Case Studies | SaaS B2B

From two clouds to one

Analysis Engine Migration

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 essential to a migration that changed the cloud without changing the product.

Preserve load-bearing engine behaviour

Six behaviours defined success: event-triggered orchestration with duplicate suppression, the 21-activity DAG, the in-process eight-way LLM fan-out, four layered retry budgets including a 24-hour external webhook budget, deterministic webhook event IDs, and replay-safe orchestrator logging. All six had to survive the move.

Migrate two regions with no flag day

The engine runs in the UK and Canada, and customer documents must never leave their region of upload. Cutover had to be reversible at each step and safe for a DORA-regulated insurance customer, with the cross-cloud call to Vertex AI Gemini preserved exactly as before.

Prove performance for SLAs

The client identified a production load test as a prerequisite for committing to enterprise SLAs, so the migrated engine had to be profiled for throughput, latency percentiles, and failure behaviour before those commitments could be made.

Solution

The work was organised into four workstreams, taking the engine from a validated build to a decommissioned Azure estate and an operable, load-tested AWS engine.

94

Azure resources retired with feature parity preserved on AWS

21

Orchestrated activities ported to Step Functions and Lambda

2 regions

Single-cloud production estate in the UK and Canada

By the numbers:

  • 94 - Azure resources retired with feature parity preserved on AWS
  • 21 - Orchestrated activities ported to Step Functions and Lambda
  • 2 regions - Single-cloud production estate in the UK and Canada
Changes

The migration consolidates the client onto a single-cloud AWS estate in both regions while preserving the engine's behaviour, with acceptance measured against output equivalence to the Azure engine and per-behaviour validation of all six load-bearing behaviours.

  • Behaviour preserved — The 21-activity DAG, eight-way LLM fan-out, layered retry budgets, deterministic event IDs, and replay-safe logging all carry over, validated against the Azure output on a sampled comparison set.
  • No flag day — The shadow, dual-write, and canary sequence keeps rollback a configuration change at the API tier, with Canada following the UK only after two stable weeks.
  • SLA-ready — A reusable load-test harness and a SLA recommendation, anchored to real throughput, latency, and failure-mode data, give the client the evidence base for enterprise SLA commitments.
  • Lean run-rate — The AWS engine runs at roughly $645 per month across both regions, with per-token LLM spend staying on the existing Vertex AI bill; the client's projects an Azure saving of $3,000 to $6,000 per region per month, to be confirmed against invoices at cutover.
  • Handover — Runbooks, CloudWatch alarms routed to Slack, and a quota dashboard are handed to the client's on-call team with a recorded knowledge-transfer session.

With the engine consolidated onto AWS, the client is positioned for the natural next step already identified as a separate future engagement: modernising the AI layer from Vertex AI Gemini onto Amazon Bedrock.

AWS Stack

AWS Step Functions Standard

For deterministic orchestration of the 21-activity analysis pipeline.

AWS Lambda

For the ported activities, including the in-process eight-way LLM fan-out.

Amazon Textract

For document OCR, replacing Azure Document Intelligence.

Amazon EventBridge Scheduler

For the 24-hour external webhook retry budget.

Amazon S3

For in-region storage of customer documents and intermediate artefacts.

Amazon CloudWatch

And Amazon SNS for alarms and quota dashboards routed to Slack.

Amazon VPC

With NAT Gateway and AWS KMS for regional networking, cross-cloud egress, and encryption.

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.