Case Studies | FinTech

Automating security due diligence with generative AI

A FinTech client

About

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

Challenge

The challenge had four focus areas, each of which shaped the design of the solution.

Accurate document processing

The system had to extract information from mixed source documents, including PDFs and images, then answer each questionnaire item precisely against the evidence provided rather than in the abstract.

Grading with reasoning

Responses had to be scored automatically against the criteria in the spreadsheet, with a short explanation of why each score was given, and the results written back into the file ready for stakeholder review.

Flexibility and correction

The client's team needed to add or remove sections and questions without breaking the workflow, and to improve any single weak answer through prompt adjustment without disturbing the others.

Maintainable and compliant by design

The whole solution had to be serverless where possible, built in Python with Terraform infrastructure, fully cost-tagged, and compliant with the client's regulatory requirements.

Solution

Source documents land in Amazon S3, where AWS Lambda functions use Amazon Textract to extract text and data from PDFs and images. That extracted content is indexed into Amazon OpenSearch Service, giving the system a vector store for fast, relevant retrieval.

2 hrs

Target turnaround for a standard due diligence questionnaire

$6.06

Estimated AI processing cost per document (Bedrock and agents)

3 weeks

Build duration from kick-off to demonstrated solution

By the numbers:

  • 2 hrs - Target turnaround for a standard due diligence questionnaire
  • $6.06 - Estimated AI processing cost per document (Bedrock and agents)
  • 3 weeks - Build duration from kick-off to demonstrated solution
Changes

The solution was delivered against the agreed success criteria and demonstrated to the client, automating the SDD questionnaire process end to end while keeping a human review step at the point of sign-off. It was benchmarked to return results for a standard questionnaire within two hours, with every resource cost-tagged for transparent reporting.

  • Accurate extraction — Amazon Textract plus a RAG pipeline turns mixed PDFs and images into evidence-grounded answers.
  • Automated grading — Each response is scored against the spreadsheet criteria with reasoning, then written back into the file automatically.
  • Correctable and flexible — Individual answers can be improved through prompt adjustment, and sections or questions can be added or removed without breaking the parser.
  • Maintainable infrastructure — A serverless, Python and Terraform build deployed into the client's own AWS account, fully cost-tagged for reporting.
  • Handover — Immersion day workshops upskilled the client's team on Amazon Bedrock and the wider solution.

This initial implementation is Level 1 of a planned roadmap. The agreed next steps move towards a human-in-the-loop review interface, model fine-tuning on that feedback, and ultimately a more fully automated onboarding journey. Estimated running cost at this stage is around $1,700 per month, a projection that depends on actual usage.

AWS Stack

Amazon Bedrock

For the large language model and AI agents that answer and grade each question.

Amazon Textract

For extracting text and data from PDFs and images.

Amazon OpenSearch Service

For vector indexing and fast similarity retrieval.

AWS Lambda

For serverless document processing and spreadsheet population in Python.

Amazon S3

For secure storage of raw documents and processed outputs.

AWS Identity

And Access Management for least-privilege access across the solution.

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.