Forrester predicts one in five US and EMEA retailers will launch customer-facing GenAI applications in 2025. Enhanced product search, personalized recommendations, and improved category navigation are top use cases. So why did automated interactions cause the US’s customer experience score to decline by 5% in 2023—the lowest since 2015—and what can retailers learn from this before making their GenAI investments?
The 2023 KPMG report highlights failure to meet customers’ expectations as the cause of decline, with overuse of technology that lacked strategic benefit to shoppers. Of 50 CIOs and CTOs in Fortune 500 enterprises questioned on their GenAI projects, most found their pilot technology addressed the wrong business need.
As we enter 2025, retailers must prioritize customer-centric GenAI strategies. Rather than adopting the latest technology as a nice-to-have, look at the business needs. Retailers should review their customer journeys, identify the room for improvement, and build or adopt solutions that fit their use case, not the other way around. Here are four lessons for retailers to consider on their journey to elevate the user experience (UX) with GenAI.
Ensure Business-Data-AI Synergy
RAND researchers found in 2024 that 80% of AI projects fail due to five key areas: misaligned goals, data deficiency, tech-first approach, infrastructure gaps, and overambitious AI.
Retailers require a solid data foundation and expertise to build the required algorithms and succeed with their GenAI investments. They should ask themselves, “How can we ensure sufficient data availability to meet the solution’s requirements? And how much of this data is proprietary?” Successful GenAI projects hinge on high-quality, relevant information. The more unique data formats the organization has, the more customizable the solution needs to be.
A third question to ask is, “What specific talent pool and operational structure changes are needed to leverage GenAI effectively?” Understanding the level of upskilling, along with the motivation, costs, and time, will help retailers decide the return on investment (ROI) for building, customizing, or managing solutions in-house.
Today, non-technical experts can work with no-code tools or hire a long-term AI partner to leverage the benefits. When selecting third-party GenAI solutions, e-commerce executives should prioritize factors beyond pricing and ROI, such as scalability, performance, data security, vendor expertise, and tech stack compatibility. A clear business case and expected outcomes are crucial before committing to any new integration.
Take an Incremental Approach
In 2024, BCG Group evaluated the adoption rate of the top e-commerce GenAI use cases; namely, content creation such as blogs, product descriptions, and product image supplementation. More advanced use cases include personalized product recommendations, dynamic pricing, and competitor analytics. Familiarize team members with systematic services before trying your hand at more complex tasks to adjust to new processes seamlessly.
Retailers should encourage their e-commerce teams to leverage out-of-the-box GenAI tools to acquaint themselves with the tool’s capabilities. Simple use cases and no-to-low-code solutions such as product descriptions and image creation are excellent starting points as they show team members the possible time savings, as well as help them adjust their operations to include frequent validation checks. Introduce weekly or biweekly reviews in the early stages to measure the tool’s progress and adjust approaches along the way. Team feedback and participation will be key to success.
As team members become more familiar, retailers can introduce new use cases. Engineers can streamline development with AI code completion assistance. Marketers can introduce AI-driven personalized upselling and cross-selling recommendations, and loyalty managers can build adaptive loyalty campaigns based on customer engagement level.
Create a Security-First Culture
Disconnected systems are weak links that can lead to security vulnerabilities, and GenAI has the potential to lower the entry barrier for low-skilled threats. Cybercriminals can use GenAI to build scripts that could be functionally malicious if used correctly, automating attacks and targeting specific vulnerabilities. Retailers should aim for a solid data foundation, streamlined workflows, and a well-connected network of applications to keep their systems safe and easy to monitor.
Cybercriminals may also use GenAI to manipulate consumers through highly convincing fake content (i.e. social engineering and phishing). Therefore, identity verification will be even more critical in 2025. Multifactor authentication, such as sending time-sensitive codes to user devices via SMS, email, or a dedicated authentication app, will help secure customer loyalty programs and shopping platforms—especially where financial information is saved.
In addition, retailers must ensure developers regularly update software, software libraries, and systems to address vulnerabilities and minimize attack surfaces. This safety-conscious, verify-first mindset should be filtered through the entire organization. By conducting regular security awareness training and simulations and encouraging employees to report suspicious activities promptly, retailers can build a security-focused culture.
AI-powered monitoring and alerting systems, such as advanced endpoint detection and response (EDR) solutions, can also help retailers detect and mitigate threats in real time. Even so, it’s important that all employees are in the habit of verifying that systems, especially cybersecurity software, are working as they should.
Be Empathetic by Design
The biggest cause of AI distrust is its use in customer support channels. Some 53% of customers would consider switching to a competitor if they found out a company was going to use AI for customer service.
Customers fear that GenAI will build a wider gap between them and support agents. They want peace of mind that their issues will be understood and resolved in the best way possible, ideally with managers who have the authority to offer complimentary gifts for their troubles. However, retailers can build these steps into their automated services. But it’s still important to start with simple tasks first. Making FAQs and online information more accessible via conversational chatbots are helpful use cases.
In the beginning, more hands on deck to respond to customer feedback, confusion, or queries will be a proactive and welcomed buffer as retailers adapt to GenAI’s capabilities. Real-time feedback from support teams will help retailers imagine all scenarios where tasks are too complex for GenAI tools. In these scenarios, chatbots must direct customers to an agent with a holding message, such as: “Offer not helpful? Contact an agent” button. Analyze this feedback daily until all possible common queries are answered simply and automatically.
It’s essential that all tasks GenAI tools undertake seamlessly transform into an agent chat that picks up where the chatbot left off if needed. It’s also critical that customer service agents remain a key part of the user journey, saving them for high-value tasks such as watching the data and identifying underlying causes of recurring customer issues. This way, retailers have a basis to propose solutions and prevent future problems with automated response channels.
Whether retailers choose to adopt GenAI or not, competitors, customers, and malicious actors will. Preparing team members with simple use cases will help them adjust to new ways of working and better understand the new potential threatscape. Retailers can leverage out-of-the-box tools and trial GenAI projects in a phased approach, building on their teams’ knowledge and expertise with more advanced algorithms each time a project is fulfilled successfully. By automating the transactional tasks and keeping an expert team of human agents, customers can enjoy quicker access to desired products and feel reassured there is an agent a call away if they need them.