Thought leadership from the frontlines of finance transformation.
AI agents are moving into the financial close, and sometime in the next few audit cycles an external auditor is going to ask a question most companies cannot answer: how is the agent's output controlled? The answers on offer — "a person reviews it," "the model checks itself" — would not survive a walkthrough if a human performed the work the same way. Finance solved this exact problem decades ago and named the solution segregation of duties. The control transfers. The profession just hasn't noticed yet.
Most mid-market companies are sitting on the single most expensive input to AI adoption — and they have it filed under "failed." The RPA program that stalled out between 2016 and 2020 left behind exactly what a six-figure "AI discovery" engagement is hired to produce: a documented map of how work flows, where it breaks, and where a human has to step in. With process mining now a $1.1B software category and CFOs raising AI budgets while headcount stays flat, the smartest move this budget cycle may be to reopen a folder nobody has touched since the steering committee stopped meeting.
India's global capability centers are out-hiring everyone and quietly losing the talent market. Four straight years of falling employer-brand perception land hardest on career development — the exact promise the capability-hub pitch depends on — and the bill is arriving as sector-leading pay premiums, not attrition. The fix is an operating-model rebuild, not a rebrand.
More and more buyers now ask an AI assistant, not Google, which vendors they should consider — 51% of B2B software buyers already begin research with a chatbot, up from 29% eleven months earlier. The sources those assistants cite overlap less than 20% with the pages that rank in search, down from roughly 70%. Ranking is no longer the prize; being cited — and described correctly — inside an answer your buyer trusts is. The shortlist already moved, and most vendors haven't noticed.
Microsoft found 75% of knowledge workers use AI at work — and 78% of those users bring their own tools. McKinsey found only 6% of companies capture 5%+ of EBIT from AI. Both numbers are right, and together they describe the defining strategic failure of the current AI cycle. AI is not failing at work — it is succeeding spectacularly for individuals. The question is whether your company becomes smarter because of it, or merely poorer for having paid for it twice.
When every competitor has access to the same intelligence, the model isn't the moat — your data is. Frontier models are converging fast, and organizations that rely solely on general-purpose APIs are renting capability, not building it. The winners are investing in domain adaptation and workflow integration to create compounding advantages their competitors can't replicate.
A Fortune 500 controller watched an AI agent process 4,200 reconciliation entries overnight — then paused the deployment. Not because it failed, but because the trust architecture around it didn't exist. Across finance, the function best suited for agentic AI is the one least willing to let agents act autonomously. The gap isn't technology. It's trust — and it manifests at three distinct levels most organizations haven't learned to distinguish.
Enterprise leaders have been choosing AI systems by benchmark scores — optimizing for raw intelligence. But when a Fortune 100 company pitted its top-ranked model against one with a wider context window, the results upended its entire AI strategy. The real constraint on enterprise AI isn't how smart the model is. It's how much of the problem it can see.
AI adoption in finance jumped from 37% to 58% between 2023 and 2024. Then it froze. In a year when worldwide AI spending surged 44% to $2.5 trillion, actual adoption ticked up a single percentage point. The bottleneck isn't awareness, budget, or executive will. It's operational readiness — and the Audit-Ready Autonomy Stack is how you break through.
A Fortune 200 company deployed AI-powered invoice matching and saw touchless processing jump from 31% to 68% in four months. Then the GBS leader looked at the bill — the monthly fee hadn't changed. The provider was capturing every dollar of productivity gain. Across the $65 billion F&A outsourcing market, legacy contracts are forfeiting the value AI creates. The fix isn't better technology. It's better contract architecture.
The CFO of a $7 billion industrial conglomerate discovered that her flagship division's reported 28% operating margin was actually 19% — inflated by allocation percentages nobody had questioned since 2015. Roughly $340 million in capital decisions had been directed to a business unit whose economics the board had fundamentally misunderstood. Across global enterprises, cost allocation remains frozen in place — and the result is not imprecision. It is systematic misdirection of capital, compounding silently over years.
Get our latest insights delivered directly to your inbox.
