FEATURE Industrial AI
MOVING FROM PREDICTION TO ACTION
Ali Mustoe-Playfair, Director of Agentic Operations at ANS, says more factory data hasn’t made manufacturers more productive. She believes
Agentic AI can put data to better use, and help optimise decision-making
M
anufacturers have spent years investing in connected machinery, sensors and digital systems to gain greater
visibility across their operations. Today, a typical factory can generate vast amounts of data about asset condition, production performance, quality and maintenance. Despite this, engineers still spend valuable time searching for information, maintenance teams often remain reactive and critical operational knowledge can be difficult to find when it’s most needed. The challenge is increasingly not whether the data exists, but whether people can bring the right information together quickly enough to make a decision. That creates a less visible form of operational inefficiency: decision latency. The time between an issue emerging, the relevant context being assembled and someone deciding what to do can become a significant source of lost productivity. Microsoft uses the term “Frontier Firms” to describe organisations where people and AI agents work together to get work done. For manufacturers, this points to a different way of thinking about AI – not as another source of information, but as a way to reduce the time and effort involved in turning existing information into decisions and action. Consider an engineer investigating a
recurring problem on a production line. The information they need may already exist, but is spread across different systems, from machine data in SCADA or MES to maintenance records, production schedules, technical procedures and engineering drawings.
The engineer must bring that information together before deciding what to do. This may take only minutes for one task,
14 September 2026 | Automation
but multiplied across thousands of decisions, it becomes a significant drag on productivity.
The same applies beyond maintenance. Root cause investigations, production planning, quality issues, compliance reporting and engineering handovers all require people to find and connect information before work can move forward. The factory may be highly automated, but decision-making can remain surprisingly manual.
From retrieving information to orchestrating action This is where agentic AI can go further than more widely used AI assistants. An AI copilot might help an engineer find a maintenance record or summarise a document. An AI agent can bring information together from different systems and help organise what needs to happen next, while leaving the final decision with the person responsible. When a production asset begins showing signs of deterioration, for example, an agent could bring together recent condition data, previous work orders, engineering documents, spare- parts availability and the production schedule. Instead of giving the engineer another alert to investigate, it can provide the information they need to understand the problem and decide what to do next. Experienced engineers know which assets are prone to certain failures, which procedures work in practice and which small warning signs are worth investigating. Some of this knowledge is documented, but much of it comes from
years of working with the same equipment and processes. Agentic AI cannot replace that expertise simply by having access to more data. Instead, it can make relevant information easier to find, helping engineers spend less time searching and more time using their experience to solve problems. This becomes particularly important when only a few experienced people know how to deal with a particular process or problem, meaning they can quickly become a bottleneck. Making the right information easier to find means less experienced engineers can handle more situations themselves, while experienced engineers remain responsible for the decisions that need their expertise. For manufacturers to become Frontier Firms, they must look beyond whether AI can predict when something might go wrong and asking what happens next. Can the right information reach the right person quickly? Can systems work together without creating another manual process? Can AI handle routine tasks while engineers remain in control of important decisions? Manufacturers do not need to start again with their technology. The data, systems and expertise they need are often already there; the challenge is making them easier to use together.
That is where agentic AI could have its
greatest impact. Not by giving manufacturers more information, but by helping people make better use of the information they already have, when it matters.
ANS
www.ans.co.uk
automationmagazine.co.uk
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40
