Retailers, beware of the AI hype
There is no topic that is currently more widely discussed than artificial intelligence (AI). Since ChatGPT launched in November 2022, the hype has not abated. Businesses across all industries have quickly realized the potential in terms of cost savings, increased productivity, improved customer experience, and more. Retail is no exception.
When it comes to commerce, AI can be used to completely transform the customer experience from start to finish. From advertising, to how customers search and buy on the website or in stores, to order processing and delivery.
Many retailers are already using the technology in creative ways. For example, eBay has introduced a ShopBot that acts as a personal shopping assistant. It helps customers sift through the listings on the site to discover the most attractive deals. Customers can interact with the ShopBot via text, voice or even by sharing a photo to indicate what they are looking for.
SVP and MD EMEA at Fluent Commerce.
The pressure to innovate
However, the reality for most retailers is that they are not yet able to capitalize on AI technology. This is due to a lack of data, both in quantity and quality. Boards and marketplaces are putting pressure on technology providers to launch new AI products. As a result, many of the ‘new AI tools’ we see on the market today aren’t new at all. They are existing technology, leveraging machine learning, that has now simply been rebranded as ‘AI Tools’.
Even retail leaders like Amazon have succumbed to this pressure. After years of touting their AI-powered Just Walk Out cashierless technology, it was recently discovered that these were in fact just cameras reviewed by real people in India. With the pressure to innovate weighing heavily, other retailers are launching new AI products before they’ve been tested and proven to work as they should, leading to further problems.
Here are three considerations retailers should keep in mind before implementing AI:
1. Clean data
Currently, few retailers have enough data to use predictive AI. However, predictive AI with bad data (or not enough data) is dangerous. It will do more harm than good, because it will lead retailers to make the wrong decisions. However, good, clean data, with the necessary speed and quantity, is hard to come by. It is often in multiple systems in different formats.
The data retailers need depends on the question they want to ask and the problem they want to solve. For example, to optimize inventory and order management, questions might include: “Which locations are at risk of out-of-stock?” “What is the optimal safety stock level for each SKU?” “How often do I ship from the ‘ideal’ location?” What percentage of orders are rejected by stores due to labor constraints?” “What was the average order fulfillment time at each location?” “What are the top 10 items with the highest excess inventory at each location?”
2. Trust and privacy
Collecting and analyzing large amounts of customer data also raises concerns about privacy breaches and cyber threats. Unauthorized access to personal data via AI can undermine trust. With this in mind, the National Retail Federation (NRF) has published its Principles for the Use of Artificial Intelligence in the Retail Sector. According to the NRF, the principles encourage appropriate and effective governance of AI, promote consumer trust, and facilitate continued innovation and beneficial uses of AI technologies.
When integrating AI into their business, retailers need to consider how it will impact the customer experience. It’s important to be as transparent as possible with customers about how the company is using AI to improve the shopping experience, while also taking steps to protect customer privacy.
3. The skills challenge
When it comes to generative AI, most companies don’t have the skills or the money to train generative AI engines. The investment required is significant in both time and money, not to mention organizational change. The Global Workforce of the Future Report 2023 found that about 70% of employees are currently working on generative AI in their workplace. Yet half of them have no experience or training in this area.
Retailers need to review their current skillsets and consider what they need to do in terms of hiring or upskilling before they can do AI well. After all, it can be expensive to jump on trends without a clear understanding of the potential benefits, challenges, and skills required. Investing in tools without people who know how to use them is wasteful. And using AI tools without the necessary skills is reckless.
The future
The potential for business optimization using AI/ML models for retailers is huge. However, it’s important that retailers understand the barriers to innovation and work to get the basics right. The first step is to get their data right based on the problem they’re trying to solve. Then they need to consider: do we have the right organizational skills to execute this effectively?
With so much pressure to innovate, retailers can become overwhelmed by the road ahead. Leveraging modern technology such as inventory data and order management systems provides the reliable, accurate data retailers need to feed into their AI/ML models. This is essential if retailers want to ensure that ‘Project AI’ is set up successfully.
We provide an overview of the best e-commerce platforms for businesses.
This article was produced as part of TechRadarPro’s Expert Insights channel, where we showcase the best and brightest minds in the technology sector today. The views expressed here are those of the author and do not necessarily represent those of TechRadarPro or Future plc. If you’re interested in contributing, you can read more here:
