National Retail Federation (NRF) 2026 – ‘Retail’s Big Show’– made one thing clear: retail is racing towards full automation. Agentic AI was everywhere, promising autonomous replenishment, dynamic pricing, and self-optimising store operations. Retailers want systems that act, not just predict – but they’re going to need item-level intelligence to get them there.
The rise of agentic AI
The AI Stage was the heart of the show – a space where retailers demonstrated autonomous agents capable of managing complex workflows without human intervention. Deployed at scale, these agents will reduce friction, increase throughput, and maximise margins in environments where labour, supply, and demand are increasingly unpredictable.
However, this level of automation relies on total inventory precision. Retailers are no longer satisfied with tracking pallets; they want to see the journey of every single item – from the factory to the checkout. Yet current AI stacks typically have only partial visibility. In the absence of item-level data, models must still ‘guess’ rather than know. That gap was noted repeatedly across panels on retail media, supply chain, and loss prevention. And the solution is item-level intelligence at scale.
The need for item-level intelligence
While item-level tracking isn’t a totally new concept, it’s typically been restricted to high-value, luxury items. This is partly down to cost – but also down to the carbon footprint of the silicon chips that enable it.
The production of these chips requires silicon of ultra-high purity, which is energy-intensive to produce, and multiple process steps at extremely high temperatures – well beyond 1000°C. It also requires vast volumes of water for cleaning and rinsing, as well as ultra-pure water (UPW) for cooling and chemical dilution. UPW is thousands of times cleaner than drinking water, and its creation is inefficient, requiring anywhere from 1,400 to 1,600 gallons of municipal water to produce 1,000 gallons of UPW.
In short, silicon chips have an outsized carbon impact.
How should brands address the need for item-level data to further train emerging AI models against a backdrop of consumers who increasingly demand sustainability as standard? And, as legislation around Digital Product Passports (DPPs) expands, how can circular models sustainably extract the data they need to fulfil their goal?
This is where low-carbon, low-cost alternatives such as FlexICs become transformational – enabling item level intelligence at scale to finally makes mass-market adoption viable, without shattering sustainability targets. Check out their credentials in this report.
The importance of sustainability
The industry is steadily moving toward circular supply chains. In fact, sustainability was baked into almost every keynote. And this shift is not just for the planet and compliance, but for profit, too. While it’s easy to be cynical about sustainability as a lever for profit, when things are profitable, companies do more of it – which can only be a good thing.
DPPs provide every product with a digital ‘birth certificate’ – a comprehensive record of a product’s lifecycle from cradle to grave. That includes origin, composition, environmental impact, and end-of-life. DPPs dominated multiple discussions, especially in fashion, beauty, and consumer electronics, with the most compelling use cases centred on:
- validating authenticity;
- enabling resale and repair;
- informing responsible sourcing;
- guiding end-of-life decisions.
Retailers are no longer treating circularity as a compliance problem; it’s emerging as a value engine, creating a clear competitive advantage for early adopters. Resale and recommerce players noted that item-level traceability – in the guise of DPPs – unlocks margin opportunities in refurbishment and secondary markets, too.
Yet despite progress, most circular initiatives still struggle with how items are identified, tracked, recaptured, and routed back into reuse loops. The same data gap that limits agentic AI also limits circularity. Waste streams remain largely ‘blind’, making automation difficult and recovery rates low.
Item-level intelligence would be a game-changer, supercharging circular systems and accelerating progress towards a low-carbon economy.
Item-level data: a scalable advantage
The biggest takeaway from NRF 2026? From agentic AI to DPPs, retail is shifting to a model defined by item-level data. Retailers that embrace this shift will gain new levers for margin, loyalty, and sustainability, in a brave, new world where:
- intelligence is ambient;
- value is continuous;
- ownership is flexible;
- data moves with items across their lifecycle.
Total inventory precision is within reach. In a market defined by volatility, low-cost, low-carbon data at item level is the advantage that scales: realising the potential of agentic AI, powering autonomous decision-making and delivering outsized gains for early adopters.
Want to know what item-level intelligence can do for your retail systems? Get in touch
