Feature: Edge AI


MCU or MPU calls for Small Language Models (SLMs) and severe quantisation techniques. Tese optimisations compress reasoning capabilities into footprints that fit within limited memories, making ‘intelligence’ a local resource rather than a remote service. And although agentic AI brings great potential to embedded systems, for engineering teams it also brings significant practical challenges. Making these systems work successfully depends on architectural choices, trade-offs and deep understanding of how autonomous behaviour interacts with compute, timing and safety constraints. Tis article treats agentic AI as an engineering discipline, shaped by deployment constraints, system design decisions and trade-offs.


From reactive to autonomous Traditional AI systems are stateless responders. Given a prompt or input, they produce an output without internal persistence or long- term objectives. An LLM might summarise a document or answer a question but does not incrementally pursue a goal or alter its strategy over multiple steps. Tis works well for discrete tasks but falls short when systems must operate continuously and interpret feedback without explicit instructions. In contrast, an agentic system can assess its environment and


Introducing agentic AI in embedded systems By Michaël Uyttersprot,


Market Segment Manager for AI, ML and vision, Avnet Silica


to a cue and produce a single output. Te next phase of AI focuses on building systems that can reason, plan and act autonomously – an approach known as “agentic AI”. While LLMs have popularised agentic concepts, deploying them at the edge requires a shiſt in scale. Moving from the cloud to an


A 18 July/August 2026 www.electronicsworld.co.uk


rtificial intelligence (AI) has evolved from rule-based systems to deep learning and large language models (LLMs). Today’s generative AI can draſt prose, translate languages and create media in response to prompts. But at their core these models remain reactive – i.e., they respond


objectives before acting. It then plans sequences of actions that move it toward the goal, continually revising those plans as new information arrives, and acts upon them. Tus, agentic AI systems are characterised by organised decision making, operating autonomously toward higher-level goals. Goals can be hierarchical, with high-level objectives decomposed into intermediate steps and constrained by operational limits such as safety, latency or power consumption. In embedded and edge systems, this structure provides a way of balancing competing requirements whilst maintaining predictable behaviour. For example, in robotics, autonomous vehicles, industrial


automation and intelligent IoT devices, systems must interpret sensor streams, integrate multi-modal data and make timely decisions under resource constraints. Here a reactive system might trigger a hard emergency stop upon detecting any vibration anomaly, whereas an agentic sensor can perform a cost-benefit reasoning task. It will evaluate the severity of the fault against the current production state, perhaps determining to complete the current high-value batch at a reduced speed before initiating a controlled shutdown, thus preventing both equipment damage and unnecessary material waste.


Deployment constraints at the edge In embedded and edge applications, agentic AI operates as a decision-making component within a larger system. It maintains state, evaluates objectives and issues actions over time. Like other subsystems, it is subject to constraints around timing, resource usage and safety, which shape how much autonomy is feasible practically. One of the most significant constraints are compute and


power. Agentic AI systems oſten require sustained processing for reasoning and planning, which can be difficult to support on


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  |  Page 41  |  Page 42  |  Page 43  |  Page 44