range of languages and extracts the fields that support an identity check, all deployed via AWS CloudFormation in the client's the client's site environment.

  • Multilingual by default — The pipeline detects the language of each bill and returns a full British English translation, flagging any untranslated components.
  • Decision-ready extraction — Name match, six-month recency, issuing entity, country match and address are returned as structured JSON a checker can act on.
  • Safe by design — An AI-misuse check screens every document before analysis, and rejects are returned with a clear rationale.
  • Built to respect limits — Amazon SQS throttles the flow to stay within Bedrock's one-hundred-calls-per-minute limit, keeping the service reliable under bursts.
  • Handover — The client received the CloudFormation templates, documentation, cost-tagging guidance and a demonstration of the service.

With automated, multilingual document screening in place, the client can extend the same pipeline to further document types and checks as its verification needs grow.

AWS Stack

Amazon S3

For secure document upload via pre-signed URLs and results storage.

Amazon API Gateway

And AWS Lambda for the upload flow and the model-call pipeline.

Amazon SQS

For throttling calls to stay within Bedrock rate limits.

Amazon DynamoDB

For pipeline metadata.

Amazon VPC

For network isolation, with Amazon CloudWatch for logging and monitoring.

AWS CloudFormation

For region-specific, infrastructure-as-code deployment.

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.