Boards expect AI savings: why almost no organisation is finding them

Somewhere in most board packs this year there is a line about AI productivity savings. It was written in good faith, usually a year or so ago. Now someone has to report against it.

New research from Gartner suggests that for most organisations the honest report will be uncomfortable. It also suggests the problem is not the technology. It is that leadership teams never agreed what AI was for.

What the research found

Gartner presented the findings at the opening of its HR Symposium in Orlando on 26 October 2026. Gartner is an analyst firm that also sells advice on these questions, so its figures are best read as strong signals rather than settled fact. The press release does not give the sample size behind its board figures.

The headline: 77% of board members expect AI to deliver productivity and cost savings, yet 97% of organisations are unable to achieve their desired cost savings through productivity alone. Only 3% exceed their savings goals that way.

Gartner's CV Viverito added that "Thirty-eight percent of companies are over budget on AI and 85% report they are planning to spend even more next year." His conclusion was blunt: "Ultimately, organizations will not generate cost savings from AI at this point."

Two further findings explain why. A July 2026 Gartner survey of 297 CHROs found that 78% have no alignment among leaders on where AI creates value. And a survey of 1,865 employees in the second quarter of 2026 found that only 21% of the time freed up by AI was redirected into more innovative work.

Gartner's own prescription is to treat AI as a way "to do things that weren't possible before", rather than as a cost-cutting programme.

The same week brought supporting signals from elsewhere. FIS, a financial technology firm, published research by Opinium among 100 senior UK finance directors: 90% have a clear business case for AI in cash and liquidity management, but only 71% can clearly measure the value their AI initiatives deliver. That is also vendor research, with a small sample. And a UK tech union's AI Workers' Inquiry, newly covered by Computer Weekly, reported workers describing management teams mandating tools they do not understand. That is a campaigning report from one sector, but it shows how savings pressure can feel from the other side of the desk.

What it means for leaders

Read together, the findings describe a chain. The board expects savings. The leadership team has not agreed where AI creates value. Time is freed, but most of it leaks into more of the same work. Finance cannot evidence the return. And the people doing the work feel pushed rather than led.

None of those links is a technology problem. Each one is a leadership decision that was skipped.

Ian Hughes makes the point about what the savings conversation misses in What Stays the Same:

"The true cost of expensive thinking was never just the salaries. It was the long list of questions you never got to ask at all."

If the only question a business asks of AI is how many hours it can remove, it will measure hours, and it will struggle to show cash. If it asks what it can now do that it could not afford to do before, it has something to choose, own and prove.

That does not mean savings are wrong. Some work should stop. But a savings target with no agreed outcome behind it is a number waiting to disappoint.

Which toolkit tools help

The AI Enablement Toolkit for Leaders is a leadership method, not software. Several of its 26 tools fit this problem directly.

- Tool 5, the value case canvas, makes the leadership team agree whose outcome is being improved, and what evidence would show it, before anyone puts a number in front of the board. - Tool 18, the proof-of-value scorecard, separates cash saved, capacity released and new value created, so each can be reported honestly rather than blended into one optimistic figure. - Tool 19, the benefit and evidence register, keeps forecasts apart from observed results. - Tool 21, the opportunity portfolio, gives a board real choices: test, scale, revise, hold or stop. - Tool 22, the leader AI enablement dashboard, tracks whether released capacity is actually redeployed, which speaks directly to Gartner's 21% finding. - Tool 3, the change listening guide, helps leaders hear what is really happening to people's time and confidence.

Where training fits

The CHRO finding is about a leadership team that has not had the conversation. That is where Esm's training and enablement is most useful: a facilitated session in which the executive team and finance agree what AI is for, which two or three outcomes matter, and how value will be evidenced. The toolkit's two-hour leadership working session is designed for exactly that.

The union's findings, and a new ZipRecruiter survey of US workers in which only 29% said they receive enough AI training, point to a second need: practical, role-based learning for the people doing the work, so that adoption is something people are equipped for rather than told to do.

Honest limits

The toolkit cannot make savings appear where the work has not changed, and it will not settle a disagreement that a leadership team is unwilling to have. It does not fix data quality or system integration, which the FIS research identifies as a constraint in finance; that is technical work. It is not employment law advice, and it takes no position on the union's policy proposals.

What the toolkit can do is turn a savings line into a set of owned, evidenced choices that a board can actually govern.

A question to take into the next board meeting

If AI is not going to pay for itself in savings this year, what is it for in your business, and does everyone around the table give the same answer?

Tags: ai-adoption, ai-roi, board-governance, leadership, workforce-planning

Sources

- https://www.gartner.com/en/newsroom/press-releases/2026-10-26-gartner-hr-research-finds-77-percent-of-board-members-expect-ai-to-deliver-productivity-and-cost-savings-yet-just-3-percent-of-organizations-exceed-cost-savings-goals-via-productivity

- https://www.fisglobal.com/about-us/media-room/press-release/2026/fis-research-uk-finance-leaders-back-ai-in-cash-management

- https://www.computerweekly.com/news/366652721/Employees-must-be-empowered-to-refuse-AI-tools-at-work-says-UTAW

- https://techworkersinquiry.org/ai/

- https://www.businesswire.com/news/home/20261026536008/en/ZipRecruiter-Research-53-of-Workers-Say-AI-Has-Made-Their-Work-More-Enjoyable