What’s holding retailers back from AI adoption?
By Jean-Matthieu Schertzer (pictured), Chief AI Officer, EagleAI
Artificial Intelligence (AI) applications have the potential to forever change many aspects of retail, from supply chains to physical and digital store operations to marketing. This is why retailers are making significant investments in this technology.
AI applications can help retailers meet consumer demand for personalised offers and deliver them efficiently at scale. There are, however, a lot of unanswered questions. One of those is, when will this technology be ready for implementation?
According to one study, AI adoption in retail is expected to surpass 80% in the next three years. Additionally, predictive and prescriptive analysis investments are expected to double over the same period of time.
While the rate of change is still impressive, what’s holding retailers back from adopting AI quicker?
Retail marketing strategies remain fundamentally unchanged, blending traditional and digital approaches. Digital advertising spending is growing rapidly, but traditional marketing isn’t obsolete either; it’s evolving.
Customers are often looking at a product in store but then buying it online, they also now expect more personalisation in their shopping experiences, both digitally and in-store. So, it’s important for retailers to adapt to meet customers’ digital expectations while recognising that in-store experiences are still crucial.
The trend is towards a seamless omnichannel experience, integrating both physical and digital elements. AI is playing a role in reimagining these strategies, as companies explore its potential in retail marketing.
The Precipice of AI-Driven Retail
The role of AI in business and society is still finding its place. Since the emergence of ChatGPT in 2022, the world’s eyes have been transfixed by generative AI without fully understanding how it will be applied or where it should be positioned.
One distinction to make before continuing is the difference between generative AI and predictive AI. Retailers using generative AI use the technology to create original content, like patterns, images, and text. Generative AI engines rely on existing data patterns to create something new.
In contrast, predictive AI uses patterns in historical data to project future outcomes. In other words, it can support strategy formulation and decision-making. Retailers already make data-driven decisions, but predictive AI’s emergence can take it to the next level.
With predictive AI poised to be a major gamechanger for retailers, we think there are three key points retailers should understand about AI adoption:
1. The need for data quantity and quality:
Predictive AI is an exciting development in retail, but it remains in its early stages. Just as future customer behaviour cannot be predicted from a single data point, usable retail AI outputs (like measuring a shopper’s brand affinity) need sufficient data to be effective. Similarly, AI models trained on poor-quality data will generate subpar outputs. Therefore, pre-processing data, from that perspective, is of paramount importance.
2. Optimal integration of AI outputs:
When implementing an AI model’s outputs, there is a trade-off between full automation (AI outputs trigger events such as emails, promotion offers sent to clients, generated images used for real-time ads, etc.) and systematic manual review.
Sometimes, the choice is obvious. However, finding the right implementation balance often requires adapting existing tools (or using purpose-built monitoring dashboards), putting common-sense guardrails in place, and enforcing manual review when AI predictions are uncertain.
3. An AI-driven virtuous cycle:
A significant driver of the relevance of AI outputs (prediction/content) is the ability to see whether predictions are correct — or not. This allows for the next round of AI system optimisation, driving the performance upwards. This continuous improvement cycle can end up being a solid competitive advantage.
The first step of the journey to AI integration might seem high, but retailers should understand that optimisations multiply quickly, and the initial performance improvements are only the beginning.
None of these limitations reduce the value of predictive and generative AI, but brands must be aware of them when integrating AI into various retail marketing niches, including when planning sales, promotions, and loyalty program alterations.
This article is an excerpt from the recent whitepaper, Eagle Eye’s AI Anthology, the definitive resource on leveraging AI for personalisation and retail marketing. To check out the full document, which is full of data and insights from our AI experts, follow this link.
