Skip to content
Critical Cloud Critical Cloud
Book a call
Case study · Software and SaaS · Critical Support

24x7 cover live in five weeks for an AI fraud-prevention platform

A Dublin company using AI to stop online fraud went from kick-off to 24x7 Critical Support in five weeks.

Talk to us All case studies
At a glance
5 weeks
from kick-off to live cover
24x7
on-call for production alerts
Every
AWS account regularly scanned
AI and fraud preventionIrelandAWSServerlessDatadogManaged service
01
The challenge

The challenge

Small product teams on serverless AWS can ship quickly, but production monitoring, on-call and runbooks often wait until the big customers arrive, and Datadog log costs climb fast once full logging is switched on. This company, which uses AI to stop online payment fraud, account takeover and bot traffic, chose to put a specialist SRE function in place as it prepared to scale. It asked Critical Cloud to lead on Datadog best practice from a fresh start.

02
What we did

What we did

  • Ran a structured onboarding: an architecture deep dive, a shared Slack channel and service desk, an agreed out-of-hours escalation route and a 24x7 on-call rota.
  • Ran a Datadog HealthScan and turned the results into a prioritised improvement backlog.
  • Put every AWS account under regular automated cloud scanning, deployed through the team's own infrastructure as code.
  • Built Datadog monitoring as code across the real-time event pipeline, with a runbook linked from each alert.
  • Cut low-value network log volume in Datadog while keeping warnings and errors.
03
The outcome

The outcome

  • 24x7 Critical Support live in October 2026, five weeks after kick-off, with alerting tuned alongside the team so production gets 24x7 response and test environments are handled in the working day.
  • Production alerts reach a 24x7 on-call team, each pointing to the affected account and its runbook.
  • A clear improvement backlog, refreshed by regular scanning across every account.
  • Leaner Datadog logging, managed in code so it can be adjusted as the business grows.
Critical Support →

This case study is anonymised, so the customer is not named.

Talk to us about Critical Support.

Tell us what you run today and what you need from it. We’ll recommend the simplest first step.

Talk to us About Critical Support