3 Mistakes Retail C-Suites and Boardrooms Are Making When it Comes to AI

Earlier this year, my company surveyed CIOs and CTOs at companies making $500M+ in annual revenue about how their brands were approaching AI. Across these executives, 84% agreed that only those AI projects that resulted in ROI would survive the year, and 82% agreed that their boards were increasingly questioning how much they were spending on AI.

In my view, this double-edged pressure is largely responsible for companies – especially retailers – turning to an old habit when it comes to incorporating the latest technology wave: what I like to call “random acts of [fill in latest technology here].” Over my career, I’ve witnessed random acts of mobile, random acts of IoT, and now we’ve arrived at random acts of AI.

More than any other prior technology wave, labeling your adoption of AI as a “tech issue” strictly under the purview of the CIO/CTO is missing the forest for the trees. For retailer leaders who are serious about gaining a competitive edge in the current AI wave, there are three major mistakes they should avoid.

Mistake 1: Letting the Corporate Antibodies Win

Because AI has the potential to be such a transformational force within a company, one of the greatest obstacles to truly getting the most out of it is a fixed perception of what a company is or can become. For retailers in particular, this tends to sound like a product-first mentality –  i.e., “we sell shirts, we’re not a technology company.”

This inflexibility is a corporate antibody – ensuring true innovation is kept at bay for the sake of maintaining the status quo. As a result, AI budgets are delegated to CTOs who are given cost-cutting marching orders. With no strategic imperative or long-term vision to guide them, unsurprisingly, they turn to familiar territory: hurried investments in bolt-on point solutions like recommendation engines, chatbots, or visual search plugins.

At best this approach leads to a small, though immediate, lift. At worst, however, these fragmented AI tools lack the backend organization to work across the organization as a whole, and therefore more often result in “rogue” experiences that frustrate customers. Ultimately, I’ve seen this end with abandoned pilots and wasted budgets. A lose-lose.

Mistake 2: Inflexible Investment Thesis

A related issue to that of the inflexible company vision is an inflexible investment thesis. When investing in AI, C-suite executives and board members need to have honest conversations as to whether they’re actually interested in AI transformation (and all that it entails) or if they would be better off saving time and money by avoiding it altogether. It helps to anchor the conversation in a concrete and widely known example like Amazon.

It’s no secret that Amazon is the powerhouse it is today because it had over a decade of making no meaningful returns. During that period, the technology and long-term vision were seen as valuable enough to warrant the investments and lack of immediate ROI. Reshaping a company around AI requires the same level of going all-in.

At a recent industry event, online retailer Stitch Fix shared that the company was running around 40 concurrent AI pilots in their dogged pursuit of personalized shopping experiences. That is what “stubborn vision” looks like when AI transformation is taken seriously. And if your eyes just bulged out of your skull thinking you don’t have the budget for that sort of thing, you’re not alone.

I recommend approaching AI like a three-tiered diversified investment portfolio: quick wins (fast, low-risk momentum builders like chatbots), staged bets (medium-term “invest and learn” initiatives like personalized customer messaging), and game-changing bets (high-risk, high-reward plays like fully-integrated AI shopping assistants that can redefine a company). Pursuing all three simultaneously builds near-term results while creating the resilience and momentum needed for larger, transformational change.

Mistake 3: AI Candy for Breakfast

Approaching AI as little more than a cost-cutting tool or marketing gimmick is a lot like eating candy for breakfast – the novelty doesn’t last much longer than the Red 40 on your tongue, and you’re going to have to fuel up on something more substantive before you know it. The clearest marker that a tool is AI candy is it prevents your business from transacting, fulfilling, or learning legibly underneath it.

So don’t start your retail AI journey with a pre-packaged chatbot – start by investing in the infrastructure that paves the way for a proprietary chatbot that learns from your own customer base, giving you a true competitive edge. This starts with getting your data agent-ready.

You’re probably already thinking about how to enrich your product data for LLM discovery, considering the ways these platforms are increasingly becoming the default starting block for search. Beyond that, you’re also going to want to invest in an intelligence layer that understands customer context (behavior, loyalty, order history, stated intent), session context (what the shopper compared, asked, hesitated over, and left in the cart), and business context (inventory, margin, promotions, and merchandising rules). By structuring and organizing data in this way, a flywheel is created that continuously learns and improves on itself.

More importantly, however, the beginning stages of your AI journey should also be where you’re building the governance and risk-tolerance architecture that will enable agents to truly act on their own. Though there is much talk these days about keeping a “human in the loop,” this is neither scalable nor cost-efficient.

For enterprise retailers, these steps are certainly easier said than done. But to take AI as a technology seriously demands putting substantial work in at the front end that could lead to vertical-disrupting innovation in the near future. Of course, there is always the option to continue with business as usual, but as AI moves toward becoming the status quo, the window to play catch up only gets smaller. I don’t recommend getting left behind.