What is agentic AI and how is it being used in industrial operations?
Every morning, in warehouses, factory floors and distribution centres across the country, a production manager walks in an hour before their shift starts.
Their systems were running all night, recording what went through, what didn't, where the bottlenecks were and which orders slipped. It tracked and logged everything. It just couldn't tell anyone what to do about it.
The manager comes in early to decipher the data and decide what the day looks like before their team arrives on site. That gap – between what operational tools know and putting it into action – is where most industrial businesses have been living for years. And the numbers reflect it.
Our own research revealed 49% of manufacturers aren't realising major operational efficiencies.
Not because the data isn't there, but because their technology isn’t designed to do something with it.
From ‘systems of record’ to ‘systems of action’
Operational systems are extraordinary recorders of truth. Every transaction, every inventory movement, every supplier invoice, every demand signal – all of it captured, timestamped and securely stored.
For businesses that have been running on these systems for ten, fifteen, twenty years or more, the depth of that operational data is remarkable. But recording something and acting on it are two entirely different processes.
Traditional business software tells you what happened, but it doesn't tell you what to do in response. And it doesn't do anything on its own. The job of interpreting data trends, identifying what matters and deciding what action to take has always fallen to a production manager or operations team, usually working off last week’s export.
Decades of operational data. Finally put to work.
AI agents built for industrial operations change all of that. These aren’t chatbots, LLMs or basic automation rules. Agents read a situation, determine optimal actions and either take or deliver the decision to whoever needs to make it, before the window to do anything useful has closed.
A demand signal spotted three days too late isn’t useful to anyone. It’s just a record of what you should have done earlier. In the same way you can’t frame a stock position that tips from slow-moving to written-off while it waits on someone's weekly report as a data problem.
The information was always there, it just never turned into an action quickly enough. Agentic AI closes that gap by acting on operational data faster than any team can manually manage.
Research from Deloitte shows that manufacturers implementing ‘smart factory’ workflows saw a 10–20% improvement in production output, a 7–20% improvement in employee productivity, and 10–15% in unlocked capacity.
What is agentic AI?
Agentic AI doesn't just analyse data. It identifies what matters and recommends or takes action based on business context.
Turning written-off stock into revenue
Picture a manufacturer sitting on a warehouse full of stock nobody is buying. Their products have been deprecated, written off the balance sheet and about to be filed under ‘dead stock.’ Rather than accept the loss, the production manager makes a decision to do something that would have previously taken 30 days of consulting work and significant overhead to even attempt.
Using AI-powered inventory management, every single obsolete SKU is cross-referenced against Google Trends. Product descriptions are matched to real-time regional demand data. Marketing campaign ideas are automatically generated, not just for the product range as a whole but also each individual SKU. These are then targeted to the regions where they are most likely to sell.
Rather than a month, the whole process takes a day and a half. The stock that had been written off as pure loss generates £100,000 of additional revenue. Not a recovered margin or a cost reduction. New revenue, from inventory that was already written off on paper.
What makes this significant isn't just the figures. The data needed to do it – the SKU records, the product descriptions, the inventory positions – was already there. Agentic AI simply acted on information that already existed, faster than any team could have done, to find value that would otherwise have stayed invisible.
The value hidden in your data
The dead stock example is an illustration of what becomes possible when a system that has spent years recording operational reality is finally able to plan and optimise its own next steps.
For decades, ERP systems have been the backbone of industrial operations. They are living records of how a business actually runs, capturing every supplier relationship, every inventory decision and every demand pattern. In most cases, these records run incredibly deep.
Generic AI tools can't access or process the whole picture. An LLM pointed at a spreadsheet export is working with a fraction of available data. Now contrast that with agentic AI built on top of your existing software that works with everything, including your full operational history and the context in which every decision has ever been made.
The difference is stark. It’s also why established manufacturers and distributors that run on reliable systems and detailed records spanning years have a significant advantage that no new entrant or generic AI can replicate. In the past, that advantage has been tough to maximise. But with agentic AI, it finally becomes actionable.
Gartner predicts that half of all supply chain management solutions will include agentic AI capabilities by 2030, while 40% of enterprise applications will be integrated with task-specific AI agents by the end of 2026, up from less than 5% today. The window to move first is open, but it won't stay that way.
The people behind the systems
It's easy to talk about agentic AI in terms of revenue generated and hours saved. The numbers are compelling enough to stand on their own. But the more significant shift is the one that doesn't show up on first glance of an ROI calculation.
That hour the production manager spends every morning reading shift data, cross-referencing reports and building a picture of the day ahead is fundamental work. The inefficiency is that they are the ones doing it. The system they rely on has always required a human in the middle to make sense of what it knows.
It's a process problem – one that has defined how industrial businesses are staffed, structured and stretched for decades.
Pulling data and generating reports isn’t where experienced managers add value. And every hour spent on it is an hour not spent on the decisions that actually require human expertise. The supplier relationships that need more attention. The production schedule that needs rethinking. The lead times that are creeping up and starting to hurt.
A different kind of morning
Agentic AI clears the path to better, quicker judgment. The manual interpretation, the early starts, the cross-referencing across systems that don't talk to each other are the tasks that existed only because the technology couldn't handle them. Now it can.
For the organisations that get this right, the impact is felt across how their people work, what they're able to focus on and how quickly the operator can respond when something changes. So when that same production manager walks into a briefing that's already been prepared, reviewed and flagged for the decisions that actually need them, suddenly it’s a different kind of morning.
And before long, a different kind of business.
Sitting on years of operational data you’ve never fully acted on? Now you can.
Manufacturers, distributors and craftspeople moving first on agentic AI are making decisions faster, finding revenue others miss and responding to change before their competitors have even spotted it.
The difference usually comes down to having the right technology in place to do the heavy lifting for you. Demand signals. Inventory positions. Production bottlenecks. Obsolete stock. Supplier lead times. The data is there, it just needs to be turned into something usable, and quickly.
With agentic AI built on top of your existing systems, Forterro transforms years of operational data into a system that records what happened and then tells you what to do next. No extra admin or early starts to make sense of last night's reports – just faster, smarter decisions at every level of your operation.