Case Studies | FinTech

Three AI and data capabilities on one AWS foundation

AI and Data Capabilities Programme

About

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

Challenge

The client identified three focus areas, each a distinct capability with its own success criteria, sitting on a common need for compliance, security and cost efficiency.

Slow onboarding of unstructured product data

Bringing new clients on meant wrestling large, heterogeneous sets of product PDFs into a canonical schema by hand. It was slow, error-prone and did not scale, so onboarding cycle time needed to come down through automated extraction and mapping.

Hard-to-interrogate policy content

Answering questions about policy and understanding the impact of new regulations such as DORA relied on manual reading. The client wanted trusted, permission-aware natural language interrogation of that content, with cited sources and support for gap analysis when new rules arrive.

A fragile calculation workbook

A large, ever-evolving Excel workbook of formulas underpinned a new product testing proposition due to launch in November. It needed to become a configurable, auditable calculation engine that runs at low cost inside the client's own AWS accounts.

Solution

Cloud Combinator delivered the programme as three parallel workstreams, each with early prototypes, explicit quality gates and jointly agreed acceptance metrics signed off before build. All three followed least-privilege IAM, encryption in transit and at rest, and infrastructure-as-code delivery.

3

AI and data workstreams delivered on one AWS foundation

10s-100s

Product PDFs ingested per batch with intelligent mapping

Human-in-

Validation gating every AI output before it is trusted

By the numbers:

  • 3 - AI and data workstreams delivered on one AWS foundation
  • 10s-100s - Product PDFs ingested per batch with intelligent mapping
  • Human-in- - Validation gating every AI output before it is trusted
Changes

The programme gave the client three working capabilities on a single, governed AWS foundation, each meeting jointly agreed acceptance criteria and each leaving a full audit trail behind its outputs.

  • Faster onboarding — Automated extraction and schema mapping, with a bulk review UI, replaced manual data entry and cut the effort of getting heterogeneous product data into the canonical schema.
  • Trusted policy answers — A permission-aware chatbot returns cited answers from the policy library and assists gap analysis when new regulations arrive, without ever making an automated compliance determination.
  • A configurable engine — The product testing workbook became a declarative, containerised engine validated against golden Excel cases, ready to support the November launch and future formula changes without redeployment.
  • Built for due diligence — Least-privilege IAM, encryption, and full run-level and answer-level audit trails give the client the transparency it needs to satisfy its own regulated clients.
  • Handover — Documentation, runbooks and a two-week hypercare and knowledge transfer window left the client's team able to operate and extend all three capabilities.

With the foundation, the review workflows and the audit patterns proven, the client is positioned to widen each capability, from richer extraction and broader policy coverage to additional calculation packs, on infrastructure the team owns and understands. The programme is a base to build on, not an one-off delivery.

AWS Stack

Amazon S3

For secure document landing and reference data storage.

Amazon Textract

For extracting text and tables from product PDFs.

Amazon Bedrock

For prompt-engineered extraction, embeddings and retrieval-augmented answers, with Bedrock Guardrails for safety.

Amazon SQS

For event-driven fan-out of ingestion and extraction work.

AWS Lambda

And Amazon ECS with AWS Fargate for serverless and containerised compute.

AWS Step Functions

For orchestrating multi-step calculation runs with retries.

Amazon RDS

For PostgreSQL and Amazon DynamoDB for structured data, metadata and vector search with pgvector.

Amazon OpenSearch Service

For permission-aware semantic search.

Amazon API Gateway

For run submission, chat and upload endpoints.

Amazon CloudWatch

For logging, metrics and monitoring across all three workstreams.

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.