Cloud and AI are now among the fastest-growing lines in the IT budget, and among the hardest to explain. Usage-based pricing, new models and teams spinning up their own tools make it difficult to say who is spending what, and why. CostPerform Cloud & AI, powered by Mavvrik, is built to close that gap. Unlike standalone FinOps tools, it brings your cloud and AI spend directly into your CostPerform model, so those costs are allocated with the same rigor as the rest of IT. In this article, we walk through how it works in five steps, from connecting your providers to setting budget guardrails and running showback and chargeback.
About Mavvrik
CostPerform Cloud & AI is powered by Mavvrik, a US-based AI and cloud cost governance and optimization platform. Mavvrik connects directly to cloud providers, AI model providers and GPU infrastructure, including on-prem and private AI environments that have no billing API, to collect cost and usage data at the source. CostPerform applies its proven cost modeling and allocation principles on top of that data, so cloud and AI spend flows straight into your financial model.
How it Works
For an enterprise user, the onboarding phase is much lighter than a traditional enterprise implementation or transformation project. The solution is designed to build on the already multidimensional data sources of the tech providers who lead enterprise AI and cloud adoption.
To create value from CostPerform Cloud & AI, we break down the process into five steps. From ingesting signals to acting on the insights:
Step 1: Ingest signals
Every cost starts as a signal at its source. CostPerform Cloud & AI connects to providers via API, automatically collecting cost, usage, and telemetry data. The Mavvrik engine connects to billing and usage data from cloud providers and AI companies. You can directly connect to Anthropic, AWS, Google Cloud (GCP), OpenAI and many more. Visit the CostPerform Cloud & AI homepage to see the full list of 30 available connections.
Step 2: Combine
This step is mostly automated. The tool combines the data from the sources you have provided. Because the APIs share live data, you get a continuously updated breakdown of your cloud and AI costs. In this phase, you can normalize tokens, requests, GPU hours, and more into one consistent usage model, implement live views, and virtually tag your data. A well-built semantic model here lays the foundation for great insights.
Step 3: Allocate
This is still a CostPerform product: we ensure that the data is not only collected and displayed, but also incorporated into your financial logic, using the same principles CostPerform applies in regulatory-grade cost reporting. You can work with showback and chargeback functionalities, capture unit economics, and simply allocate shared costs by team, business unit or product.
Step 4: Govern
The governance phase is the link from collecting the data to implementing the changes in the organization. You can set up budget guardrails by team, project, model or agent. You can forecast future spend and margin impact from actual usage. This phase may also include notifications via Slack or Microsoft Teams, or ticket integrations with Jira or ServiceNow.
Step 5: Act
The data doesn’t exist in isolation; it must be brought in to drive decision-making. The solution can proactively flag over-provisioned resources or spending anomalies on certain models or providers. You get the whole cost picture, and can carry your total AI and cloud spend into the wider CostPerform model, identifying the bill of IT, product and customer profitability and beyond for the whole IT department and organization.
It starts with an assessment
To learn more about the integration process and the path to value of CostPerform Cloud & AI in your business, contact us to plan a free assessment. After a 45-minute call with our team, we provide a customized 6-page report with quick wins and optimization potential outlined, helping you learn how the 5-step implementation listed above will work at your organization. For further details, download CostPerform’s True Cost of AI & Cloud whitepaper.