AI and Graph in the Retail Industry
Macro-economic forces over the last few years expedited the journey of technology adoption for the retail industry, affecting most aspects of a retail organization. While sub-categories within retail – apparel, grocery, convenience stores, consumer packaged goods (CPG) – accelerate in different areas at different paces, specific challenges and opportunities span the entire industry.
Our Graph + AI Summit session, AI and Graph in the Retail Industry, explored some of the recent and new digital imperatives, the ongoing challenges with wrangling and managing data, where the opportunities are, and how graph plays a significant role.
The panel, held during the May 2022 Graph + AI Summit, featured Dinand Tinholt, vice president insights & data, Retail/CPG at Capgemini; Wei Manfredi, global architecture leader – retail/CPG at Google Cloud; Manish Sood, chief technology officer, founder, and chairman at Reltio; and moderator Deena Lawrence, retail and CPG industry strategy lead at TigerGraph.
The Issues Facing the Retail Industry
The COVID-19 pandemic is just one of the forces that ushered in and augmented a host of issues for retail organizations. How do you understand, and quickly react to, customer behavior when it changed overnight and continues to change? How do you offer personalized experiences? How can you blend the channels that consumers use when there are more than ever before?
It’s not just on the consumer side either. Organizations have more data while using more systems and applications than ever before. More business units need access to the same datasets than ever before. Supply chains have experienced more disruptions than ever before. And one of the most significant issues facing retail – speed, or the need for everything in real-time, including transaction payments, insights, recommendations, and fraud detection – has never been more real.
Suffice it to say staying efficient, nimble, and profitable has never been more complex.
Setting a data strategy and managing data effectively at an enterprise level and scale set the stage for unprecedented ROI; yet, with great rewards come trials along the way.
Lack of Business Stakeholder Alignment and Executing in Isolation
Everyone thinks, “If we have better data, we’ll have better outcomes,” said Manish. “But quantifying the value and tying it to the value delivered to business stakeholders is the part where either it takes too long to get to that type of baseline, or the lack of alignment forces people to go down the wrong path … so making sure [to know] what is the business value and what is the acceleration to that business value will lead to better outcomes.”
The data from all your applications isn’t “useful at all unless you use it for some business goal,” agreed Dinand. “Defining your strategy, how you want to become a data-powered enterprise, what business value … when we want to create that goal, that ambition is usually easy to formulate and roadmap, but executing it in collaboration between those pillars [data, business value, and IT implementation], that’s going to be crucial.”
One aspect that makes data strategy execution difficult is that many data systems still reside and function in silos. “To leverage, you copy data over and over again to create the models and insights,” stated Wei. “Multiple copies later, you’ve lost traceability of where the data came from and who leveraged your data.
“If you really want to get into data management, make sure you have visibility where the data came from and who’s using live data,” continued Wei.
Opportunities Abound Globally for Retail
While the retail landscape is certainly turbulent after navigating a global pandemic and due to other market factors, there are also some incredible opportunities thanks to advances in technology.
A Juniper study projects global retail spending on AI will reach $7.3 billion annually by this year, up from $2 billion spend in 2018. Investments in AI-powered predictive and prescriptive analytics will more than double in that same period.
Artificial Intelligence (AI)
AI is already a digital must. Most retail organizations use AI to a varying degree: conversational AI, chatbots, services enhancements, online product recommendations, intelligent digital assistants, and more. Developing features include customizable products, cashier-less shopping, and emotion chatbots.
Many of the challenges mentioned earlier revolve around relationships within your data. Traditional technology, such as relational databases, falls short when uncovering insights in your data, especially when relationships matter. Graph databases provide a level of business value, making this technology a new digital must-have. Each of our panelists has compelling reasons for enterprise retail organizations to implement graph analytics.
Wei Manfredi – “…Where graph shines is really being able to derive non-obvious relationships from the existing relationship, [being] able to find a correlation. And that’s where the recommendation side is all about leveraging that graph type of relationship.”
Manish Sood – “Data has to be organized as an organizational institutional asset, which leads us to the interconnected nature of the data. Graph is one of the ways in which we are going to make that come to fruition.”
Dinand Tinholt – “What I love about graph technology is that it allows you to create an understanding, in a very complex world, which will allow organizations to use that to increase their adaptiveness and resilience.”
Collaborative Data Ecosystems
Collaborative data ecosystems (CDE) are also making strides in the retail industry, though more slowly than AI and graph technology. CDE is about “how can we get more value from our data by combining it with that of other organizations in our value chain,” explained Dinand. “You can have different levels, different tiers of sharing. You can open up the insights while not disclosing all of the data.”
Watch the Recording for More Insights
Our panel covered many fascinating topics. Find out what various retail sub-categories focus on, what ‘farm-to-fork’ really entails, innovative ways to use graph technology, the importance of data connectedness, the revenue impact of collaborative data ecosystems, and more by watching this fascinating Graph + AI session.