Most retailers are already “doing AI.” The work that turns that into value is quieter, harder, and mostly not about the technology.
By Henry Sands, Associate Partner Columbus Consulting
Bring up AI in almost any retail leadership meeting today and someone will say, “We’re already doing AI.” And they’re right. Pilots are live, copilots are drafting, teams are experimenting. That’s real progress, and it’s worth being proud of.
But “doing AI” and getting value from AI are not the same thing. 2026 wasn’t the year to go big; it was the year to get aligned. That still holds. What’s changed is the next question. It’s no longer “should we start?” It’s “how do we turn all this motion into something the business can actually absorb?” And the honest answer starts with a part of the story that’s easy to miss.
Why the hardest part hides in plain sight.
Most boards have, very sensibly, handed AI to the CIO. And the CIO, just as sensibly, leads with what they know best: technology, platforms, and data. That work is essential, and nothing else is possible without it.
But tech and data are only one side of the coin. The other side is the organization: the decisions, the trust, the people, and an entirely new way of working alongside intelligent systems. That side rarely has an owner in the room. It isn’t anyone’s failure; it’s a blind spot built into how the work got assigned. And it’s precisely where value is won or lost.
Value follows decisions, not technology.
It’s natural to start with technology. It’s tangible, and it’s what the vendors are selling. But AI doesn’t follow the org chart; it follows the flow of decisions. The better starting point is to ask which decisions actually move the business: in buying, planning, pricing, allocation, replenishment. Once those are clear, you can decide which ones AI should assist, which it can make, and which it can eventually run. Decisions first, technology second. That’s what turns AI from a science project into a business case a CFO will fund.
People move from doing the work to guiding it.
Here’s a shift that’s easy to underestimate. Your best people gradually stop being the doers and become the supervisors. The planner who once pushed the units begins reviewing an agent’s recommendations, working the exceptions, and sharpening the rules it follows. That’s a genuinely different job, and people need help growing into it. Plan for that shift, be clear about where humans stay involved and where the system can run on its own, and teams lean in rather than route around it.
Your experts’ judgment is the ingredient no vendor can sell you.
Your best merchants and planners carry judgment that has never been written down: why you never markdown or promo that item, which vendor slips in Q4, which store always runs hot. That hard-won instinct is exactly what a safe AI system needs, and it lives in people’s heads, not in the data. The real opportunity of the coming year is drawing that knowledge out and building it into the guardrails and escalation paths that keep AI trustworthy. The retailers who reward people for sharing what they know will build systems people are comfortable trusting. The rest will keep a human in every loop, and wonder where the value went.
Trust is earned before autonomy is granted.
Trust in AI isn’t a launch-day event; it grows. It grows by starting small, proving value, and keeping an honest eye on how the system performs, where it’s consistently right, and where it still hands the decision back. Autonomy isn’t a switch. It’s earned, one reliable handoff at a time. Watch the results hold steady and confidence compounds; skip the evidence and every conversation stalls in opinion.
And it all rests on an aligned C-suite.
None of this survives a divided leadership team. When the CIO, the CFO, and the merchants share one view of the ambition, the value at stake, and how the organization will work, AI moves. When they don’t, even good pilots quietly die. Alignment beats enthusiasm; it always has.
Don’t go it alone.
The good news is that you don’t have to work this out alone, and you don’t need to be an AI expert to lead it well. This is exactly what our AI offerings are built for. Grounded in a proven, externally validated approach to agentic AI, adapted from Harvard’s framework and rewritten for retail, experts like Columbus Consulting can help leadership teams climb the ladder deliberately, from GenAI to assists to agents to autonomous workflows, with the organization, governance, and trust built in from day one. Most engagements begin in weeks, not quarters: an AI Readiness assessment to show exactly where you stand across data, architecture, organization, and governance, followed by a clear strategy and roadmap that sequences the decisions worth tackling first.
Start with clarity. Understand where you are. Build the roadmap together. Because the retailers who win the next 12 months won’t be the ones doing the most AI. They’ll be the ones who turn it into value the business can trust.
For more on AI in retail: https://www.columbusconsulting.com/generative-ai-at-a-glance/