Case Studies | FinTech

Turning compliance calls into automated audit trails

PoC - Audio Transcription and Compliance Check

About

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

Challenge

The challenge had three focus areas.

Secure, high-volume ingestion

Sensitive audio needed to reach AWS securely and at scale, with every file tracked. The design had to accept files from the client's existing tooling through Zapier, generate short-lived pre-signed upload URLs, and log metadata for every file so that nothing could be lost or processed without an audit trail.

Accurate transcription with cost headroom

Transcription is the foundation of the whole pipeline, so it had to be accurate, but it also drives the majority of running cost at volume. The

Regulator-aware compliance reporting

A raw transcript is not enough. Each call needed to be assessed against FCA frameworks and returned as a structured report covering call summary, GDPR consent, script adherence and sentiment. This required a language model grounded in the relevant regulatory documentation rather than answering from general knowledge.

Solution

An administrator first loads the relevant FCA and policy documentation into an Amazon S3 bucket that acts as the data source for an Amazon Bedrock knowledge base. To submit a call, an user makes a POST request to receive a pre-signed URL, then uploads the audio file directly to Amazon S3 with its metadata attached. That upload event triggers the first AWS Step Functions workflow, which runs transcription through either Amazon Transcribe or a SageMaker-hosted model and writes the transcript back to S3.

20,000

Recorded calls per month the pipeline was designed to process (target)

~1,000

Files per day at target production scale

~66%

Lower projected monthly run cost using the SageMaker path versus managed transcription

By the numbers:

  • 20,000 - Recorded calls per month the pipeline was designed to process (target)
  • ~1,000 - Files per day at target production scale
  • ~66% - Lower projected monthly run cost using the SageMaker path versus managed transcription
Changes

The proof of concept demonstrated a complete, secure path from uploaded audio to a regulator-aware compliance report, meeting the success criteria set at the start of the engagement. The pipeline was built to be deployed as infrastructure as code with AWS CloudFormation, giving the client a repeatable foundation rather than an one-off script.

  • Secure ingestion proven — Pre-signed URL uploads to encrypted S3, with per-file metadata and status logged to DynamoDB for a complete audit trail.
  • Choice of transcription engine — Both Amazon Transcribe and a self-hosted SageMaker model were designed in, letting the client trade managed simplicity against a materially lower cost profile at volume.
  • Grounded compliance reports — A Bedrock knowledge base kept the language model anchored to FCA documentation, returning structured JSON reports for summary, consent, script adherence and sentiment.
  • Built for scale and resilience — Event-driven Step Functions, S3 intelligent tiering with a six-month archive policy and cross-region replication were designed to carry the service to production volumes.
  • Handover — Cloud Combinator provided a documented GitHub repository, Loom walkthroughs and knowledge-transfer sessions so the client's team could run and extend the solution.

With the pipeline proven, the client has a clear route to move from proof of concept to a production compliance service, deployed in their own AWS account and ready to scale as call volumes grow.

AWS Stack

Amazon S3

For secure, encrypted audio storage with intelligent tiering and cross-region replication.

Amazon Transcribe

And Amazon SageMaker for the two evaluated speech-to-text routes.

Amazon Bedrock

For a knowledge-base-grounded language model that produces the compliance reports.

AWS Step Functions

For orchestrating the transcription and compliance workflows.

AWS Lambda

For pre-signed URL generation and pipeline glue logic.

Amazon DynamoDB

For per-file metadata, status logging and report storage.

Amazon API Gateway

For the upload and retrieval endpoints, with AWS CloudFormation deploying the stack.

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.