The Customer Optimization Gap: Why AI Won’t Fix Retail’s CX Problem on its Own

Retail enterprises have never had more tools to reach customers, and customers have never rated them lower. In June 2025, Forrester’s Global CX Index reported that customer experience (CX) quality in North America had hit an all-time low across 469 brands. In the same window, martech still claimed roughly a fifth of marketing budgets at large enterprises, and artificial intelligence-powered personalization sat at the top of every retail road map.

Call this the customer optimization gap: the widening distance between how retailers optimize the channels they manage and how customers experience the brand across all of them. The gap isn’t new. What is new is what’s required to close it. The next layer of work isn't another martech investment, it’s an operating model and AI architecture rebuild.

The Gap Doesn’t Close From Inside Marketing

Most retail organizations were built around channels before omnichannel was a word. Email, e-commerce, paid media, loyalty, and stores each run with their own P&L, their own key performance indicators, and their own definition of a good quarter. Each function makes reasonable decisions inside its own lane. The customer experiences the sum of those decisions.

That sum doesn’t add up to a brand. It adds up to whatever pattern the channel calendars happen to produce. The dashboards in each function read green at the moment the customer is already gone. In March 2025, Kohl’s then-new CEO told investors that the third pillar of his turnaround would be "removing friction in Kohl’s omnichannel offerings." Q4 digital comparable sales had declined 13.4 percent. Kohl’s wasn’t running bad channel teams. It was running good channel teams accountable to the wrong unit of measurement.

This is an org design problem before it is a martech problem. The retailers closing the gap are reorganizing accountability around customer journeys: a single owner for the loyalty member curbside pickup experience, new household onboarding flow, and post-purchase service relationship. The journey owner answers for what the customer experience is. The channel teams answer to the journey owner.

AI Without Architecture is Just Faster Fragmentation

The AI capabilities maturing this year — agentic commerce, unified messaging APIs, autonomous shopping agents — are not closing the customer optimization gap in most retail stacks. They're widening it. Each new capability runs against its own data, optimizes against its own model, and answers to its own team. The customer experiences five AI-driven touchpoints that don’t know about each other.

What separates retailers extracting value from AI is not more AI. It is the architecture under it: unified customer data foundations, governance that follows the customer rather than the channel, and human judgment embedded at every layer where an AI decision touches the customer relationship. Forrester’s 2025 release of a Total Experience Score points the same direction. The measurement has to follow the customer, and so does the AI.

Where the Work Starts

This is not a multiyear transformation. The retailers building this practice are running 90-day cycles. Pick one customer journey. Build a shared scorecard across the channel teams that touch it. Instrument the data. Act on what it tells you. Measure what moves at the customer level: repeat purchase, share of wallet, advocacy. Pick the next journey.

The customer optimization gap is not a CX problem dressed in AI clothing. It is an operating model and architecture problem dressed in CX clothing. The retailers that figure out which problem they're actually solving will be the ones whose dashboards finally tell them what their customers have already decided.

BingYune Chen is CEO of Active Digital, where he helps enterprises architect and scale applied intelligence — built on decades of advising global brands like Marriott, Google, and Levi’s.