The AI conversation in retail has reached a critical inflection point. Everyone claims to be “AI-powered,” but walk the stores or talk to planners, and you’ll find that not much has changed day-to-day. This gap between hype and implementation comes down to, not technology limitations, but execution strategy.
In a recent conversation on Omni Talk’s Retail Technology Spotlight Series, Mahesh Vijayaraghavan (VP of Retail & Consumer Goods at Experion Technologies) and Siraj Alimohamed (Global Head of Data & AI) cut through the noise with Chris to reveal what’s actually working in retail AI implementation, and why most organizations are approaching it backwards.
The State of AI in Retail: Overhyped and Underimplemented
Mahesh frames the current state bluntly: AI in retail is overhyped at the highest level but undervalued where it really matters. The hype isn’t the problem. The execution gap is. As he puts it, “If AI is still in the dashboard, it’s overhyped. But if it’s driving a decision at 6:00 a.m. in the morning before the store opens, then it’s kind of underappreciated.”
This distinction reveals the core challenge facing retailers. Board presentations celebrate AI initiatives, but the technology hasn’t penetrated the operational layer where it could transform forecasting, inventory decisions, shrink reduction, and pricing.
Siraj offers a slightly different perspective: the hype is actually appropriate because retail pushes the frontier of what technology delivers. The barrier to advanced analytics has dropped dramatically. You no longer need teams of PhDs to extract real-time insights or predict demand trends. The gap, to him, is in implementation readiness across data infrastructure, operating models, and organizational alignment.
The Pilot Fatigue Problem
Retailers are drowning in pilots. NRF showcases hundreds of proof-of-concepts, but how many scale beyond the initial test? Siraj acknowledges that pilots serve a purpose. Organizations need to learn before committing to multi-million dollar, multi-year contracts. But the industry has developed pilot fatigue, running endless tests without the infrastructure to scale successful ones or quickly kill unsuccessful experiments.
The solution?
Move from pilots to platforms. Experion’s approach centers on building AI-ready platforms first, then bringing use cases onto that foundation. This enables rapid testing, prioritization based on business value and data availability, and seamless scaling when initiatives prove their worth. Organizations that skip this step find themselves trapped in circular pilot programs, never building the muscle memory needed for sustained AI implementation.
Process Optimization Before Automation
When asked whether retailers should automate repetitive tasks or first question whether those tasks are even needed, Siraj lands firmly in the “question first” camp. Most organizations carry 30-40% operational fat, processes continued simply because “the guy before you did that.” Automating these inefficiencies just creates “automation chaos.”
His advice: “Stop, pause and think because you want to try and improve the process that you’re trying to do.” If a process previously required five people, that doesn’t mean you need five AI agents doing the same work. Focus on outcomes and explore how AI can achieve them more efficiently and creatively than the legacy approach.
This thinking challenges the common assumption that AI is transformative. The fundamental outcomes haven’t changed. Retailers still need to get products to shelves for customers to buy them. What changes is how efficiently and intelligently you achieve those outcomes. Re-engineering AI to mimic human processes backward-engineers innovation and restricts the creative potential of the technology.
Internal Efficiency vs. Consumer-Facing AI
Where should retailers start? Internal operations or customer-facing experiences? Siraj admits he “blows hot and cold” on this question but offers a strategic framework: pursue your biggest gains first, but don’t pocket the savings. Internal processes often yield the fastest efficiency improvements, but the biggest brand benefits come from consumer-facing applications.
His recommendation is to find internal efficiencies first, then reinvest those savings back into consumer experiences. This creates a flywheel where operational improvements fund customer-facing innovation, which in turn drives revenue that supports further optimization. Taking internal savings straight to the P&L misses the opportunity to compound AI benefits across the organization.
What Retailers Are Actually Asking For
Mahesh shares insights from recent client conversations. A Canadian retailer with multiple brands highlighted three priorities: personalization, agentic AI capabilities, and operational efficiency. But these high-level buckets mask a deeper challenge. Most retailers know what they want but aren’t sure where to begin.
On personalization, the questions have evolved. Retailers aren’t chasing hyper-personalization for its own sake anymore. They’re asking: “How do I personalize without blowing up my margins? How do I personalize across brands? How do I do it in real time?” Personalization is moving from marketing experiments to operational personalization across offers, assortments, and content.
For agentic AI, clients want systems that act within guardrails, not just recommend actions. This means agents that monitor inventory and trigger replenishment, flag anomalies before humans notice them, and drive decisions without constant human intervention.
On efficiency, retailers face tight margin pressure and ask: “How do I run a leaner program without degrading customer experience? Where can I bring automation that removes friction, not people?” The focus has shifted from cost-cutting to decision velocity. Making better decisions faster.
The Digital Experience Differentiator
As AI becomes table stakes, digital experience emerges as the key differentiator. But Mahesh warns this isn’t about prettier screens or more features. It’s about how intelligent, predictable, and respectful interactions feel. The retailers that stand out won’t be the ones shouting “AI-powered!” They’ll be the ones making experiences feel smart in ways that anticipate intent, remove friction, and adapt without being creepy.
Trust isn’t built through one magical AI moment. It’s built through consistent, explainable, predictable interactions, especially when things go wrong. If customers and employees can’t ask “why did this happen?” and get clear answers, AI breaks trust, and experience suffers for a long time.
Retention comes down to reducing cognitive load. For customers, this means fewer decisions and fewer surprises. For employees, it means less manual work, fewer transaction overrides, and more confidence in systems. The best digital experiences don’t demand attention. They earn it by making life easier.
Building AI as a Muscle, Not a Project
Mahesh closes with perhaps the most important insight: “AI isn’t a project. It’s like a muscle. You build it over time.” This framing challenges the project mentality that treats AI as a one-time implementation with a defined beginning and end.
Building AI muscle requires consistent effort across data infrastructure, organizational capabilities, process re-engineering, and cultural adaptation. It means establishing platforms that enable rapid experimentation, developing frameworks for prioritizing use cases based on business value, and creating feedback loops that turn learnings from successful and failed initiatives into organizational knowledge.
The best AI implementations will be the ones retailers never talk about, where intelligence is embedded so seamlessly into systems, applications, and experiences that it just works. Customers walk into stores or log into apps and simply find what they need, when they need it, without friction or frustration.
The Path Forward
For retail leaders navigating AI implementation, the message from Experion Technologies is clear: start with platforms, not pilots. Question processes before automating them. Find internal efficiencies first, then reinvest savings into customer experiences. Focus on outcomes, not technology showcases. And treat AI as a muscle you build over time, not a project with a finish line.
The retailers that win won’t be the ones with the most AI pilots or the loudest AI marketing. They’ll be the ones who’ve built the operational muscle to deploy AI where it matters, in the 6 a.m. inventory decisions, the real-time personalization engines, and the seamless digital experiences that customers trust and employees rely on.
If you enjoyed this article, and want to listen to Chris Walton’s entire discussion with Mahesh and Siraj, you can do so via the platform of your choice below:
Apple Podcasts | Spotify | SoundCloud | Amazon Music
Be careful out there,
– Chris, Anne, and the Omni Talk team
P.S. See our past 8 years of wonderful Spotlight Series podcast guests, featuring roughly 200 movers and shakers in retail, by clicking here
Music by hooksounds.com
*Sponsored Content*
Omni Talk® is the retail blog for retailers, written by retailers. Chris Walton and Anne Mezzenga founded Omni Talk® in 2017 and have quickly turned it into one of the fastest growing blogs in retail.