Key takeaways
- AI spending in retail is rising fast: 97% of retailers plan to increase budgets this year, but only 6% of organizations are generating what McKinsey defines as genuine enterprise-level value from AI.
- The barrier is organizational, not technical. BCG’s research puts 70% of AI value in people, processes, and operating model — the areas most retailers have left largely untouched
- The companies where AI is working made one structural decision differently: they stopped layering AI onto existing workflows and redesigned those workflows around AI.
Retail has never spent more on AI. The numbers don’t lie. Deloitte’s 2026 survey of 200 retail executives found that 82% plan to increase AI investment in the next 12 months. AI spending in the sector is projected to reach $19.9 billion globally this year, up from $6.4 billion in 2021. NVIDIA’s 2026 State of AI survey, which polled over 3,200 respondents across industries, found that 97% of retailers plan to increase their AI spending in the next fiscal year. By any measure of investment intent, AI adoption in retail has become a near-universal commitment.
The returns, however, are a different matter.
AI adoption in retail is outpacing its own returns
The most instructive document on this subject right now comes from Bain. Its 2026 Automation and AI Pathfinder Survey, conducted with 951 respondents, has a fitting, explanatory title: “Your AI budget is growing. Your returns aren’t.” The numbers are specific: 37% of companies targeted cost reductions of 11–20% from AI investments. Nearly 40% of those that measured actual outcomes landed below 10% instead. The technology worked. The savings didn’t arrive. And 90% of those same companies are now increasing their budgets again.
Deloitte’s findings reinforce this. About half of retail executives invest less than 0.5% of revenue in AI, despite categorizing it as a strategic priority. Enterprise-wide deployment sits at between 7% and 10%. The conclusion is plain: retailers are piloting broadly but scaling almost nothing.
A study by consultancy firm valantic conducted in partnership with the Handelsblatt Research Institute, surveyed 1,000 business decision-makers including more than 100 retail executives. The finding? While 45% of retailers describe themselves as AI pioneers, only one in three achieves measurable economic value from those investments. McKinsey’s State of AI 2025, drawing on 1,993 respondents across 105 countries, puts the high-performer share at just 6% — the fraction of organizations where AI is genuinely moving the bottom line. The pattern holds across every major piece of research on this subject. AI spending is rising. Enterprise-scale results are not.
Technology is not the problem
This is where most diagnoses go wrong, and where retail is particularly exposed.
When AI tools for business fail to deliver at scale, the instinct is to assume the tools need upgrading or that the data infrastructure wasn’t ready. These are real problems, but they are symptoms rather than causes. The consistent finding across Bain, Deloitte, BCG, and McKinsey is that the barrier is in fact organizational and not technical.
Case in point: BCG has quantified where real AI value comes from. Its research puts 10% in the algorithm itself, 20% in the technology required to implement it, and 70% in the people, processes, and operating model built around it. Most retailers are investing heavily in the 30% while leaving the 70% largely untouched.
Bain identifies data access and integration as the single biggest barrier to AI progress, cited by 41% of respondents and ranking above budget constraints and skills gaps. It draws a sharp distinction between organizations that treat this as an IT problem and those that treat it as a board-level business prerequisite. The ones delivering on their targets belong to the second group. Stanislas Vignon, Head of Insights at LVMH, put it plainly in the IBM Institute for Business Value’s 2026 retail study: “AI is not a magic wand. If you don’t have the right data, it doesn’t work.’
Where AI is really working
McKinsey’s research into AI-native companies, — businesses built around AI from the ground up, rather than retrofitted to accommodate it — offers the clearest illustration of what a structurally different approach looks like.
In one case, a 20-person agriculture technology company has paused all external hiring because AI now handles more than half of tasks across nearly every business function. The people it employs focus on scientific judgment and partner relationships. At a DevSecOps platform, non-technical staff fix software bugs and rename product features without involving the engineering team at all.
What do these companies have in common? They made the structural decision to use AI as the operating layer the business runs on.
Among McKinsey’s 6% of genuine high performers, the distinguishing factor is not better technology, it is workflow redesign. High performers are more than three times as likely to be pursuing transformational, enterprise-level change rather than incremental efficiency gains in AI in retail. They stopped asking what AI could do for their existing processes and started asking which processes still needed to exist.
Where this leaves retail
The valantic/Handelsblatt study includes one figure that deserves more attention than it typically receives: 83% of retail executives believe the current AI investment boom will end before 2030, mainly because most companies will fail to generate real business value from it. The same study finds that 80% believe they will lose competitiveness entirely without it. What they are skeptical about is whether most organizations have the appetite for what generating that value requires.
That skepticism is well-founded. The current model — buy the tools, run the pilots, report deployment as progress — will not hold. The evidence from Bain, Deloitte, McKinsey, and BCG is consistent on this: AI spending is rising while enterprise-scale results remain the exception. The market is beginning to understand that the gap between the two is not a technology problem waiting to be solved by the next release.
Retail has so far treated AI adoption as a procurement decision. The companies on the right side of the returns gap treated it as something closer to an organizational redesign. Those are not the same undertaking, and the difference in outcomes is showing.
