From recorded calls to recruiter insight

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

The challenge had three focus areas that the proof of concept needed to address.

A recruiter's call audio is uploaded securely to Amazon S3 using a pre-signed URL issued by Amazon API Gateway and an AWS Lambda function, with metadata logged to Amazon DynamoDB. The upload triggers an AWS Step Functions workflow that runs an Amazon Transcribe job to produce a text transcript, tracking job progress and results as it goes.

The proof of concept met its acceptance criteria, demonstrating that recruiter calls can be transcribed and analysed automatically on AWS and that the resulting insights are suitable for measuring team performance, giving the client a validated blueprint for a production rollout.

• Automated end to end — An uploaded call moves through transcription and analysis without manual intervention, from pre-signed upload to stored insight.
• Objective analytics — Amazon Bedrock extracts sentiment and success indicators, and Amazon QuickSight turns them into dashboards for recruiter performance and candidate outcomes.
• Secure and governed — Uploads are encrypted at rest, backed by cross-region replication and a clear lifecycle policy, keeping sensitive audio protected and cost-controlled.
• Serverless and scalable — A Step Functions and Lambda architecture handles concurrent transcription jobs within service quotas and scales with call volume.
• Handover — Cloud Combinator delivered the working pipeline with a demonstration, materials and documentation so the client could evaluate a full production build.

With the pipeline proven, the natural next step (projected) is a production deployment in the client's own AWS account with the full QuickSight dashboard suite and DynamoDB-backed reporting, giving managers ongoing, objective visibility of call quality across the whole recruitment team.