Even last year, the term ‘agentic AI’ would have been unfamiliar to many. But the technology is starting to gain real visibility and traction this year. In sectors like retail and manufacturing in particular, which involve complex supply chain and product pricing decisions, it’s already well placed to provide immediate impact. But what exactly is it?
Agentic AI builds on the recent breakthroughs in large language models (LLMs) like ChatGPT and decision-making frameworks. The key differentiator between agentic AI and traditional AI systems is that agentic AI can act autonomously in real-world systems.
A traditional AI model, for example, responds to instructions or answers questions one at a time. It can analyse a range of real-time and historical data to produce forecasts and detect patterns – an optimisation engine then uses business logic to offer recommendations like optimal stock levels or pricing suggestions. Once this takes place, a supply chain manager can review these suggestions to inform their decisions and then handle the logistics themselves.
But AI agents, the models which make up agentic AI, can now perform this latter step independently. The LLM acts as the brain and allows them to plan tasks in relation to a retailer’s profile and then select the best tools to execute recommended actions without manual intervention.
In a retail context, for example, this means AI agents can automatically adjust inventory levels or pricing markdowns and then place replenishment orders themselves, rather than simply suggesting the best course of action to take. Impressively, they are able to adapt in real time to market conditions and learn from the actions they take.
The arrival of agentic AI fits a broader shift in enterprise software: automation moving from the back office to the front lines. It’s a groundbreaking step that will transform operations in retail and manufacturing. But what exactly does this look like in practice?
Smarter merchandising and dynamic pricing
Two key areas where agentic AI can make a major impact in retail is with merchandising and commercial pricing. The first area, merchandising, centres on optimising the entire merchandising lifecycle, from the initial supplier, to the distribution centre, to the store.
AI agents can use recommendations for metrics like buying, pricing and markdowns to negotiate with suppliers and place purchase orders. They can then oversee existing orders and make any amends, while also ensuring optimal stock levels by automatically replenishing stores. The AI agents can even handle promotional campaigns from their inception all the way to their execution.
Human oversight is required to ensure recommendations are accurate and the right actions are taking place. But AI agents can significantly help teams to move faster, test more and reduce the manual effort behind critical decisions. This enables retailers to respond to market trends with speed and agility.
I work with a company that provides equipment rental and support services. It relies on an excellent customer experience – the right products have to be available wherever and whenever customers ask for them. Not only that, but their customer promise means they need to be delivered or collected within four hours.
But with 3,500 products across 200 depots, and with products varying from a simple cut-off saw to lighting towers, it has a complex network to manage. The process of forecasting demand is not just a one-off sale: teams have to take into account how long the tool will be hired for and its anticipated condition on its return.
So, its supply chain teams already use AI to access demand forecasts for each of their products and recommendations for how they should be distributed across the depot network, when to replenish products, and by how much. Consequently, they’ve been able to improve their inventory savings while satisfying more demand too.
Now, however, the company’s CEO, George Foster-Jones, has explained how “advances in agentic AI” will allow them to take the AI optimisation of their operational performance and commercial activities “even further”. Ultimately, his “vision is to have specialised AIs” working alongside his commercial teams “to optimise every price we set”.
This brings us to the second area: dynamic pricing. Any pricing decision, be it list prices or real-time quotes, can be managed with agentic AI.
Here, by balancing revenue, margin and inventory goals, AI agents can automatically execute pricing decisions and support commercial teams. For example, not only can they automatically apply optimal pricing to quotes through an existing CRM system, but they can also offer instant pricing guidance to customer service teams and customers and assist with live negotiations by suggesting optimal price points.
In such a margin-sensitive and volatile environment, this dynamic pricing at scale boosts competitiveness. The speed and accuracy of agentic AI decision-making and actions (which provide the best trade-off between stock levels and pricing in relation to demand) will place retailers using AI agents far ahead of those who aren’t.
Supply chain autonomy: building resilience amid volatility
Inventory management is the cornerstone of supply chain operations. The aim of agentic inventory management is to continuously optimise decisions across procurement, production and fulfilment. By using data analysis from each stock-keeping unit, location and process, the technology can drive smarter and faster decisions across the supply chain.
Picture AI agents running scenario simulations to try out various supply chain strategies. Like with merchandising, they can independently negotiate with suppliers, place, manage and adjust orders, and automatically replenish stock at different locations. All of these actions take place within guardrails chosen by humans and supply chain managers – any exceptions will trigger an alert to the user.
Take another example of a health and wellness company I work with. It oversees the logistics for snacks, beverages and meals from its portfolio of brands. One of its main challenges is knowing how inventory decisions impact service levels and on-shelf availability. So, again, its supply chain teams use AI to optimise the flow of goods across procurement, production and fulfilment.
Yet the VP of International Supply Chain at the company, James Cranfield, has expressed how agentic inventory management will enable them “to anticipate and respond to demand volatility more effectively”, ensuring their “customers get the products they love while reducing inefficiencies across our operations”.
But why is the transition to agentic AI actually needed?
Why this matters
Many businesses are still grappling with implementing AI into their operations and can be unsure over how to extract the best value from the technology. Therefore, the idea of trying to introduce and understand agentic AI seems even more complex – there’s a lot of hype around the term which can blur its practical value. But a range of external factors are making the adoption of technologies like agentic AI not simply a useful business asset, but an essential one.
Economic headwinds mean companies are expected to achieve more with less; costs are still rising and margins are shrinking. Supply chains are becoming ever more complex – more suppliers, trade routes, digital platforms – and facing greater risks – cyberattacks, extreme weather events. Alongside this, consumer expectations are rising, with people accustomed to more choice and timely deliveries, and consumer demand remains highly unpredictable.
With so much uncertainty and market pressures, the efficiencies and optimal decision-making unlocked through AI agents could be transformative. There has never been such a need for real-time adaptability and insights.
But for many companies, key commercial decisions – knowing what to order, where to house stock, how much to sell it for – are still dependent on disconnected systems and dashboards, even spreadsheets and gut feel. Companies reliant on these processes could be losing millions in lost profits through inefficient inventory management and sub-optimal pricing.
There are also well-publicised talent shortages in analytics and operations roles. A 2025 survey put ‘IT & data’ and ‘operations/logistics’ as the top two skills organisations are having the most difficulty in finding. Agentic AI can fill this skills gap in part, easing the burden on supply chain professionals and helping them to work more efficiently and productively.
Outpacing the competition
Agentic AI addresses a real gap: scaling commercial decisions at pace, without scaling headcount. The fact that AI agents can take actions on behalf of humans is revolutionary. Supply chain professionals can then focus on high-level strategy and commercial decisions, instead of getting bogged down in making inventory and pricing decisions.
Above all, the technology can build organisational agility and enable faster adaptation to market shifts. As enterprises brace for an increasingly uncertain future, agentic AI could mark the difference between reacting late and acting first. The retailers and manufacturers ready to hand over the reins may find themselves not just surviving, but outpacing the competition.
Author
Richard Potter, CEO and co-founder at Peak.
Peak’s AI optimises inventories and pricing for global industry leaders including Nike, The Body Shop, Marshalls and Eurocell. With a core belief that businesses need their own AI – built for their business, with their data – Peak’s pre-built AI products can be configured to fit unique requirements.
