Most AI conversations in retail focus on tools when the real question is around decision quality.
Despite widespread adoption of AI tools, many retailers are quietly disappointed with the results. In fact, a recent CEO survey by professional services network PwC found that more than half of respondents reported no financial return from AI investments. The problem is rarely the technology itself, it’s how AI is being used.
The newly published Retail Gazette BIG AI Report 2026 makes this clear. Throughout the report, Felix Hoffmann, CEO & Co‑founder of 7Learnings, shares expert commentary on where AI is delivering real value today and where expectations consistently outpace outcomes. The conclusion is simple: AI only creates value when it improves decisions.
Below are four core lessons retailers should take away.
1. AI is decision automation, not the automation of people
One of the most persistent myths around AI is that its value lies in replacing human work. In practice, the highest-impact use cases look very different, as AI works best when it automates decisions, not roles.
Retail is driven by thousands of micro-decisions every day: pricing adjustments, promotion timing, markdown depth, stock allocation, campaign spend. Historically, many of these decisions have been reactive, manual, or based on incomplete information.
Predictive AI changes that by allowing retailers to simulate outcomes before acting; forecasting how price changes, promotions, or inventory moves will affect margin, sell‑through, and cash. Humans still set strategy and guardrails, while AI provides foresight at a scale no team could manage manually.
When AI fails, it’s often because it’s layered on top of existing workflows without changing how decisions are actually made.
2. The real shift is from reacting to markets to target‑driven steering
Most retailers still operate reactively:
- Competitors change prices → we respond
- Stock builds up → we discount
- Demand drops → we push promotions
AI enables a fundamentally different approach: target‑driven steering. Instead of reacting to signals after the fact, retailers define the objective first, for example:
- Maximise profit
- Increase full‑price sell‑through
- Accelerate stock rotation
AI then calculates the best path to reach that target, continuously adjusting as conditions change. Think of it as Google Maps for retail decisions: the destination is fixed, the route adapts in real time.
This is where functions such as pricing become strategic rather than tactical. Prices are no longer adjusted based on market movements, but rather on the retailer’s understanding of how each decision will impact the outcome.
3. Most retailers need clearer use cases
A common pattern highlighted in the Retail Gazette report is pilot fatigue. Retailers experiment with multiple AI tools across departments, but struggle to scale any of them. The root cause is almost always unclear success criteria.
High‑performing retailers take a different approach:
- Start with one or two concrete use cases
- Tie them to measurable commercial KPIs
- Prove ROI early
- Scale what works
In pricing, this might mean focusing first on markdown timing or elasticity modelling rather than trying to optimise everything at once. In inventory, it could be improving forecast accuracy for a single category before rolling out more broadly.
4. As AI becomes agentic, trust becomes the new battleground
Looking ahead, the report also explores the rise of agentic AI, that is, systems that don’t just recommend actions, but execute them autonomously. As this trend accelerates, retailers won’t just compete for customer attention; they’ll compete for algorithmic trust.
AI agents comparing prices, availability, and delivery options on behalf of consumers will favour retailers that are predictable, transparent, and consistent. Decision quality will increasingly shape visibility and conversion. In this environment, poor decisions will be penalised by algorithms.
What it takes to make AI deliver real ROI
The central message of the Retail Gazette BIG AI Report 2026 is that the AI race is about being effective. Retailers that succeed with AI treat it as part of their operating rhythm. They invest in decision quality, define targets clearly, and use AI to design outcomes rather than chase them.
You can read the full Retail Gazette BIG AI Report 2026 here: https://www.retailgazette.co.uk/the-big-ai-report-2026/
A note from Felix: Why this report matters now
By Felix Hoffmann, CEO & Co-founder, 7Learnings
The real question around AI is decision quality. That’s the core idea I kept coming back to while contributing commentary throughout the Retail Gazette BIG AI Report 2026.
Over the past years, I’ve spoken to hundreds of pricing leaders, commercial directors, and CEOs. Almost all of them are experimenting with AI in some form, yet many are unsatisfied with the impact. This is because AI is too often applied without a clear decision problem in mind.
AI should replace uncertainty
Retail decisions have become too complex to manage reactively. Pricing, promotions, stock, and demand signals interact in ways that no spreadsheet or weekly meeting can fully capture.
What predictive AI enables is foresight: the ability to understand the impact of a decision before committing to it. That’s why I describe AI as decision automation, not people automation. Humans remain responsible for strategy, guardrails, and accountability. AI provides the probabilistic understanding that makes those decisions defensible.
From reacting to markets to steering outcomes
Many retailers still change prices because competitors have moved, or discount because stock is building up. These are reactions, not strategies.
Target-driven steering flips that logic. You define the objective first (margin, sell-through, stock rotation), and let AI continuously calculate the best route to get there as conditions change. As a result, pricing stops being a tactical response and becomes a strategic lever.
Why pilots fail and what works instead
A recurring theme in the report is pilot fatigue. AI projects launch with ambition, but stall when results are hard to prove or scale. In my experience, the retailers who succeed follow three simple rules:
- Start with one clearly defined decision
- Tie it to commercial KPIs, not technical metrics
- Prove value early, then expand
Better focus is almost always what is required.
Competing for algorithmic trust
As agentic AI becomes more common, retailers will increasingly be evaluated not just by customers, but by algorithms acting on their behalf. In that world, decision quality becomes visible. Inconsistent pricing, unclear availability, or erratic promotions will be penalised by systems optimising for trust and predictability.
This is why AI maturity is becoming a core part of brand perception.
Tools don’t matter if decisions don’t improve
The reason this report resonated with me is that it avoids extremes. AI is neither a silver bullet nor an existential threat; it’s a capability that rewards clarity and punishes confusion.
Retailers who treat AI as part of their operating rhythm make better decisions, faster, and with greater confidence. It’s a practical, grounded view of where AI delivers real value today and what it takes to get there.
