Case Studies | SaaS B2B

Turning live conversations into structured insight

Theme and Sub-Theme Extraction

About

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

Challenge

The challenge had three focus areas, each of which had to work together in a single pipeline.

Reliable real-time transcription

Live conversations needed to be captured accurately as they happened, with audio streamed to the service and transcribed without manual intervention. If transcription was slow or lossy, everything downstream would inherit the gap, so this had to be dependable and automatic.

Turning unstructured content into usable data

Meetings do not only happen in speech. Users upload documents too, in a range of formats, and those needed to be read and converted into

Consistent AI analysis with an audit trail

The heart of the feature was extracting categories, themes, sub-themes and stakeholders, with supporting evidence, from every input and returning it as a single structured JSON. Users also needed the option to let the model discover categories or to supply their own, and every session had to be logged securely so activity could be audited and troubleshot.

Solution

Each input path is handled through Amazon API Gateway. A client requests a pre-signed URL and uploads a document to Amazon S3, which triggers a processing Lambda that runs the file through Amazon Textract when needed, then passes the extracted text to Amazon Bedrock for analysis, storing the structured output back in S3. For live meetings, the client obtains temporary credentials and streams encoded audio to Amazon Transcribe, receiving the transcript in real time.

4

Input paths unified: live audio, uploaded audio, text and documents

3 min

Analysis refresh cadence during a live meeting

600

Users the pipeline is sized to serve (design target)

By the numbers:

  • 4 - Input paths unified: live audio, uploaded audio, text and documents
  • 3 min - Analysis refresh cadence during a live meeting
  • 600 - Users the pipeline is sized to serve (design target)
Changes

Cloud Combinator delivered the pipeline against the agreed success criteria: real-time transcription with Amazon Transcribe, automatic text extraction with Amazon Textract, structured theme and stakeholder analysis with Amazon Bedrock, and secure session logging in Amazon DocumentDB, all deployed as repeatable CloudFormation stacks.

  • One consistent output — Every input path resolves to a single JSON of categories, themes, sub-themes and stakeholders with supporting evidence, updated as new data is processed.
  • User-controlled analysis — Prompts to Bedrock let users either have categories generated by the model or supply their own, keeping people in control of the workflow.
  • Resilient by design — Amazon DocumentDB was configured with multi-AZ availability and snapshot backups, and S3 storage was replicated to a second region with lifecycle rules moving older files to S3 Glacier.
  • Deployed as code — The environment was stood up through AWS CloudFormation, so it can be rebuilt and evolved consistently as the client moves towards production.
  • Handover — Cloud Combinator provided a reference front end for testing live transcription, along with demonstration calls, Loom walkthroughs and documentation so the client's team could operate and extend the service.

The solution gives the client the start of their AI capability and a well-understood foundation to build on. With the pipeline, the logging model and the deployment templates in place, the client is positioned to harden the service towards production and extend it as their platform grows.

AWS Stack

Amazon Transcribe

For real-time, automatic transcription of streamed meeting audio.

Amazon Textract

For extracting text from uploaded documents so it can be analysed.

Amazon Bedrock

For AI analysis that extracts categories, themes, sub-themes and stakeholders.

Amazon DocumentDB

For secure, auditable logging of every transcription and analysis session.

Amazon S3

With S3 Glacier for durable storage, lifecycle management and cross-region backup.

AWS Lambda

For the serverless functions that orchestrate each processing step.

Amazon API Gateway

For the REST and WebSocket endpoints that connect the client to the pipeline.

AWS CloudFormation

For repeatable, infrastructure-as-code deployment across the environment.

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.