Key takeaway
- FinOps is a cultural practice for managing cloud and technology spend. Learn the six principles, three phases and Crawl, Walk, Run maturity levels.
FinOps, short for financial operations, is an operational framework and cultural practice that maximises the business value of technology, supports data-driven decisions and creates financial accountability across engineering, finance and business teams. That wording comes from the FinOps Foundation framework, and it explains why FinOps is more than cost cutting: it is about deciding what technology spend is worth.
This guide is for cloud, finance and platform leaders who run, or are starting, a FinOps practice. It covers the framework phases and principles, who takes part, what capabilities a practice builds, how maturity grows, and how AI agents change the work. It sits alongside our overview of FinOps and our guides to FinOps for data management and FinOps for Snowflake.
Why does cloud spending need FinOps?
In cloud computing, teams provision resources on demand, so cost follows usage and changes continuously. Static yearly budgets, built for fixed infrastructure, struggle with that. FinOps responds by putting finance, technology and business people in the same conversation, and by making usage and cost data visible to the people who control them. The related practice of DataFinOps applies the same thinking to data platforms.
What are the three FinOps phases?
The FinOps Foundation says FinOps is performed "by working iteratively on the Framework Capabilities through three phases: Inform, Optimize and Operate" (FinOps Foundation, Phases). The phases repeat, and teams can run them at their own pace, with different people working on different phases.
| Phase | What happens | Who leads the work |
|---|---|---|
| Inform | Teams review technology cost, usage and efficiency data to understand the current state. Accurate allocation supports reporting, forecasting and benchmarking, including real-time monitoring where it is available | Finance and engineering together |
| Optimize | Teams find ways to improve efficiency and value, through usage changes such as rightsizing and rate changes such as negotiating better pricing. Options are documented and prioritised, and "Optimization options may compete" | Engineering for usage; procurement and leadership for rates |
| Operate | Teams implement the chosen changes. The phase depends on accountability across engineering, finance and business groups, who make incremental decisions from Inform data | Engineering, finance and business groups |
The Foundation notes that cloud and AI pricing changes often, so teams must "continuously revisit the Inform phase to validate the impact of optimization actions". Operate feeds back into the earlier phases, so the cycle never ends.
What are the six FinOps principles?
The framework lists six principles (FinOps Foundation, Principles). Older articles may name different principles, so check the current framework page.
| Principle | What it means in practice (FinOps Foundation wording) |
|---|---|
| Teams need to collaborate | Finance, technology, product and leaders work together to manage each technology category |
| Business value drives technology decisions | Unit economic and value-based metrics demonstrate business impact |
| Everyone takes ownership for their technology usage | Accountability of usage and cost is pushed to the edge |
| FinOps data should be accessible, timely, and accurate | Process and share cost data as soon as it becomes available |
| FinOps should be enabled centrally | A centralised FinOps function encourages, evangelises and enables best practices in a shared accountability model |
| Take advantage of the variable cost model of the cloud | Make technology decisions based on cost model considerations as well as technical requirements |
Who takes part in FinOps?
The framework names core personas (FinOps practitioner, engineering, finance, leadership, procurement and product) and allied personas (IT asset management, IT financial management, IT service management, security and sustainability) (FinOps Framework).
What capabilities does a FinOps practice build?
The framework groups its capabilities into four domains, and it also defines scopes: segments of spend, tied to business constructs such as products or cost centres, that a practice can manage separately.
| Domain | Capabilities named by the framework |
|---|---|
| Understand usage and cost | Data ingestion, allocation, reporting and analytics, anomaly management |
| Quantify business value | Planning and estimating, forecasting, budgeting, KPIs and benchmarking, unit economics |
| Optimise usage and cost | Architecting and workload placement, usage optimisation, rate optimisation, licensing and SaaS, sustainability |
| Manage the FinOps practice | Executive strategy alignment, practice operations, governance, policy and risk, education, invoicing and chargeback, assessment, automation, tools and services, intersecting disciplines |
How does FinOps maturity grow?
The framework describes maturity as "Crawl, Walk, Run": organisations start small and grow in scale, scope and complexity. Assess each capability separately, because a team may run anomaly management while still crawling on forecasting.
How do AI agents change FinOps?
FinOps already relies on automation, and AI agents extend it. An agent can watch cost and usage data, flag anomalies, draft a rightsizing proposal with the evidence behind it and route it to the owner of that workload. Because AI workloads add their own pricing patterns, the Foundation says teams must keep revisiting the Inform phase. Decision rights matter most: write down which savings an agent may apply alone, such as shutting an idle test environment, and which need a human owner. Machine learning and artificial intelligence methods help with forecasting, but forecasts still need review against business plans. To build and run agents for cost and cloud operations, see ElixirClaw, the AI agents computing platform.
What should you do first?
- Choose one scope, such as a product or business unit, and name an accountable owner for its cloud spend.
- Make cost and usage data visible to that owner, and agree how costs are allocated.
- List optimisation options, prioritise them against business goals and assign each an owner.
- Set a regular review cadence and revisit the data after every change to check its effect.
- Decide which actions an agent may take alone, and record the evidence it must attach.
To map one cloud spending decision your agents make, and the data and policy it depends on, book a Context Workshop. See what happens in the workshop.
Frequently asked questions
What does FinOps stand for?
FinOps is short for financial operations. The FinOps Foundation describes it as an operational framework and cultural practice that maximises the business value of technology and creates financial accountability across engineering, finance and business teams.
What are the phases of FinOps?
The three phases are Inform, Optimize and Operate. Teams work through them iteratively, reviewing cost and usage data, choosing and prioritising improvements, then implementing them and looping back to measure the result.
What are the FinOps principles?
The six principles are that teams need to collaborate, business value drives technology decisions, everyone takes ownership for their technology usage, FinOps data should be accessible, timely and accurate, FinOps should be enabled centrally, and teams should take advantage of the variable cost model of the cloud.
What does Crawl, Walk, Run mean in FinOps?
It is the maturity approach in the FinOps Framework. Organisations start small and grow in scale, scope and complexity as their practice develops.
Is FinOps only about cutting costs?
No. The framework centres on business value. Optimisation covers both usage and rates, but decisions are meant to be driven by the value the spend delivers rather than by the lowest bill.
Follow XenonStack on LinkedIn: XenonStack on LinkedIn for new guides on agentic AI, data and platform engineering.
Ready to put AI agents to work? Explore ElixirClaw, the agentic OS platform to build, orchestrate and govern AI agents.
Explore the XenonStack platform family: ElixirData · ElixirClaw · ElixirHub · NexaStack · MetaSecure · Neural AI · AgentGRC · AgentSRE · AgentAnalyst · AgentInstruct · AgentRAI · AgentScrum · AgentSearch · AgentForce · AI Aviator · SAIF Aviator · CohortIQ · XenonStack AI · XenonStack UAE