Scaling Agentic AI In Global Revenue Operations
Gargi Ray, VP Finance, Group Revenue Operations, Synopsys Inc.
The MIT 2025 State of AI in Business report delivered a sobering reality check that sent shockwaves through the C-suite: 95% of enterprise AI projects fail to move from prototype to production.
For the modern finance executive, this staggering failure rate isn’t an indictment of large language models (LLMs); it is a fundamental failure of operational effectiveness.
Many companies have spent the last few years enamored by the “magic” of generative AI. However, we are now hitting the hard wall of enterprise integration, where that magic is being replaced by the nonnegotiable requirements of absolute accuracy and fiscal accountability. Pilots are deceptively easy to stand up because they thrive on clean, curated data—a luxury that does not exist in the real world.
Across the Fortune 500, many organizations are trapped in a cycle of designing for technical feasibility rather than economic viability or operational scale, asking, “Can AI answer this question?” rather than asking, “How can AI be embedded seamlessly into our legacy systems of record?” Most initiatives remain isolated tasks on a single database. Bolt-on AI agents sit outside the core financial workflow. They are easy to demo in a boardroom but can be nearly impossible to scale because they lack the “connective tissue” to the enterprise’s true systems of record.
From Retrieval To Reasoning: A Case Study
Consider the difference between a simple query and an enterprise-grade one. A simple query might be: “Summarize the key clauses of this contract.” This is an example of isolated retrieval—useful, but low-stakes.
Now, take a complex business question: “What are the payment terms in this contract? How many days late has this customer paid on average for past invoices, and how will that behavioral lag affect our Q3 cash forecast?” This query requires cross-functional intelligence. It cannot be answered by the contract database alone. It requires semantic mapping to understand that “net 30” in a PDF must be calibrated by “Avg_Days_Late” in a SQL table. This is no longer a search; it is a chain of reasoning:
Agent A (Legal): Extracts the net-30 payment terms from the contract
Agent B (Credit and Collections): Queries 24 months of payment history to calculate the behavioral lag and identifies $2 million in upcoming billings from the ERP
The Orchestrator: Adjusts the $2 million inflow by the lag and updates the Q3 cash position with deterministic precision.
Considerations For Success
Companies design pilots in clean sandboxes, but when they hit the “real world” of inconsistent ERP records, fragmented CRM entries and M&A outliers, the accuracy disappears. The hidden costs—manual oversight to fix hallucinations and the compute required to process messy data—frequently outpace the value the AI provides.
To increase our chances of moving into the successful 5%, my company underwent a strategic pivot toward clean, integrated databases and a robust semantic and governance layer. Success requires an architectural mindset: The entire flow must be mapped, and the traceability of information established, before implementation begins.
As I look at the successful deployment of agentic AI at my company, the lesson is clear: You cannot automate deep-dive insights without integrated databases and established traceability. The value gap is bridged only when the AI stops being an experiment and starts being the infrastructure. For finance, AI must be as reliable as the ledger itself. It must handle the edge cases—the acquisitions, the outliers and the messy historical records—that a clean-room pilot ignores. In 2026, true success is defined not by the sophistication of the LLM but by the robustness of the data architecture that supports it, the adaptability to dynamic business models, security and governance of the data and proprietary semantics and workflows.
The information provided here is not investment, tax or financial advice. You should consult with a licensed professional for advice concerning your specific situation.
Forbes Finance Council is an invitation-only organization for executives in successful accounting, financial planning and wealth management firms. Do I qualify?
