To help retailers and brands plan for 2019, Researcher Claudia Tajima and I are interviewing experts within Forrester for our series, “Applying 2019 Predictions To Retail.” Last week, we spoke to Sam Stern about how employee experience will impact retailers in 2019. This week, Claudia interviewed Michele Goetz, principal analyst on Forrester’s enterprise architecture team and an expert in AI, on her 2019 AI predictions report. Here’s what Michele thinks retailers and brands can expect and should focus on regarding AI in 2019.
Claudia: In general, what should be top of mind this year for retailers in terms of artificial intelligence?
Michele: Retailers don’t fully realize that they have to do more curation to determine what should educate the AI systems [in order] to execute retail professionals’ jobs. Take personalization, for example. AI in personalization is much more intent-driven, meaning that marketers need a more robust understanding of customer needs, wants, influences, and relationships. You cannot just log customer behavior that is pulled from search, traffic on-site, or captured from what comes in from the store. Why? AI doesn’t care about one-time customer behavior: AI cares about everything. We are looking at what customers prefer and why, and the data must represent that.
Claudia: What should retailers do to bypass data “doldrums” that can drown their AI efforts?
Michele: Retailers should think more intentionally about structured data. They should compare how they currently gather data with simulations in the store environment to see if those two things diverge. Digital twins can help simulate your business as it happens in virtual time. You can apply different types of behaviors in simulation to drive better outcomes. Data strategy needs to support creating the experience first, and then the data can be used for insight to influence the experience.
Claudia: As the race for AI talent picks up, do you have any recommendations for retailers using AI in their recruiting processes?
Michele: AI is really good at identifying the nuances of talent. [Therefore], retailers should consider using AI to improve traditional recruiting tactics such as using a job board or LinkedIn and searching for certain keywords. In a traditional human–based approach, you are applying your own biases to the recruiting process. AI helps you strip away your own bias. How? For example, AI can search for prospects who have created new experiences for customers or those who have created new digital assets. This AI-driven approach enables you to look for the underlying qualities and experiences of prospective recruits because AI requires you to pull apart existing processes and recreate them to be more effective and efficient.
Claudia: How will the increased number of non-human “digital workers” — from robotic process automation (RPA) and AI — impact retailers?
Michele: Retailers will need to reconsider what they do with inventory and fulfillment today. Companies that have ship–to–store and store pick-up like Kohl’s, Lowe’s, and Walmart have applied machine-learning practices to expose inventory visibility by tapping into warehouse and store inventory data.
Claudia: Across industries, the demand for “explainable” AI is growing. How should retailers approach this trend and building out their AI strategy?
Michele: Explainable AI provides a window into why a machine makes a decision or takes an action. Retailers need to really think about how to put in place appropriate mechanisms to determine if they are training AI appropriately. But there are also performance metrics to track and rules to apply [in order] to keep AI from going rogue. Think of AI as another worker. We see businesses placing more demand on AI vendors to fulfill requirements and to be partners in defining requirements for any given retail risk area (e.g., privacy).
Claudia: Will retailers need to consider bringing human expertise back into the decision-making loop?
Michele: Business leaders are realizing that some aspects of machine learning can’t just run on their own. Machine learning still needs strong governance and management. With AI, there is always going to be a balance between oversight and the power of a machine where you can’t give the machine free rein. In customer service, think of AI as a virtual agent that you can incorporate side by side with the real agents you have. When comparing how personalization and next-best offers occur, tap humans to consider if the machine is starting to over–bias on giving free gifts because that action improves customer experience (CX) scores. Situations like these require a human to course-correct or educate and extend machine-learning capabilities.
Claudia: For retailers looking into buying AI solutions, what should they be aware — and wary — of?
Michele: Don’t buy AI technology until you know what job it will do and how it will execute that job. You must define your desired CX and operating model up front and then determine what needs to be acquired and incorporated into what you already have. You must be customer-driven. For example, partner with more experienced AI consultancies that push you into the right mindset. These service firms should ensure that you have a good CX strategy and understand how to be customer-obsessed (e.g., being project-, not platform-, oriented). Finally, remember that without the right data and analytics governance, you won’t have the right AI governance.
Stay tuned for next week’s interview, where we’ll speak to Emily Collins about loyalty. For more information on where the retail industry is headed and how to prepare your business, join us for our complimentary webinars: Future Of Retail Three-Part Series.