AI has introduced new tools for seeing what's actually happening on the shelf, tools that didn't exist even five years ago, and the economics finally justify using them.

Retail wages are up 52% over the past decade while the cost of tools such as electronic shelf labels has dropped 67% over the same stretch, and AI adoption in retail operations has gone from under 1% of retailers a decade ago to 69% today.

CPGs and retailers alike are piloting shelf monitoring cameras, predictive replenishment, and AI flagged phantom inventory detection in response to a problem that has quietly persisted for decades: on-shelf availability.

Yet despite this surge, few organizations are capturing that value at scale. A recent survey of nearly 40 senior CPG and retail executives found that 75% of CPGs remain stuck in pilots and exploration, with only 18% scaling meaningful impact. Retailers are more bifurcated: 45% are scaling real impact, while a nearly equal share, 40%, have barely begun.

On-shelf availability remains one of the least measured, least owned metrics in the entire end-to-end value chain.

State of play: real value, unevenly captured

For years, CPGs and retailers have managed the shelf indirectly through service level and case fill metrics that effectively stop at the distribution center door. AI is now making it possible to measure and act on what happens after that, at the individual store and SKU level, in close to real time, while prioritizing the gaps that matter most.

But the maturity of adoption tells a very different story from the maturity of the technology. Most organizations remain stuck running shelf monitoring pilots in a handful of stores, celebrating promising results, and then struggling to translate those results into a scalable operating model.

Across the industry, on-shelf availability still averages around 92%. In other words, roughly one out of every twelve times a shopper reaches for a product, it simply is not there. Despite a decade of investment and innovation aimed at solving the problem, that number has barely moved.

Part of the challenge is measurement itself. More than half of executives surveyed say they do not measure the ROI of their AI investments in this space at all. That makes it nearly impossible to determine whether a successful pilot is genuinely worth scaling or simply generating activity. Then there is the ownership problem. Supply chain teams own fill rate. Retail execution teams own the physical shelf. Yet almost no one owns the gap between the two.

Questions that matter today

Addressing the following questions can help CPG leaders pressure test whether their shelf availability efforts are built to create a lasting competitive advantage.

1. If we had to report our on-shelf availability tomorrow, rather than our service level or case fill rate, could we? And would we trust the number?
2. Are our OSA goals truly ambitious, or are we simply aiming for incremental improvement while remaining at the industry average?
3. When a shelf goes empty, who is responsible for knowing about the issue, who is responsible for fixing it, and who is accountable for measuring the ROI?
4. Are we trying to solve OSA alone, or are we building the shared data, workflows, and retailer partnerships required to solve it at scale?
5. Do we know exactly how much poor on-shelf availability is costing us in lost sales, and what we stand to gain by improving it?

Considerations to turn shelf availability into a competitive advantage

Corresponding to the questions above, here are five considerations can help translate intent into results.

1. Make on-shelf availability a leadership metric. What gets measured at the top is more likely to get fixed on the floor.
2. Set an ambitious target and connect it to revenue. Aim meaningfully above the industry average and clearly define what achieving that improvement is worth to your business.
3. Start small, then scale. Focus on a limited number of priority categories or regions, redesign execution end to end, and prove the model before scaling.
4. Create clear accountability for the DC to shelf gap. Assign a single owner and align supply chain, sales, and retail execution around the shared outcome.
5. Build measurement and transparency in from day one. Track the investment behind every improvement against the revenue it recovers, and ensure retail partners and field teams understand what is changing and why.

For CPG leaders, the next step is straightforward to state and much harder to execute: stop treating on-shelf availability as a supply chain footnote and start treating it as a commercial priority, with a clear owner, an ambitious target, and its own place in the P&L conversation.

Sources: AI in CPG and Retail: How Winners Are Pulling Ahead, Board Brief PDF: State of AI for CPG, Press release: CPG and Retail Leaders Are Bullish on AI, yet Most Haven't Scaled It Where It Matters Most