As AI hype matures into real deployment, Infios is focusing on embedded agents that augment human decision-making and close the gap between insight and action.
It was no surprise that a significant number of the many conversations I had at LogiMAT referred to AI in one sense or another. It’s clear that AI is no longer a future ambition but is a present-day expectation. However, beneath the noise a more nuanced shift is taking place. The conversation is moving beyond what AI can do, towards what it is delivering inside real supply chain operations.
For Aadil Kazmi, Head of AI Product Development at Infios, that distinction is critical. Speaking to me during a busy day on the company’s stand, he described a market that is not just curious about AI, but increasingly focused on practical outcomes.
“Everyone is definitely interested in AI,” he said. “The conversations we’re having are really around two themes: how we’re bringing AI into our products and services, and how those capabilities actually empower people rather than replace them.”
Embedded, Not Imposed
Infios’ approach to AI is deliberately grounded in how supply chain teams already work. Rather than introducing standalone tools or requiring new workflows, the company is focusing on what it calls ‘embedded AI’.
“Embedded AI really just means that we build agents that work where people are already used to working,” Kazmi (pictured below with Peter) explained. “So people don’t have to learn a new screen or change their workflows – the agents meet them where they already work.”
This matters more than it might first appear. In an industry where operational systems are deeply entrenched and change management is notoriously complex, reducing friction is often the difference between adoption and abandonment. Alongside this, Infios is positioning AI firmly as a support mechanism rather than a replacement for human expertise.
“Our agents aren’t looking to automate the full human,” Kazmi said. “Quite the opposite. We take a very ‘human-in-the-loop’ approach, where the person is always in control.”
In practice, that means users can delegate specific tasks to AI agents while retaining oversight. If the system encounters an unfamiliar scenario or crosses predefined thresholds, it escalates back to a human operator – much like an exception in a warehouse or parcel sorting flow being diverted for manual handling or checking.
From Chatbots to Agents
Part of the challenge in these conversations is language. Terms like ‘AI’, ‘LLMs’ and ‘agents’ are often used interchangeably, creating confusion for end users. I’m as guilty as the next person when it comes to this. But Kazmi helps me to clear this up by drawing a clear distinction: “In the beginning, we had simple LLMs delivered as chatbots,” he said. “You ask a question; it gives you an output. It’s not doing anything more than that.”
These early systems were limited to their training data and unable to interact with live operational environments. AI agents represent the next step. “An agent is really an LLM with access to tools and reasoning loops,” he explained. “Tools allow it to interact with the real world – fetching order statuses, triggering workflows, even communicating externally. And reasoning loops allow it to gather information, reassess, and improve its output.” In other words, agents are not just answering questions – they are participating in processes.
Three Tangible Benefits
While the technology is evolving rapidly, Kazmi is clear that customer interest ultimately comes down to measurable outcomes. Based on current deployments, he identifies three core benefits.
The first is increased productivity or capacity. “A good example is our driver check call agent. Many customers weren’t able to perform check calls across all loads because it’s too expensive and time-consuming. With AI agents, they can now achieve full coverage.”
The second is cost reduction. “One customer deployed our order entry agent, and it processes inbound orders in seconds versus what took upwards of 10 minutes manually,” he noted. “So, the cost of executing that workflow fell enormously.”
The third is improved and more consistent customer experience. “Now that these agents are handling check calls, order entries, reporting and even catching exceptions before they occur, all of that translates into a more reliable experience for the end customer.”
Taken together, these benefits point to a broader shift: AI is not just optimising isolated tasks but reshaping the economics and expectations of supply chain operations.
Disruption Advantage
Supply chains have always been vulnerable to disruption, but recent years have amplified both the frequency and the impact of unexpected events – from geopolitical tensions to infrastructure blockages. In this context, AI’s ability to compress response times is becoming a defining advantage.
“What AI has been able to do is supercharge teams’ abilities to sense disruptions, decide what to do, and act in real time,” Kazmi adds.
Where responses once took hours or even days, AI-enabled systems can now react almost immediately. However, he is quick to point out that this capability depends on a critical foundation. “The precursor to all this is integrating agents into the systems where your data lives, and deploying them safely and reliably.” Once that foundation is in place, organisations can move towards a more autonomous model of operations: detecting disruptions as they emerge, evaluating possible responses, and executing actions within defined guardrails.
As Kazmi puts it, AI may not be able to prevent a crisis such as the blockage of the Suez Canal, but it can fundamentally improve how companies respond. “It allows us to bring visibility and transparency into the promises that we make every single day.”
From Visibility to Execution
For much of the past decade, visibility has been the dominant goal in supply chain technology. Knowing where goods are, and what is happening across the network, was seen as the key to better performance. That goal, Kazmi argues, has largely been achieved. “Visibility was probably one of the most important vectors to manage over the last decade,” he said. “Most organisations have already overcome that hurdle.” The new challenge is what comes next. “What good is visibility if you can’t act on it? That’s the gap we’re filling.”
This shift – from insight to execution – is emerging as a central theme across the industry. It also aligns closely with Infios’ positioning around ‘intelligent supply chain execution’, combining AI agents, machine learning models and optimisation tools into a unified operational layer.
AI as a Co-Worker
Despite the technological complexity, Kazmi emphasises that successful AI adoption is as much about people as it is about systems. “It’s very critical to think about how your agents will work with your existing team members,” he said.
Framing AI as a ‘co-worker’ is one way to approach this shift. Rather than viewing automation as a threat, organisations can position AI as an extension of their workforce – taking on repetitive or time-sensitive tasks while leaving humans to focus on higher-value decision-making. However, this requires careful planning around roles, responsibilities and trust.
“The onus is on us to figure out how to integrate these agents into our current workforce,” Kazmi stated. Equally important is choosing the right technology partner.
“Partnering with a vendor that integrates deeply into your stack is key. You need to bring the agents right to where your teams are already working to reduce the friction of change management.”
The Next Frontier: Context
Looking ahead, Kazmi believes the next major leap in enterprise AI will come from improvements in memory and context management. “While current deployments are delivering immediate ROI, the next jump will come from bringing deeper context to these agents,” he said.
Context, in this sense, goes beyond raw data. It encompasses the reasoning, priorities and situational awareness that underpin human decision-making. “As a human being, when I come to work in the morning, why am I making the decisions that I’m making? That’s super important.” Embedding that level of understanding into AI systems could unlock a new phase of capability – moving from reactive assistance to more proactive and strategic support.
Defining Moment
For an industry long defined by complexity and constraint, the convergence of AI, data and execution capability represents a significant inflection point. As Kazmi sees it, the opportunity is not just to do things faster or cheaper, but to fundamentally rethink how supply chains operate.
“The world is no longer predictable,” he said. “Execution is becoming trickier and trickier to manage.” In that environment, the winners will not be those with the most visibility, but those with the ability to act on it intelligently, consistently and at speed. Increasingly, that now means working not just with AI, but alongside it.
