LeadershipNot yet confirmed elsewhere1 publisher2 min readPublished
IFS survey finds just 4% of organisations mandate human approval of AI-supported decisions
IFS's survey of more than 2,200 senior decision makers in 14 countries found only 4% of organisations require human approval of AI decisions. IFS published it beside its own AI platform. The gap it describes is also its sales case.
The Board Room · Leadership desk
More than 70 per cent of organisations claimed confidence in their safety and data policies. 11 per cent have validated AI outputs for safety-, quality- or mission-critical processes. 11 per cent have control mechanisms for AI use with sensitive or export-controlled data.
| Figure | Value | Note | Claim |
|---|---|---|---|
| Claim confidence in safety and data policies, over | 70% | 5 | |
| Have validated AI outputs for critical processes | 11% | safety-, quality- or mission-critical | 6 |
| Have controls for AI use with sensitive data | 11% | or export-controlled data | 7 |
What happened
- According to the index, 43 per cent of frontline workers trained and introduced agentic digital workers into their daily operations over the past 12 months.
- Executive leadership's main AI effort over the same period was upskilling management on its own AI literacy, at 41 per cent.
- Over 90 per cent of surveyed organisations reported reliability and performance gains, among them a 90 per cent increase in asset availability.
- Only 11 per cent of organisations have validated AI outputs for safety-, quality- or mission-critical processes.
Why it matters
- contradiction More than 70 per cent claim confidence in their safety and data policies, against 4 per cent with mandatory human approval, a gap of at least 66 points. Stated confidence says little about whether controls exist.
- decision Leaders must choose which AI-supported decisions get a named human approver, and each approval step costs some of the speed the frontline is gaining.
- exposure Supply chain and maintenance planning, where failures cost most, have embedded industrial AI at only 20 per cent of organisations. The thin approval rate matters more as adoption reaches them.
The index is IFS's own research, released at IFS's own conference [1] beside the product the company puts forward as its answer to the gap [14]. The numbers may well be right, but the remedy in the press material is the vendor's remedy. Cathie Hall, IFS's Chief Product and Customer Officer, said: "The real challenge for leadership isn't driving adoption, it's having the insights to clear legacy roadblocks and delivering the secure industrial harness, data pipelines and governance required to scale these wins safely and globally." [11]
The respondents were senior decision makers [2], yet the adoption figures describe what frontline workers did [3]. The coverage does not say how those figures were gathered or what the 43 per cent is a share of.
The speed claim has a second limit. A single survey records levels at one point in time. Adoption and oversight rates sit side by side in it [3][8], so it shows that governance is thin where adoption is under way, and it cannot show whether the gap is widening.
The returns data has its own ambiguity. Over 90 per cent of organisations reported gains, including a 93 per cent improvement in first-time-right performance [10]. The wording supports two readings: 93 per cent of organisations reporting an improvement, or an improvement 93 per cent in size.
IFS Intelligence acts within the rules the business sets [13]. The business therefore still has to write them, and the index says few have written the hard ones: 11 per cent have control mechanisms for AI use with sensitive or export-controlled data [7]. In our view, a platform can enforce an approval rule only after a customer has decided one is needed.
The trade-off is speed against review. Hall said: "To compete, leadership must match the frontline's pace." [12] A mandatory approval step slows the decisions the frontline is accelerating, and leaving it out leaves outputs unchecked. We think the figures support a short list of approval rules this quarter, starting with supply chain and maintenance planning [9]. Next quarter's cost is throughput: as AI moves into those functions, each approval step sets how much of the reported gain survives.
What to watch
- Whether IFS publishes how the frontline figures were collected and what the 43 per cent and 93 per cent are shares of.
- Whether the next edition of the IFS Industrial Index shows the 4 per cent approval rate moving.
- Whether the 20 per cent embedding AI in supply chain and maintenance planning rises before approval rules are written for those functions.
Clarity's read
What the record supports and how the coverage leans. The claims behind it follow.
Reality
- Evidence35
- Adoption30
- Hype gap+40
- Incentives85
- Confidence55
Claim ledger
Ranked by verification strength, evidence, and original report placement.
- [1]
IFS used its annual Unleashed conference in Orlando, Florida, to roll out IFS Intelligence, alongside the release of the IFS Industrial Index 2026.
- [2]
The IFS Industrial Index 2026 is a survey of more than 2,200 senior decision makers across 14 countries.
- [3]
According to the index, 43 per cent of frontline workers have trained and introduced agentic digital workers into their day-to-day operations over the past 12 months.
- [4]
Executive leadership's primary AI effort over the same 12 months was upskilling management to improve their own AI literacy, at 41 per cent.
- [5]
More than 70 per cent of organisations claimed confidence in their safety and data policies.
- [6]
Only 11 per cent of organisations have validated AI outputs for safety-, quality- or mission-critical processes.
- [7]
Only 11 per cent of organisations have established control mechanisms for AI use with sensitive or export-controlled data.
- [8]
Only 4 per cent of organisations have implemented mandatory human oversight or approval requirements for AI-supported decisions.
- [9]
Just 20 per cent of organisations have embedded industrial AI into supply chain management or maintenance planning, described as the functions where operational failures carry the highest consequences.
- [10]
Over 90 per cent of global organisations surveyed reported gains in reliability and performance, including a 93 per cent improvement in first-time-right performance, a 90 per cent increase in asset availability and an 88 per cent improvement in on-time delivery schedule adherence.
- [11]
Cathie Hall, IFS Chief Product and Customer Officer, said: "The real challenge for leadership isn't driving adoption, it's having the insights to clear legacy roadblocks and delivering the secure industrial harness, data pipelines and governance required to scale these wins safely and globally."
ReportedSupportedSource: Cathie Hall, IFS Chief Product and Customer Officer, quoted by itwire.comView cited source - [12]
Cathie Hall said: "To compete, leadership must match the frontline's pace."
ReportedSupportedSource: Cathie Hall, IFS Chief Product and Customer Officer, quoted by itwire.comView cited source - [13]
IFS Intelligence is built around the data an operation already holds, including assets, work orders, schedules and history, draws on live data across IFS and third-party systems, and acts within the rules the business sets.
- [14]
IFS Intelligence is the product IFS is putting forward as its answer to the gap between what industrial operations know and what they can act on.
- [15]
The gap between the share claiming confidence in safety and data policies (more than 70 per cent) and the share with mandatory human approval of AI-supported decisions (4 per cent) is at least 66 percentage points.
Sources
1 independent publisher whose own reporting we read for this story.
- itwire.comIFS Unveils New AI Platform For Industrial Operations
1 article · October 13, 2026
Topics and entities
Follow any of these and your For You feed starts watching them — no settings page required.