A conversation with Matt Bertucci of Lenovo.
For years, artificial intelligence in retail was a headline feature without a business case. At the National Retail Federation (NRF) show three years ago, “AI-powered” was the industry’s favorite sticker slapped on everything from chatbots to planograms but often without a clear link to measurable impact. The result? Retail executives walked away intrigued, but not necessarily convinced.
That dynamic has shifted.
In a recent conversation with Matt Bertucci, Lenovo’s Director of Retail Solutions for North America, Latin America, and, and EMEA, one theme emerged with clarity: the retail AI story has moved from novelty to necessity. We are now in what industry insiders are calling the agentic AI era – systems that don’t just respond to prompts, but learn, refine, and self-optimize in near real time. This is where AI starts behaving less like a static tool and more like a dynamic co-worker.
From Curiosity to Competence
Looking back, the timeline is telling:
- Year One: AI was the “shiny object” of NRF. Vendors showcased it for optics, with little proof it could drive operational or financial results.
- Year Two: Generative AI and large language models started producing genuinely useful pilots – natural language search, product recommendations, and basic content generation began to find their way into store and e-commerce environments.
- Year Three (Now): AI agents capable of iterative improvement – asking follow-up questions, refining outputs, and delivering outcome-specific solutions – are moving into production.
The key shift is that deployments are increasingly tied to hard ROI metrics: higher basket sizes, faster transactions, reduced returns, improved staff productivity. AI is no longer a speculative investment – it’s a performance lever.
Augmentation Over Automation
The retail AI debate has often been clouded by fear: Will this replace jobs? The reality playing out is far more pragmatic. AI is taking on the repetitive, time-intensive, or customer-avoidant tasks – the things that drain labor without adding much value.
That frees human associates to focus on higher-value, emotionally intelligent interactions. In Bertucci’s words, “It’s about augmenting what’s being done for multiple reasons and being able to, as a store, run efficiently and at lower cost, because your margins are always going to be razor thin.”
In other words, AI in retail works best when it’s a force multiplier for people – not a substitute.
Case Study: Digital Signage as a Revenue Engine
One of the clearest examples is in digital signage. For years, the move from paper to screens was a slow crawl – expensive to implement, hard to keep relevant, and easy to ignore.
AI changed the equation.
With dynamic content triggers – from customer proximity detection to time-of-day product rotation – signage is now personal, contextual, and tied directly to conversion events. A customer walking past a screen might see an offer shift in real time based on what’s in their loyalty profile or current weather patterns.
This isn’t just “nice to have” creative. Retailers can now attribute measurable lift to these moments, turning what was once an expense line item into an ROI-positive engagement asset.
Returns: From Cost Center to Conversion Engine
Returns remain one of retail’s biggest profit leaks, especially in e-commerce. By the time you account for reverse logistics, restocking, and markdowns, returns can quietly erase much of a retailer’s margin.
Lenovo’s AI-powered returns kiosks are designed to turn that weakness into a strength. Here’s how:
- AI-assisted online initiation: When customers start a return online, the system asks reason-specific questions (size, fit, color) to both speed the process and collect actionable product data.
- QR-based store interaction: Customers receive a scannable code to use at an in-store kiosk.
- On-site replacement or upsell: The kiosk can direct customers to where a replacement item is already waiting or suggest complementary items with a discount incentive.
- Automated triage: Returned items are instantly sorted for resale, repackaging, or write-off.
The economics here are compelling. Studies show more than 50% of customers who enter a store to make a return end up buying something else. Streamlining the process reduces frustration, increases dwell time, and raises the odds of a new purchase. Even a small uptick in that secondary basket can translate into significant incremental revenue at scale.
The Infrastructure Layer: AI Factories
All of this – from dynamic signage to real-time returns processing – depends on heavy compute power. That’s where Lenovo’s “AI factory” concept comes in.
Think of it as a modular architecture: GPU towers handle the high-intensity tasks of large language and visual models, while CPU towers manage orchestration and integration into store systems. The deployment model is flexible – on-premises for latency-sensitive workloads, cloud for scale, or hybrid for both.
For retailers, especially mid-tier players without enterprise IT budgets, this matters. It means AI at scale is no longer reserved for the big-box elite.
The Next 12 Months: Conversational Retail
If the last three years have been about proving AI’s value in specific tasks, the next year will be about deepening the quality of interaction. Bertucci predicts the next wave will focus on making AI conversations more human, more natural, and more linguistically flexible.
That means:
- Multilingual fluency without translation lag.
- Context continuity, where AI remembers a customer’s last interaction and builds on it.
- Ambient assistance, where help is available without explicit activation from fitting room mirrors to product displays.
These aren’t incremental tweaks, they’re experience shifts that could redefine how customers perceive service, both online and in-store.
Analyst Take: What This Means for Retail Leaders
- Measure for outcomes, not activity. AI projects should be tied to specific, revenue-linked KPIs from the start – basket size, conversion rates, transaction speed, inventory turns. Pilots that don’t measure impact risk being cut before they scale.
- Design for augmentation. The most effective AI deployments don’t remove staff; they reallocate human capacity toward high-touch, high-value customer interactions.
- Own the infrastructure conversation. Whether cloud, on-prem, or hybrid, the compute strategy needs to be part of the executive agenda. Without it, AI ambitions will stall.
- Focus on hidden profit pools. Returns, markdown optimization, and cross-selling are ripe for AI’s data-driven precision. Start where the margin leaks are.
- Think customer-centric, not feature-centric. AI’s real value is in removing friction and adding relevance. The technology is the means – not the end.
Bottom Line:
AI in retail has crossed a critical threshold. It’s no longer about proving the concept – it’s about scaling the wins. Retailers who treat AI as an experience driver and an efficiency engine will have an edge in a market where margins will always be thin and customer expectations will only rise.
As the next chapter unfolds, the leaders won’t be those who simply deploy AI. They’ll be the ones who integrate it so seamlessly into the retail environment that customers stop thinking of it as technology – and start experiencing it as service.
