THE GAP BETWEEN INTENTION AND CONFIDENCE

New research from The Futurum Group, commissioned by industrial software provider IFS, finds that industrial workers lose 41 per cent of their time to manual, repetitive tasks, producing what the report terms a capacity gap: the widening distance between the work industrial organisations need to do and the actual capacity of their teams. Among aerospace and defence decision-makers, 77 per cent say they have delayed or avoided a strategic initiative because their team lacked the capacity to pursue it.

The response identified is a shift toward agentic digital workers, defined as AI agents that autonomously monitor, decide and execute operational tasks, escalating to humans only where judgement is required. Adoption is described as accelerating, with two businesses investing in digital workers over the next twelve months for every one that is not, and 66 per cent saying they are likely to invest within the year.

Set against that intention is a single figure that reframes the whole result. Just 5.7 per cent of decision-makers trust AI to act fully autonomously. IFS states the position plainly: trust, not technology, is the holdup.

WHAT A TWELVE-TO-ONE RATIO ACTUALLY DESCRIBES

The distance between those two numbers deserves attention, because it is not a contradiction so much as a description of how the technology is actually being bought. Roughly twelve times as many organisations intend to deploy autonomous agents as are willing to let them operate without supervision. What that indicates is that nobody is purchasing autonomy. They are purchasing throughput on tasks they have already decided do not require judgement, with a human retained at the point where judgement begins.

IFS’s own product architecture reflects the same split. The company states its Loops Agentic Platform enables enterprises to deploy digital workers that automate 60 per cent of agentic transactions end to end, with the remaining 40 per cent including human review and approval checkpoints, which it attributes directly to industrial customers’ desire for transparency and validation. A vendor building a 40 per cent human checkpoint rate into its platform by design is not describing a limitation of the technology. It is describing what the market will accept.

Somya Kapoor, Chief Executive of IFS Loops, said the capacity gap was a high-stakes problem in industrial operations, and that when a purchase order, maintenance schedule or supplier delivery falls behind because there are not enough hours in the day, the cost appears as downtime, missed deliveries or idle equipment. Closing it, she said, takes more than general-purpose AI, and requires digital workers built for the job and integrated into the systems already running the business.

WHERE AEROSPACE AND DEFENCE IS APPLYING IT

The research identifies the top three agentic AI use cases in aerospace and defence as resource and schedule planning, project knowledge management, and sourcing and subcontracting. The common characteristic is instructive: each involves high volumes of well-defined transactions across long timescales and multiple parties, and none is a task where a wrong answer is immediately unsafe. These are the parts of an aerospace business where the work is voluminous, the rules are documented and the consequence of an error is commercial rather than physical.

Project knowledge management is arguably the most significant of the three, given the workforce change the report identifies. Experienced engineers and operators are retiring across factories, utilities and supply chains, taking with them the undocumented knowledge that keeps complex programmes running. That is a problem World Airnews Daily encountered in concrete form earlier this month at GE Aerospace’s Lynn site, where two operators with roughly 64 years of service between them worked with management to convert their individual knowledge of a heat-treat oven into documented standard work. The digital worker approach addresses the same loss from a different direction.

DEPLOYMENTS IN PRODUCTION

IFS cites customers running purpose-built digital workers in production rather than in pilot. Kitron Group and KLN Family Brands are described as reclaiming dozens of hours a week on core procurement and order-to-pay workflows, with several pushing toward fully automated processing. Kitron is rolling out purchase-to-order digital workers across all 13 of its sites by the end of 2026, built into software the business already uses, so that existing guardrails, rules and approval processes remain intact.

One detail from that rollout is worth isolating. A digital worker surfaced a decade-old part-number error that a manual process had missed for years. IFS presents this as evidence that digital workers can find and fix data problems as they run, rather than requiring a data cleanup project before deployment. The underlying point is about consistency rather than intelligence: a system that applies the same check to every transaction will eventually encounter a record that a human, checking selectively across ten years, never happened to look at.

On pace, Ependion brought a Supply Order Manager digital worker live company-wide in a ten-week progressive rollout, while Kitron is on track to have all 13 sites live in under seven months. The full report, titled Bridging the Capacity Gap, is published by The Futurum Group.