Udam Dewaraja - Founder & CEO of StitcherAI. Founded FOCUS, a finance standard adopted by thousands of enterprises.
getty
The CFO asks about a cost overrun.
The IT finance team didn't make the decision that caused it. It wasn't consulted. But it has to explain it.
This scene plays out daily around technology spend, especially with AI costs rising exponentially. It’s a structural problem that no IT finance dashboard, training program or executive mandate has fixed.
That's because the spending mechanism itself has changed. IT costs are no longer set centrally, in budget cycles. They accrue continuously, as developers provision cloud, teams adopt SaaS, business units roll out AI across their workflows and agent fleets trigger inference at machine speed.
Each decision generates cost at a speed and volume no periodic review was designed to capture. By the time the impact surfaces, the feature is live and customers are using it. The IT finance team operates in hindsight: chasing cost overruns and investigating spikes that were locked in weeks before anyone noticed.
This does more than waste money. It spreads discord. The IT finance team becomes the bottleneck and the antagonist. Engineering and product see IT finance questioning decisions without understanding context. Meanwhile, IT finance has no time for higher-order work like enabling AI transformation. Instead, it is consumed by reactive firefighting. Leadership then questions why it's not more strategic.
The longer this pattern holds, the harder it is to break. Gartner projects worldwide IT spending will surpass $6 trillion this year. Fast-growing categories are AI, cloud and SaaS. But data from each of these sources arrives in different formats, schemas and languages. Only the IT finance team can interpret it. So, even as more people make spending decisions, the accountability sticks with IT finance.
Treating this as a people problem won’t fix it. Leadership wants engineers to care about costs, product managers to think about margins and business leaders to review consumption. But asking people to change behavior without giving them the capability to act differently never works.
An engineer focused on cloud costs can't see the cost of an architecture choice made by a different engineer. A product manager attending monthly cost reviews still can't immediately access margin-per-feature data. A business leader committed to "cost-aware decisions" still receives the consumption bill three weeks after new features roll out.
People can't make cost-informed decisions without data in their context and language, at the moment the decision happens.
Three preconditions are needed to distribute accountability. They are:
Every provider of IT services delivers cost data differently. Each cloud has their taxonomy. SaaS and AI providers deliver aggregated costs lacking organizational context. Internal IT costs sit in finance system categories that don't map to how the rest of the organization thinks about spending. Each source is internally consistent. None of them interoperate.
That’s why cost management centralizes by default. Someone has to learn every schema and translate. That someone is IT finance, which then becomes a bottleneck.
When data speaks one language, the IT finance team stops being the translator. A product manager can interpret cost data without understanding cloud billing constructs. An engineering lead can review spending without reconciling three providers’ naming conventions. The barrier that forced centralization disappears. I’ve spoken about this before with the FinOps Open Cost and Usage Specification (FOCUS), which is a standardized scheme for cost and usage data that uses common dimensions and metrics that normalize data regardless of source.
A common language creates consistency, but it doesn't create decisions.
Standardized cost data is still useless if it lands in a dashboard three weeks after the architecture choice. The data has to arrive in business terms (products, teams, revenue streams, margin per feature). It needs to include IT finance data on business-aligned costs. And it all has to arrive where the decision is being made: in the IDE, in the planning ticket, in the Slack conversation so that humans and agents can make cost-aware decisions from the start.
This way, the bills land aggregated. The decisions are informed and local. Until that gap closes, accountability stays with whomever can read the bills.
In the CapEx era, cost governance was finance's responsibility. In the OpEx era, it's everyone's.
Leadership has to say so explicitly. Engineering owns the unit economics of what they build. Product owns the margin of what they ship. Business units own the cost efficiency of the growth they drive. IT finance defines the cost model, sets governance and ensures data quality. Outcomes belong to whomever made the decision.
The culture shift that training programs couldn't force becomes a byproduct of the operating model. Not the other way around.
With distributed accountability, the CFO asks about AI investment ROI. Instead of a two-week data assembly project, the answer is live: cost-per-customer trending down while revenue-per-customer trends up. An engineer evaluates two architectures and sees the cost-per-inference difference in their AI chat session, during the development cycle, not weeks later. Product managers adjust feature rollouts because they see margin impacts. Business unit leaders reallocate toward higher-value customers based on near-realtime profitability data.
Humans govern. Agents execute. Both operate on the same financial truth.
None of these interactions require IT finance to assemble data or defend numbers. They’re all proactive.
Decisions are already distributed. Accountability isn't. That's the gap to close. Enterprises that embrace it will achieve more, faster, with IT spend.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。