Azure is easy to enter one portal change at a time.
A resource gets created for a deadline, a network rule is widened to unblock a deployment, and a second subscription copies the first by hand. The estate runs, but nobody can explain which choices are deliberate or reproduce them safely.
Azure consulting earns its place when it closes a specific decision and leaves an operating system behind: architecture, infrastructure code, validation evidence, cost ownership, and a team that knows what to change next.
Planning anchors
- £5K
- starting point for a focused Azure engagement
- Architecture, cost, or migration-planning boundary
- 4-12 weeks
- common window for a bounded first engagement
- Final timing follows workload and access
- 1 workload
- recommended first migration or architecture boundary
- Expand after the path is validated
These are RaftLabs planning figures, not Azure benchmarks or promised savings. RaftLabs does not currently publish an Azure-specific case study with independently audited cost or migration outcomes. We therefore use this page to state scope, price, and evidence boundaries rather than borrow results from unrelated cloud work.
Azure consulting fits when Microsoft-specific choices materially affect the workload.
Choose general migration planning when the target provider is still open, or embedded DevOps capacity when the backlog and architecture are already owned.
A fit01Microsoft identity, SQL Server, Azure services, or licensing are real workload dependencies.
02A team can provide account access, workload owners, usage data, and a decision-maker.
03The first engagement can be bounded to an assessment, architecture, cost issue, or workload.
Not a fit01The cloud provider has not been chosen and the workload needs a neutral assessment.
02The request is ongoing ticket capacity without a defined consulting outcome.
03The buyer expects guaranteed savings, compliance certification, or a risk-free cutover.
Scope
What the first Azure engagement can cover
01Architecture and landing boundary
Map subscription structure, identity, networking, compute, data, environments,
deployment, monitoring, and ownership for the first workload. Record
alternatives and the reasons behind service choices before provisioning.
02Workload migration planning
Inventory dependencies, choose rehost, replatform, refactor, retain, or
retire, and define data movement, cutover, validation, and rollback.
Implementation begins only after owners accept the path and interruption
window.
03Identity, access, and delivery controls
Connect Entra ID and application access where required, separate environments,
define least-privilege roles, and place infrastructure and deployments behind
reviewable code rather than personal portal knowledge.
04Cost and operating review
Find resources without owners, sizing assumptions, reservation candidates,
missing budgets, and services whose operating burden exceeds their value.
Recommendations state dependencies and expected trade-offs; they do not
promise a savings percentage.
Provider-specific advice vs provider-neutral migration
| Azure consulting | Cloud migration services |
|---|
| Starting question | How should this workload run on Azure? | Where and how should this estate move? |
| Provider | Azure is a justified target | Provider may still be evaluated |
| Typical boundary | Architecture, cost, identity, or one workload | Application, data, cutover, and operating portfolio |
| Primary proof | Reviewable Azure decisions and implemented controls | Rehearsed movement, validation, rollback, and reconciliation |
| Starting price | Focused engagement from £5K | Focused first workload from $25K |
Azure consulting can feed a migration, but the pages should not make the same promise. This service is for a Microsoft-specific decision. Cloud migration is for moving and validating an existing workload, while cloud application development is for a new product designed around cloud services.
Delivery
From Azure question to an operable first change
Four steps keep architecture, implementation, and ownership connected.
- Step 1
01Define the Azure decision
Name the workload, business deadline, current estate, Microsoft dependencies,
and decision the engagement must support. Agree client owners, access, known
constraints, and what is outside the first boundary.
- Step 2
02Inspect architecture and access
Review subscriptions, identity, networking, data, integrations, cost,
deployment, and operational ownership at the agreed boundary. Confirm where
documentation and the live estate disagree.
- Step 3
03Build or change one boundary
Implement the approved architecture or migration through code, reviews,
validation, and a written rollback path. Keep environment differences explicit
and avoid unmanaged portal changes.
- Step 4
04Validate and hand over
Test the workload and controls, reconcile cost assumptions, and leave
diagrams, code, decisions, and runbooks with the operating team. Record any
residual risk and the next sensible boundary.
- Identity is treated as a late integration
- Entra ID, service identities, external users, and conditional access can shape the architecture. Include their owners and test paths early.
- The live estate differs from its diagram
- Inventory actual resources, traffic, dependencies, and configuration before committing to a migration or deletion plan.
- Platform certification is confused with customer compliance
- Translate agreed requirements into technical controls and evidence, while leaving legal interpretation and audit opinions to the client's qualified advisers.
- A cost target has no workload assumption
- Forecast against current and expected usage, service tiers, licensing, retention, and support. Cloud spend changes when those assumptions change.
Scope and price
A focused Azure engagement starts at £5,000.
Begin with one architecture, cost, or migration decision and expand only after the first boundary is understood.
Azure consumption, Microsoft licences, third-party tools, and audit services are separate unless the written proposal includes them.
Starting investment
Starts at £5K
A bounded first engagement commonly takes 4 to 12 weeks. Full workload delivery is priced after identity, data, integration, validation, and rollback are reviewed.
Decisions before provisioning
The first scope records alternatives, chosen services, assumptions, client
dependencies, and approval owners before the environment changes.
No compliance promise
Technical controls and evidence can be delivered, but RaftLabs does not
certify compliance or replace legal and audit advisers.
Related cloud paths
Choose the service that owns the decision