Case Studies | SaaS B2B

A multi-agent Generative AI platform, built on AWS

Generative AI Platform

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 about turning a broad ambition for generative AI into a concrete, safe, and demonstrable platform.

Delivering four distinct AI capabilities as one platform

Image generation, code generation, internet research with citations, and description writing are very different tasks. The platform needed to handle all four through a coherent, modular design rather than four disconnected tools, so capabilities could be combined and extended over time.

Keeping generative output safe and compliant

Generative features carry real risk around unsafe content, personal data, and off-brand output. Every model interaction needed

Proving AWS as the foundation for production

As a proof of concept, the work had to demonstrate that AWS services were suitable for the continued development of the client's service, and leave the team with the documentation and understanding to take it further.

Solution

The engagement was delivered as a sequence of milestones, each building toward a validated, handed-over platform.

4

Modular AI capabilities: image generation, code, research, and descriptions

100%

Of Amazon Bedrock calls governed by AgentCore guardrails

~$2,160/mo

Indicative AWS run-rate modelled at 2,000 users

By the numbers:

  • 4 - Modular AI capabilities: image generation, code, research, and descriptions
  • 100% - Of Amazon Bedrock calls governed by AgentCore guardrails
  • ~$2,160/mo - Indicative AWS run-rate modelled at 2,000 users
Changes

The proof of concept delivered a working, governed Generative AI platform that produces images, code, research, and descriptions through a supervisor-and-worker agent design, with safety guardrails on every model call, ready to hand over to the client's engineers.

  • Four capabilities, one platform — Image generation to 1024x1024, code snippets with docs and example I/O, cited research, and validated descriptions, all delivered through a single modular agent architecture.
  • A supervisor-and-worker design — A supervisor agent routes each request to the right specialist worker agent, each powered by the best-suited Amazon Bedrock model for the task.
  • Guardrails on every call — AgentCore applies topic filters, block lists, PII protection, and refusal handling to 100% of Bedrock calls, with structured violation payloads and full logging.
  • Serverless and encrypted — API Gateway, Lambda, and Step Functions orchestrate the flow, with S3 under SSE-KMS for outputs and DynamoDB for requests and metadata, all provisioned via CloudFormation.
  • Handover — The client received a detailed runbook, incident-response procedures, a recorded demo and technical walkthrough, and a knowledge-transfer workshop for its engineers.

With the core capabilities and guardrails proven, the client has a validated foundation for its AI features and a clear path to production. The platform was built as a first version, and Cloud Combinator can guide the client on the steps to take it to production grade as engagement grows. Cost figures above are indicative, modelled on projected usage, and subject to final model selection and workload patterns.

AWS Stack

Amazon Bedrock

For text, code, image, and embedding generation across the modular workflows.

Amazon Bedrock AgentCore

For the agent runtime and guardrails applied to every model call.

Amazon API Gateway

And AWS Lambda for the serverless request and handler layer.

AWS Step Functions

For orchestrating multi-step, long-running, and parallel jobs.

Amazon S3

With SSE-KMS for encrypted storage of generated images and exports.

Amazon DynamoDB

For tracking requests and metadata.

AWS CloudFormation

For provisioning the entire environment as infrastructure as code.

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.