Retailers are under immense pressure. As tariffs inflate global product costs and consumer confidence dips, margins are shrinking and the fight for customer loyalty is fiercer than ever.
Amid this uncertainty, new AI tools and strong customer data offer a way to turn pressure into a competitive advantage. According to McKinsey, generative artificial intelligence (genAI) could unlock between $240 billion and $390 billion in value for retailers — the equivalent of a 1.2–1.9% increase in industry-wide margins.
In marketing, brands can use AI to convert omnichannel data into sharper insights and scale content creation to match how individual audience segments actually browse, shop, and buy. When deployed properly, these tools help retailers come closer to the one-to-one personalization that drives engagement and increases customer lifetime value (CLV).
But AI integration isn't a switch you flip. Retailers face significant hurdles. Here are four key challenges and how to overcome them:
1. Building a strong data foundation
As retailers embrace omnichannel strategies, they're collecting more customer data from more sources than ever before.
This means brands are having to track activity across multiple channels — websites, apps, loyalty programs, physical stores — while contending with existing pain points like duplicate customer accounts and outdated contact details.
Duplicate accounts, outdated contact info, and disconnected systems lead to fragmented customer views. Feeding flawed data into AI leads to unreliable outputs: "garbage in, garbage out."
To apply AI successfully, retailers need a clear, reliable view of their customers. This begins with Identity Resolution – the process of resolving inconsistencies and stitching together existing records into a single, accurate customer profile. Identity resolution gives teams a solid foundation to understand customer behavior and measure value with confidence.
Once in place, these profiles can be fed into broader business tools like customer data platforms, supporting everything from audience discovery to channel strategy. The key is identity: your AI initiatives are more likely to succeed if you truly know your customer.
2. Eliminating data and communications silos
Retailers aren’t short on data; they’re short on alignment. Different teams manage different systems using different formats, often in isolation. AI can’t generate full-picture insights if it only sees part of the data.
Even with the right infrastructure, siloed teams can stall progress. Each department may have different goals for AI, and without coordination, efforts may clash or stall. Alignment on data sharing, priority use cases, and success metrics is essential.
Fortunately, AI can also foster collaboration. For instance, marketing no longer has to wait on IT for data pulls. With the right tools, teams can directly access insights and generate tailored content in real time—streamlining workflows and uniting efforts.
3. Right-sizing spending and goals
Hazy objectives and ill-defined scopes are known killers of digital transformation. The more specific your AI goals are, the greater your chances of success.
Your organization needs to decide early on what it wants AI to achieve. Are you trying to increase CLV? Acquire new customers? Improve conversion in priority segments? The goal doesn’t need to be revolutionary, but it does need to be clear.
You also need to right-size spending. Delivering personalized marketing collateral to each of your 300,000 customers is more time-consuming and expensive than some AI vendors want you to believe. That level of personalization requires time, resources, and a data environment that many teams haven't built yet. AI can definitely help, but only if it's targeting a clearly-defined challenge.
A smarter, more scalable approach is to use AI to create sharper audience segments and quickly tailor messaging to each one.
Say a retailer wants to promote a seasonal sale. Instead of sending the same generic message to every customer, marketers can use generative AI to tailor communications based on different segments' browsing habits and price sensitivity. One version might highlight high-ticket items for frequent shoppers, while another promotes bundle deals for budget-conscious buyers.
With the right data and prompts, a retailer can spin a single promotion into tailored messages for different customer segments without having to start from scratch each time.
4. Reducing risk and promoting responsible use
AI is not without risks. If the underlying data is flawed or incomplete — or if the organization lacks the necessary safeguards — AI models can produce inaccurate outputs that distort reality and misguide strategy.
Responsible use starts with education. Teams need to understand how and where AI adds value, where its limitations lie, and what safeguards should be in place to ensure its outputs are reliable. Human oversight should be built into the process so that outputs are reviewed in context before they're acted on.
Good data hygiene is just as critical. If customer records are out of date or poorly stitched together, any AI model trained on them will carry these weaknesses forward. Gaps in data accuracy create blind spots in AI's understanding; over time, those blind spots can lead to poor assumptions shaping key decisions.
AI delivers – with the right foundation
Retailers are moving fast to capitalize on AI's promise. But true transformation starts with clean, unified customer data. For a deeper look at building AI-ready infrastructure, download Amperity’s latest retail guide, “New Opportunities, High Stakes: Maximizing the Value of AI and Customer Data in Retail” to discover how to turn your customer data into your greatest AI asset.