Onaro launches Meridian to track AI labor costs

Onaro on September 8 introduced Meridian, a FinOps system of record designed to help enterprises meter AI usage, assign costs, and document the return on AI agents. The Greenwich, Connecticut company is targeting finance, risk, compliance, and engineering teams as AI spending becomes harder to track across tools and vendors.

Why it matters: - Enterprise AI spending is spreading across model providers, gateways, cloud platforms, and internal tools, making costs harder to track and control. - Onaro says Meridian is built to give finance and AI teams a record they can use for budgets, chargeback, ROI analysis, and audit support. - The launch comes as KPMG reported only 7% of leaders had established ROI from AI investments, and Zylo found 78% of IT leaders saw unexpected charges tied to consumption-based or AI pricing models.

What happened: - Onaro announced Meridian on September 8, 2026, in Greenwich, Connecticut. - The company describes Meridian as a system of record for what enterprise AI labor costs and returns. - Meridian is designed to meter AI usage, attribute spend to agents, owners, workflows, and business units, and produce financial and operating records.

The details: - Meridian is organized around three functions: meter, control, and prove. - The platform connects in read-only mode to model providers, LLM gateways, platform billing sources, and internal systems. - Meridian captures AI usage and spending, then attributes that spend to accountable owners and workflows. - The system is built to identify unattributed costs, cost drift, and concentration risks. - Meridian supports budgets, anomaly alerts, and policy actions through existing organizational controls. - The recordkeeping layer supports chargeback files, cost-per-outcome analysis, ROI analysis, and audit evidence exports. - Onaro says the platform is available to corporate clients as a managed service or private deployment. - Onboarding, support, and custom integration options are included. - The first 20 customers join a founding cohort with preferred, rate-locked pricing. - Organizations can start with a free, two-week read-only Agent Spend Assessment. - The assessment connects to selected AI and billing sources, produces an Agent Spend Report, and ends with a live CFO read-out. - Onaro says the report may surface unattributed spend, idle agents, cost drift, attribution readiness, and opportunity estimates without write access to production systems.

Between the lines: - Onaro is positioning Meridian less like a monitoring dashboard and more like financial infrastructure for AI labor. - The company is targeting a core enterprise problem: AI agents can create spend without clear ownership or easy accounting treatment. - The pitch also reflects a broader shift from AI experimentation to operational governance, where finance and compliance teams want evidence, not just usage data.

What's next: - Onaro expects enterprises to start with the assessment before moving into Meridian deployments. - The company is offering managed service and private deployment paths for organizations that want more control over integration and data handling. - The founding cohort and rate-locked pricing suggest Onaro is trying to lock in early enterprise customers as the product rolls out.

The bottom line: - Meridian is Onaro's attempt to make AI labor measurable, allocable, and auditable for the finance function.