When generative AI went mainstream, it reached everyone at the same time — globally, indiscriminately, and without instructions. Most people did what felt natural: they asked it questions. And for a lot of them, that's where the story ended.
In this episode of #shifthappens, Venchi APAC CEO Marco Galimberti talks through what happens when organizations move AI past the search bar and what keeps them from getting there. The patterns he's seeing on the ground say more about where enterprise AI is headed than most strategy documents do. What comes across clearly is that the gap between using AI and putting it to work is wider than most organizations realize. Closing that gap starts with treating AI as part of how work gets done, not just how questions get answered.
Two Sides of the Same Tool
On the operations side, AI is already producing results. Process automation, workflow optimization, and supply chain analysis are saving thousands of working hours by redirecting capacity. The point isn't reduction; it's doing more with the same team. “We can do more things with the same amount of people, and that's already extra productivity that you gain immediately,” he shares. Repetitive tasks are being automated: data entry, manual reconciliation, and spreadsheet reports. Removing that burden opens room for people to take on more meaningful work and grow into new responsibilities.
The customer-facing side moves more slowly, but the potential is just as concrete. AI is starting to change how brands read and respond to consumers — anticipating preferences and running price sensitivity models that used to require external consultants and weeks of lead time. Marco describes a future where retail stops relying on guesswork. No more irrelevant newsletters. No more sales associates pushing products that miss the mark. Instead, brands get closer to what customers actually want and not what the company has decided to sell.
He points to Uber as one example already in motion. When he landed at the Hong Kong airport, the app picked up his location and offered a ride before he cleared the runway. He was able to link his flight number, so a car would be ready when he landed. It's a small thing, but it says something larger about where customer experience is going: AI moving from reactive to anticipatory — and in Uber's case, probably generating revenue it wouldn't have captured otherwise.
Where the Operating Model Shifts
Talking about AI in broad terms is easy. Applying it means confronting specific gaps in how organizations use the tool, read its output, and decide what to do next.
Go Beyond Using AI as the Search Bar
Most users treat AI as a faster way to find answers — ask a question, get a response, and move on. Marco calls that the baseline and not the end goal. The enterprise application starts when AI moves into how work actually gets done: how decisions are shaped, how workflows are structured, and how teams pressure-test their thinking.
His own approach reflects this. He uses AI as a reasoning partner — testing hypotheses, validating assumptions, and sharing what Copilot returns with colleagues directly in group chats. No hedging, no apology. His point is that a second opinion, even from a machine, beats operating on a single perspective. That shift in culture, where adoption turns into habit rather than novelty, is what separates organizations that experiment from those that execute.
Let AI Find What You Cannot See
No single team can manually surface patterns across pricing, supply chain, and consumer behavior at the speed or scale AI can. Marco's organization runs agents that analyze pricing elasticity, sales profitability, and store-level performance across global markets. The connections those tools surface across geographies, product lines, and timeframes are ones that would take multiple people hours to piece together, if they spotted them at all.
The value isn't in automating what people already do. It's in uncovering what they've been missing. As AI systems take on more autonomous roles in the enterprise, the ability to surface those hidden connections becomes what separates incremental efficiency from a fundamentally different way of operating.
Track Less, Understand More
Venchi tracks weather near stores, foot traffic, returning customer behavior, supplier lead times, production output, and more. He draws a personal parallel: his Garmin and Whoop log sleep hours, water intake, heart rate by day of the week, and every workout for the past six months. The data exists, but the meaning and impact often do not.
Organizations have gone from collecting nothing to collecting everything, and the challenge now isn't adding more dashboards to the stack — it's turning that overabundance into outputs that actually inform decisions. Marco warns that tracking more doesn't mean understanding more. Overabundance distracts as easily as it informs. His hope is that AI becomes the layer that focuses attention on what matters, rather than adding more noise to the pile.
Verify Before You Validate
AI hallucinates, and the people reviewing the numbers need the domain expertise to catch what the model misses. Marco is upfront about that tradeoff. If he asks Copilot to run a sales analysis on a specific country, he can tell when the numbers don't add up because he recalls those figures. A junior colleague who joined three months ago might not.
That's why he sees AI as more effective when paired with experienced hands — not because juniors can't use it, but because seniors bring the context to spot errors the model won't flag. It's also why he pushes for stronger fundamentals in education, the kind of baseline that prepares people to question what AI produces rather than accept it at face value.
Leaders Set the Pace
Marco's closing message is directed at leadership: this time, you're the ones driving the change. “We have to be the one carrying the burden of really giving good examples to the team and the organization on how we use ourselves AI every day,” he says. Leaders need to use AI in front of their teams, talk about it without qualifying, and make it clear that running a budget through Copilot before a meeting is diligence, not a shortcut.
The fear around AI and jobs isn't new. Marco shares how AutoCAD was supposed to end architecture, how Excel was going to replace accountants, and, going further back, how the calculator made newspaper headlines for threatening the labor market. Every time, people found their footing. Marco expects the same — as long as organizations move past the point where enthusiasm stalls against operational reality and treat AI as what it is: a tool that changes how work gets done, not just how questions get answered.
Soundtrack of Shift
Marco's Soundtrack of Shift pick is Mötley Crüe's “Same Old Situation”. His reasoning: every time a major shift arrives, the same split happens — one side leans in, the other defaults to fear. The song captures that recurring cycle and the choice it puts in front of everyone. Marco's take is optimistic: if more people chose to move past the discomfort, the outcome would be better for all of us.
Explore more soundtracks shaping how leaders approach change and transformation today.
Episode Resources
#shifthappens Research: The State of AI Report
#shifthappens Insights:
- Leading Through AI Adoption Delays Among Enterprises
- The AI Governance Blind Spot Leaders Are Missing
- PoCs Aren't the Problem: Why AI Never Reaches Production
#shifthappens Podcasts:
