A cloud migration is one of the few IT projects every part of the business feels. Done well, nobody notices the move, only the faster systems and lower bills afterwards. Done poorly, it means outages, surprise invoices and security gaps. This checklist captures what we cover on every Azure migration.
Be specific about why you are moving: retiring a data centre lease, improving resilience, enabling AI workloads or reducing cost. Goals decide trade-offs later.
Catalogue every server, database, application and integration. Tools such as Azure Migrate can collect performance data so you size resources on real usage rather than guesswork.
Find which applications talk to each other, which share databases and which rely on hard-coded IP addresses. Unmapped dependencies cause most migration-day failures.
For each system decide whether to rehost (lift and shift), replatform (move to managed services such as Azure SQL), refactor, replace with SaaS or retire. Not everything deserves the same treatment.
Set up the foundations first: subscriptions and management groups, networking, identity with Microsoft Entra ID, policies, logging and backup. A good landing zone means every workload arrives into a secure, consistent environment.
Apply least-privilege access, multi-factor authentication, encryption, network segmentation and Microsoft Defender for Cloud. Map controls to any frameworks you must meet, such as HIPAA, SOC 2 or PCI DSS.
Turn on budgets, alerts and a tagging standard (owner, environment, cost centre) before resources are created. Retrofitting tags later is painful.
Group workloads into waves, starting with lower-risk systems to build confidence. For each wave, write a cutover runbook and a rollback plan with clear go and no-go criteria.
Use continuous replication so the target stays in sync with the source. Run functional, performance and failover tests in Azure before any users move.
Schedule cutover during low-traffic windows, lower DNS TTLs in advance, freeze changes, perform the final sync and switch traffic. With good replication, downtime can often be measured in minutes or avoided entirely for stateless services.
After a few weeks of real usage, resize over-provisioned resources, use reserved instances or savings plans for steady workloads and shut down non-production environments outside working hours.
Set up monitoring, patching and backup as an ongoing service, then look for modernization opportunities such as containers, serverless and data platforms that make AI projects easier to deliver.
Planning a move? Our cloud migration services cover every step on this list, and our managed support team can run the environment afterwards.
Small environments can move in a few weeks. Mid-size estates with dozens of applications typically take three to six months when migrated in planned waves.
Many workloads can be moved with little or no user-facing downtime using continuous replication, careful DNS planning and off-peak cutovers. Some legacy systems still need a short maintenance window.
It can, but only with right-sizing, reserved capacity, licensing benefits and good cost governance. Lift-and-shift without optimization often costs more than staying on-premises.
A practical AI readiness roadmap for mid-size businesses: pick the right use cases, fix data gaps, prove ROI in a short pilot, and scale AI into production safely.
Read articleAgentic AI explained in plain language: how AI agents differ from chatbots, where they create value today, the risks to manage, and how to launch your first agent safely.
Read articleHow to roll out Microsoft 365 Copilot successfully: licensing, data permissions and oversharing clean-up, pilot groups, user training, and how to measure adoption and ROI.
Read articleBook a free 30-minute consultation and we'll map your quickest wins in AI, cloud and Microsoft 365.