Case Studies | SaaS B2B

An AI insight layer for revenue teams

AI Insight Layer MVP

About

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

Challenge

The design work centred on three focus areas that decided whether the MVP was worth building.

Synthesising scattered knowledge

The first focus was the core differentiator: taking a large volume of informal, inconsistent input and turning it into clean, categorised, queryable common knowledge. This is a synthesis problem, not a data-entry one, and it needed an AI step that classifies each note, extracts the underlying reason, scores sentiment and makes the result shared rather than siloed.

Handling PII under GDPR

The second focus was the tension between voice capture and holding no attributable personal data. Free-text and voice notes inevitably contain PII, so the

Proving it affordably

The final focus was cost and speed. As a proof of concept on synthetic data, the build could carry no idle infrastructure and no standing bill, while still being a credible path to production. The architecture had to be serverless end to end and expressible as a single infrastructure-as-code template.

Solution

The product follows a five-stage pipeline, Inputs, Synthesis, Insight, Output and Outcome, which maps almost one to one onto an AWS GenAI pipeline. Cloud Combinator structured delivery into phases that each end in something demonstrable.

5

Stage insight pipeline mapped to AWS services

~70%

Synthesis accuracy target on synthetic data

~90%

Lower vector-store cost using S3 Vectors vs OpenSearch Serverless

By the numbers:

  • 5 - Stage insight pipeline mapped to AWS services
  • ~70% - Synthesis accuracy target on synthetic data
  • ~90% - Lower vector-store cost using S3 Vectors vs OpenSearch Serverless
Changes

Cloud Combinator delivered a complete, defensible solution design: a scoped MVP held to an explicit MoSCoW line, a serverless AWS architecture with a full bill of materials, a data model and API surface, and a costed, fundable build plan ready to become a deployable SAM template. The figures below are design targets and architectural characteristics, not delivered production results.

  • Scoped MVP — A clear MoSCoW definition of what the proof of concept includes and excludes, holding the line against scope creep such as multi-turn interrogation and predictive analytics.
  • Serverless architecture — An end-to-end serverless design with no VPC, no NAT Gateway and no always-on database, dominated only by pay-as-you-go Bedrock token usage at demo scale.
  • Compliance by design — PII flagged and redacted at the synthesis step, no attributable personal data stored, and all processing kept in the London region through an EU inference profile.
  • Cost-aware choices — Amazon S3 Vectors selected over OpenSearch Serverless to avoid a standing minimum of roughly 260 USD per month, keeping the proof of concept close to zero idle cost.
  • Handover — A design ready to become a deployable SAM template, with the bill of materials, IAM model, event sources and per-environment parameters all defined for the build team.

With the architecture proven on paper and sized for AWS GenAI proof-of-concept funding, the client has a clear, low-risk path from design to a working MVP, and from a demonstrable proof of concept to a production insight platform.

AWS Stack

Amazon Bedrock

For the Claude-based synthesis step that cleans, classifies and reasons over each captured note.

Amazon Bedrock Knowledge Bases

For retrieval-augmented common knowledge that makes synthesised insight shared and queryable.

Amazon Transcribe

For turning voice notes into text at the capture stage.

Amazon Comprehend

For detecting and redacting personal data before storage.

AWS Lambda

And Amazon API Gateway for serverless, event-driven compute and a single authenticated entry point.

Amazon DynamoDB

And Amazon S3, with S3 Vectors, for low-cost insight storage and the vector store behind the Knowledge Base.

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.