Thanks to AI, the retail industry has never moved faster.
With each new week, a new model release, AI-native startup, or shift in customer behavior changes how retailers should operate to maximize revenues. According to Amperity’s 2025 State of AI in Retail report, nearly half of retailers (45%) already use AI daily or several times per week, and 97% plan to maintain or increase AI investments in the year ahead. Yet only 11% feel strongly prepared to deploy AI at scale, and less than half (43%) have brought AI into customer-facing experiences where it can have the most business impact.
Retailers understand the potential. Gartner predicts that more than 80% of enterprises will be using generative AI in production in 2026, up from under 5% just a few years ago. McKinsey estimates that generative AI may unlock $400-660 billion in annual value across retail and CPG, driven largely by improvements in customer engagement, forecasting accuracy, and operational decision-making.
Yet retailers are still figuring out what true AI readiness looks like. AI delivers compounding results - for better and for worse. As a result, your strengths and weaknesses (from data quality to customer profiles to workflows) will become more visible with each new application.
One thing is clear: at this rate of change, the things that feel innovative today will be taken for granted two years from now. It’s important to examine where retail is headed - what “good” looks like today, what should be possible in 2026, and what a realistic future looks like by the end of 2027.
Last year: Robust profiles, natural language interfaces
2025 was a year of experimentation, and of rediscovering a familiar truth: every AI use case is only as good as the customer data underneath it.
In our conversations with retailers, identity remains the single biggest blocker to AI success: conflicting counts of “how many customers we actually have,” fragmented data across channels, and uncertainty about which signals are trustworthy enough for modeling. The data from our State of AI in Retail report backs this up: 58% of retailers say their customer data is fragmented or incomplete, and only 23% use AI in production to resolve customer identities today.
Retailers understand that before AI can accelerate value, it must first fix the fundamental problem: building complete, trustworthy profiles that update in real time.
Once your identity foundation is strong, generative AI becomes a workflow accelerant rather than a risk. Natural language interfaces let marketers describe an audience, strategy, or journey - then watch the system assemble the segments, logic, and orchestration automatically. Tasks that once required hours of clicking through tools now take minutes, returning time to actual strategic thinking.
In 2025, leading retailers used AI to unify profiles and maintain them continuously, querying customer histories using natural language, generating audiences and journeys in minutes instead of hours, and experimenting faster, with tighter loops between insight and execution. These capabilities are table stakes for 2026.
What this unlocks is agility - the ability to respond instantly when customer behavior shifts or when a competitor launches a new product. Retail is too dynamic for quarterly planning cycles. AI gives brands the speed to launch an offensive or defensive strategy in days, not months.
This year: Proactive, AI-driven actions
If 2025 was about operational lift, this year will be about intelligent orchestration.
Historically, marketers have had to stitch together churn models, product recommendation engines, LTV scores, and predictive segments manually. Now, we should see AI begin to unify those pieces of intelligence into actionable recommendations with human oversight. As a result, the marketer becomes more of a business architect: deciding goals, setting guardrails, and letting the system proactively monitor the customer base.
After establishing the baseline use of AI for unified profiles, organizations can move on to improving workflows. By the end of 2026, organizations will be able to continue their AI use to enable:
Proactive detection of customer behavior shifts (Think, “Retention dropped 5% in the last 72 hours.”)
Root-cause analysis (“A store closure within 20 miles of top-value customers.”)
Strategy recommendations (“Include 3 alternate store locations based on proximity and loyalty tier of customers within 20 miles.”)
Instead of asking, “What happened, and what should we do?” the system will communicate, “Here’s what changed, why it changed, and the three best strategies to fix it.” Retailers remain in control - approving or rejecting recommendations - but the heavy lift of analysis and ideation becomes continuous and automated.
This experience will only be perfected by late 2026 and into 2027, as retailers reap the rewards of a unified data foundation and as natural language interfaces mature into goal-oriented AI systems rather than isolated tools.
In two years: Organization-wide customer intelligence
By late 2027, retailers will experience the shift from AI as a marketing accelerator to AI as a company-wide nervous system. The defining change at this stage is that customer intelligence stops belonging to a single department and instead becomes an input that shapes decisions across merchandising, operations, supply chain, service, and digital commerce.
Rather than working from disconnected versions of the customer, the entire enterprise begins operating from the same real-time understanding of who customers are, how they behave, and what they need next.
In this future, customer intelligence will inform operational decisions that traditionally lived outside the marketing ecosystem. Inventory planners won’t just forecast based on sales history - they’ll incorporate churn risk, emerging preference signals, or shifts in price sensitivity from high-value segments. Store operations will adjust staffing models and service strategies not only by foot traffic but by predicted customer intent.
Customer journeys don’t begin or end in a brand-controlled channel. Customer interactions will flow through external AI agents and conversational interfaces, collapsing awareness, consideration, and purchase into a single continuous dialogue. With agentic commerce, consumers will soon research products, compare alternatives, receive personalized recommendations (and ads), and complete transactions without ever visiting a website.
This shift is coming - retailers need to see into the future and start taking steps to prepare. Success will depend on a brand’s ability to send the right intelligence to these agents and ensure that wherever a customer buys, the retailer shows up with the most accurate product data, offers, and service context.
Identity will still be the foundation. A retailer can only power cross-organizational intelligence if its customer profiles are accurate, unified, and continuously refreshed. In two years, AI will be capable of adjusting strategies across departments and influencing purchase flows happening outside of brand-owned environments - but only for retailers that have built the data foundation to support that level of orchestration. For everyone else, the gaps in identity and data quality will become even more limiting.
AI innovation: Built on identity
In 2025, 2026, 2027, and beyond, one theme never changes: AI reflects the state of your organization. If your customer data is fragmented, AI will amplify that fragmentation. If identity resolution is weak, personalization, predictions, and agentic experiences will miss the mark.
But if your identity foundation is strong, AI will compound value across workflows, channels, and business units. The retailers who will win the next two years are already doing three things today:
Strengthening identity resolution so every customer profile is complete and unified
Using natural-language interfaces to accelerate experimentation and remove bottlenecks
Preparing for a world of agentic commerce, where customer intelligence must serve both owned and external channels
AI will reshape the entire customer journey, and only the retailers who build the right foundation and data strategy first will reap the rewards AI has to offer.