Firms Deploying Agents
Substantiate outcome-based services.
Connect the work your agents deliver to the results clients purchase and the fees you charge.
Financial Services
EY Extends Revenue Assurance for AI Agents With NVIDIA Verifiable AI on Vera Rubin
Explore the SolutionObjective
EY, a global leader in assurance, tax, and consulting, extends its long-standing standards of audit integrity to work performed by AI agents. Using NVIDIA’s Verifiable AI on Vera Rubin, EY helps firms prove that agents deployed in client environments delivered the work they billed for, with hardware-backed records that trace each outcome to the contract and the client’s acceptance under revenue standards like ASC 606. EY’s expertise and NVIDIA’s Verifiable AI set a new standard of assurance for agent-delivered work, ushering in outcome-based business models that both firms and their clients can trust.
- Customer
- EY
- Partner
- EQTY Lab
- Use case
- Financial Audit
- Revenue Recognition
- Agentic AI
- Technologies
- NVIDIA Verifiable AI
- NVIDIA Vera Rubin
Backing Every Agent Outcome With Proof Clients Can Trust
AI agents introduce new opportunities for instrumenting expert services in client environments, delivering white glove capabilities on demand. As providers contribute to this emerging market, clients need assurance that the agents operate within their boundaries and without compromising control of their data.
EY reimagines audit standards for agent-delivered work by connecting the work an agent performs, the result a client accepted, and the recorded revenue for an auditor. NVIDIA’s Verifiable AI on Vera Rubin and EQTY Lab produce hardware-backed records that anyone can verify independently. Those records help establish what work took place and whether the evidence has been tampered with.
With real-time evidence, firms can substantiate delivery to both clients and auditors, establishing a stronger basis for selling higher-value services and taking on more consequential work.
A service provider and customer agree on the deliverable, data, benchmark, and fee. An agent produces the output. An evaluator checks the data and performance, and the customer records acceptance. A signed work record connects the accepted output to the provider’s revenue record.
The auditor examines the contract, performance obligations, transaction price, price allocation, and fulfillment and timing. Each step is linked to the relevant evidence in the workflow.
Applying ASC 606 to Agent-Delivered Work
Verifiable AI creates a durable record of agent activity whose origin and integrity can be checked independently, even after the original runtime is no longer active. These records support many forms of assurance. One important use is examining the revenue earned by agent-delivered services.
ASC 606’s five-step revenue recognition model applies whether people or agents deliver the service. Companies identify the contract and its performance obligations, determine and allocate the transaction price, and recognize revenue when or as those obligations are satisfied. EY is bringing its audit expertise to the evidence needed for agent work within that established framework.
1
Identify the contract
Connect the agent’s work to the agreed engagement and its terms.
2
Identify the performance obligations
Map the delivered result to the promised service and its acceptance criteria.
3
Determine the transaction price
Compare the evaluated result with the fee terms, including any conditions tied to performance.
4
Allocate the transaction price
Connect the evidence to each obligation and its allocated amount.
5
Recognize revenue as obligations are satisfied
Use delivery, evaluation, and acceptance records to support the assessment of fulfillment and timing.
Assurance rooted in silicon opens new ways to audit revenue from agent-delivered work. Tools such as the Pre-Audit Report Generator connect contracts, signed work records, evaluations, client acceptance, and ledger entries to help auditors assess fulfillment and verify the integrity of the supporting evidence.
A Stronger Basis for Agent-Earned Revenue
Verifiable evidence gives each party a clearer basis for its own decisions.
Finance & Revenue Teams
Bring delivery evidence into the close.
Reconcile contracts, accepted results, and ledger entries to support revenue recognition and cutoff.
Clients
Know what you are accepting.
Review evaluated results against agreed criteria while retaining control of your systems and data.
Financial Auditors
Verify the supporting record.
Check the integrity and relevance of agent evidence, then apply judgment to the revenue it supports.
Looking Forward
As firms deliver more services through agents in client environments, they need a consistent basis for defining outcomes, confirming acceptance, and substantiating revenue. A shared record of the work can carry evidence of delivery between providers, clients, and auditors while customers retain control of their systems and data.
EY’s work on trust in autonomous AI points toward assurance that evolves alongside the technology. Connecting agent activity to portable evidence moves that trust into financial workflows, supporting a market in which increasingly consequential work can be commissioned, accepted, and audited.
Solution Highlights
- Hardware-backed records that connect agent activity to its output and support independent integrity checks.
- A separate evaluator checks approved data, performance, and the supporting work record.
- Client acceptance linked to the delivered result, contract, and fee terms.
- Evidence organized for revenue review under ASC 606, with accounting and audit judgments retained.
Explore how EY brings trusted agents into your financial workflows
Learn More