Azure FinOps
Take back control of the Azure bill without slowing teams down
FinOps is not a one-off savings hunt: it is a practice that makes costs visible, attributable and manageable by the people who generate them. I set up the tooling, the rules and the rituals, then train your teams.
Who is it for?
CIOs, platform leads and finance controllers facing an Azure bill growing faster than usage, costs that cannot be split across teams, or AI workloads whose cost per token escapes any tracking.
What I deliver
- Visibility: mandatory tagging, Cost Management, exports and dashboards per team, product and environment
- Fast optimization: right-sizing, off-hours shutdown, storage tiers, orphaned resource cleanup
- Commitments: reservations, savings plans, Azure Hybrid Benefit and break-even analysis
- Governance: budgets, alerts, policies blocking out-of-scope SKUs, monthly anomaly review
- AI FinOps: tokens, costs and tools measured per person, team and model; caps and model fallback
- Rituals and training so the practice outlives the engagement
Method in 4 steps
- 1
Measure
Reading the bill, tagging and cost allocation; spotting the resources that concentrate spend; KQL queries on orphaned resources, including Capacity Reservation Groups with no VMs.
Deliverable: Cost baseline and a costed list of optimizations
- 2
Quick wins
Risk-free actions: start & stop of non-production environments, Azure Hybrid Benefit, removal of orphaned resources. Right-sizing is handled separately, after about three months of metrics.
Deliverable: Savings delivered and measured
- 3
Structure
Budgets and alerts per team, tagging enforced by Azure Policy, commitments (reservations, savings plans) sized on real usage, Infracost in the pipelines.
Deliverable: Budgets, tagging policies and estimated cost in pull requests
- 4
Culture
Recurring FinOps review with finance and product teams, shared indicators, training. Extension to AI inference costs: tokens per team and per model.
Deliverable: Rituals and teams accountable for their cost
Proof points
Allocated costs
every euro tied to a team and a product, the prerequisite for steering
Agentic AI FinOps
tokens, costs and tools measured per person, team and model, presented at conferences
Continuous optimization
improvement loops built into architecture reviews, not a forgotten audit
- Read the guide
Guide · 18 min read
Taking an AI agent platform to production on Azure with Microsoft Foundry and API Management
Guide: take AI agents to production on Azure with Microsoft Foundry and API Management as the AI gateway. Key-less identity, token quotas, MCP, guardrails.
- Read the guide
Guide · 15 min read
Generative AI FinOps on Azure: understanding, measuring and holding token costs
Guide: control generative AI costs on Azure. Token cost model, deployment types, PTUs and reservations, caching, per-team quotas, showback.
Cost, timeline and deliverables of a typical engagement
Three formats, from the shortest to the most complete. Durations are orders of magnitude observed on my engagements; the exact scope and quote are set after the first 30-minute call, within 48 hours.
3 to 5 days
Express FinOps review
Analysis of the bill and cost exports, top 10 immediate savings with figures, reservations and savings plans to commit, debrief with budget owners.
3 to 6 weeks
FinOps framework set-up
Allocation model (tags, resource groups, subscriptions), budgets and alerts, per-team dashboards, sizing rules, automated start & stop of non-production environments.
3 months, part time
Quarterly coaching
Monthly cost review with the teams, commitment tracking, FinOps built into pipelines and architecture reviews, until the team is autonomous.
Fixed price for the review and the framework, capped time-and-materials for the coaching. The daily rate and conditions (remote, on site, duration) are stated in the proposal. No licence resale and no managed services: you own the code and the tenant.
- Knowledge transfer included at every step (pair reviews, workshops, documentation)
- Code and pipelines delivered in your repository, under your licence
- Weekly progress review with decision makers
Frequently asked questions
How much can we save?
On a never-optimized scope, 20 to 35% savings are common in the first year (right-sizing, commitments, off-hours environments). The real win is then staying in control over time.
Do we need a third-party FinOps tool?
Rarely at first. Azure Cost Management, well-structured exports and disciplined tagging cover most needs. A third-party tool makes sense for multi-cloud or very large estates.
How do we track AI usage costs?
By instrumenting calls (gateway, APIM or agentgateway) to measure tokens and tools per user and per model, with caps, alerts and automatic fallback to a cheaper model when relevant.
Discuss this with me
A first 30-minute call is free and without commitment.
Other services
Related articles
Hands-on Azure, AKS, FinOps and agentic AI write-ups published on blog.smartcloudarchitect.fr.
All articles on the blog· 9 min read
FinOps de l'IA agentique : mesurer jetons, coûts et outils par personne, équipe et modèle· 12 min read
Des MCP d'entreprise pour les agents : Azure MCP Server, Microsoft Learn et une couche tickets avec approbation humaine· 12 min read
« Agent Ready » : onboarder une équipe sur une plateforme agentique, sans ticket et avec des droits bornés· 11 min read
Garde-fous IA portables : NeMo Guardrails derrière agentgateway, en fail-closed