Nascent AI innovations hold promise for e-tailers

By

Matthieu Guinebault

Translated by
Nicola Mira

Published
Oct 23, 2018

Customisation, logistics, visual searching, marketplace management, chatbots and more: a study by KPMG for France's e-commerce federation (FEVAD) has assessed current AI tools and their potential applications to e-tailing. It is an area in which AI is currently deployed at a basic level, but one where there are huge possibilities for a more advanced usage.

There are three levels of AI. The first level consists of AI tools that are able to obey to commands, while second-level AI tools also incorporate machine-learning capabilities. Third-level AI involves the ability to think and learn autonomously, to make predictive analyses and put forward hypotheses. Capabilities which, according to KPMG, could pave the way to “screen-less, voice-activated e-tail, incorporating solutions which are potentially very different from current virtual assistants, with 3D visualisation, virtual reality shopping and instant drone-operated delivery based on predictive demand algorithms.”

Currently, a third of all e-tailers have reportedly began to test or use AI solutions that are related to five specific domains. The most common is that of improving the shopping experience through customisation: 70% of the solutions dedicated to recommendation and personalised marketing make use of AI tools. These would eventually enable the large-scale deployment of predictive marketing, which notably allows to tap specific consumer categories at the right time and on the right product, leading them to make a purchase.

Logistics is the domain which is most mature and up-to-date in terms of AI applications. Besides countless warehouse robot applications, AI facilitates the deployment of an omni-channel approach by allowing physical and online stock to be easily combined, accelerating click-and-collect operations and, more recently, paving the way to ship-from-store deliveries, in other words, the fulfilment of an online order using the stock of a store close by. According to KPMG, another major application of AI is visual search, with the widespread use of “visual Shazams”, notably for fashion, enabling the search for products similar to those found on a picture. This visual analysis capability will gradually become more and more embedded in social media too.

But one of the domains which holds the greatest promise in terms of future AI applications is that of data base management. “The big marketplaces run by leading pure players have understood that the core competitive capability for the future will be based on AI and big data,” according to KPMG, which highlighted the growing number of AI tools devoted to classifying the products on offer as finely as possible, to make them stand out more clearly in online searches. For the time being however, the AI application to which consumers are most exposed to is the chatbot. Chatbots have very quickly become ubiquitous, and are programmed to respond autonomously to recurring questions and to hand the query over to a human responder when questions are more complex. An innovation which is linked to that of personal voice assistants, which Google and Amazon are very much relying on.

“Thanks to the explosive increase in data availability and the exponential growth in the power of computers, AI is no longer a futuristic concept, and e-tailers have become aware of this,” said Marc Lolivier, the General Manager of FEVAD. “Nevertheless, e-tailers see that there still are obstacles to an extensive use of AI: the integration with existing IT systems, the time needed to develop and implement effective algorithms, machine errors which necessitate human intervention, etc.” added Lolivier, who warned that “we are still at the dawn of [AI] development.”