The news
Fragmented AI tools are eating into returns, according to FICO's (FICO) State of Responsible AI 2026 report, conducted by Corinium among more than 1,000 senior technology, data and risk executives and reported by CIO Dive on October 7. Only 8.2% of organizations have adopted a single shared AI deployment platform, while 71.2% have partially done so and about one in five are working toward one.
Returns are modest. In the FICO survey, 5% said AI significantly exceeded ROI expectations, 43% said it somewhat exceeded them and 37% said it met them. Security concerns were cited by 40% as a barrier to returns from agentic AI, and 38% named data integration and quality as the biggest barrier, CIO Dive reported.
FICO Chief Analytics Officer Scott Zoldi told CIO Dive that sprawl often means shadow projects and systems left running that raise costs along with hyperscaler token use. He said governance teams stop anywhere from 50% to 90% of AI deployments because they are not sound, and that a shared standard is when ROI starts to develop.
Adoption is also lopsided. A Channel V Media report, based on a June survey of 7,675 U.S. adults including 1,960 executives or business owners, found 53.2% of executives use generative AI compared with 37.9% of other workers, and executives were 2.7 times more likely to use agentic AI (17.5% vs. 6.6%), CIO Dive reported. A separate Emergn study of 350 senior leaders found 18% admitted pretending to understand more about AI than they do.
Wall Street is divided on what this means for the spending cycle. In a report released October 7, Citi Wealth said the AI buildout is still in its early-to-middle stages, citing semiconductor demand that exceeds near-term supply, PYMNTS reported. In a note dated Monday, October 5, Bank of America (BAC) strategists said capex strength may be more priced in than not and called for a selective pivot toward consumer-driven spending, according to PYMNTS.
The numbers
- Firms on a single shared AI platform
- 8.2% (FICO)
- AI significantly exceeded ROI expectations
- 5% (FICO)
- Executives vs. other workers using generative AI
- 53.2% vs. 37.9%
- Leaders who admit pretending to understand AI
- 18% (Emergn)
- Projected hyperscaler AI infrastructure spend, 2027
- $1.2 trillion (Goldman Sachs, via PYMNTS)
Why CEOs should care
CIOs and CFOs should treat platform sprawl as a cost line. Ask for an inventory of every AI tool, model contract and running workload, who owns each, and what token spend each drives. FICO's data suggests the firms that standardize on a shared deployment and governance platform are the ones that clear review and see returns, so consolidation may be cheaper than another pilot.
Boards should question whether adoption is real below the C-suite. When executives use generative AI far more than staff, and nearly one in five senior leaders admit overstating their AI knowledge, ROI projections built on broad workforce productivity deserve scrutiny. Training budgets and usage metrics by team are fair asks.
For investors and finance chiefs planning 2027 budgets, the split between Citi Wealth and Bank of America is a reminder that big AI spending assumptions are contested. Goldman Sachs projects hyperscaler AI infrastructure spending of $1.2 trillion in 2027, up from $800 billion in 2026, PYMNTS reported. Build plans that hold up if vendor pricing, interest rates or demand shift. This article is not investment advice.
The bigger picture
The pattern is familiar from earlier technology waves: spending runs ahead of the operating discipline needed to capture value. Citi Wealth listed higher interest rates, more capable open-weight models, regulatory pressure and rising infrastructure costs as reasons the cycle faces scrutiny, while still seeing a long runway. Enterprises that tighten governance now will be better placed whichever side of that debate proves right.
What’s next
Watch upcoming quarterly results from hyperscalers and chipmakers for signs of capex growth holding or slowing, and whether more enterprises report consolidating AI tools onto shared platforms in 2027 budget plans.
What “Fact-checked” means
Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.
- What we checked
- Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
- How
- A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
- Who
- The checks are made by our newsroom, as steps kept separate from the writing, under rules set by our editor, Hussein Mukhtar. A story the checks still flag is held for the editor, who decides whether it is fixed, published or dropped.
- If something is wrong
- “Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error
