top of page
  • Writer's pictureJulia Cames

AI reconciles us with the art of customer conversation

Although the technologies behind artificial intelligence are numerous and diverse, deep learning, big data, and other "mining", one of its most effective and mature applications resembles an ancestral practice: the art of conversation.

The chatbot, a good basic

Having recently interacted with a chatbot, I have to admit that I have gone back on my preconceived idea of the subject, so pleasant was the experience. The brand, a travel booking platform, had presumably got the right people on the other end of the chat, and my issue was resolved within minutes - having certainly been qualified by a good bot. And on second thoughts, I'm not quite sure I was dealing with a 'person' at any point. It doesn't really matter, though, because the conversation happened and the experience was positive. With an estimated 5 billion people in contact with chatbots by 2021, and a projected savings of 8 billion by Juniper Research, the numbers for this technology are very encouraging and support the experience.

The transformation is therefore taking place, since in relatively "transactional" areas, the effectiveness, relevance, and, finally, the intuitiveness of these conversational agents (their French name) are no longer in question. 70% of respondents to a Forbes study described their experience with chatbots as positive overall. The technology has thus succeeded in transforming a relatively cold channel into an everyday service tool. How did this happen, and what do the next few years hold in store?

The field of application of artificial intelligences would therefore be eminently semantic, with real-time contextual analysis certainly best characterising this principle. From the simple tree structure followed on the basis of keyword recognition, AI is now able to distinguish the emotional intention of the correspondent, first of all simply by studying the register of his lexical field (even the typography, absence of punctuation, use of capital letters etc.). Then, AIs must quickly "learn" to adapt, particularly to the linguistic and stylistic specificities of different communities. Another example, instantaneous translation must also be considered, especially by e-tailers addressing multicultural urban communities or of course internationals.

Technology and emotions

A few months ago, an engineer became famous for having declared himself convinced of the sensitivity of artificial intelligence - and was "incidentally" dismissed soon afterwards. This conclusion was inevitably inspired by conversations with his AI, which he presumably found to be of excellent quality. The cinema has also taken up the subject of the emotion of AIs, and these examples remind us, if there was any need, of the need for ethical supervision of these technologies. On the net, manifestos are flourishing in this sense, most of them agreeing on the fact that AI should be at least fair, inclusive and explicit, and ideally secure and responsible. And this is all the more true in an era of deep fakes and the risk of massive disinformation that we have witnessed in recent weeks and that is enabled by NLP (for Natural Language Processing) - but that's another subject.

Connected trolley and Blablacaisse

The elegant paradox that emerges, and indeed the image applies to most aspects of digital transformation, lies in the increasing cohabitation between technology and the humanisation of relationships: Rather than taking us away from them, and by handling the tasks considered the least complex, systems would ultimately bring us closer to ourselves and others; even if it is anecdotal next to the intensity of AI inviting itself into brand/consumer relationships, the connected shopping trolley and Blablacaisse come to illustrate this point.

All around us, automatic checkouts are flourishing, even foreshadowing the forthcoming large-scale emergence of shopping carts or other connected shopping bags. The latter automatically identifies products as they are selected and simultaneously enriches customer information, thereby eliminating the need to go to the checkout. At the same time, initiatives are appearing, such as the Blablacaisse, a symbolic embodiment of "taking one's time" and respect for the rhythm and desires of each individual. The same goes for deliveries, with neighbours or shopkeepers becoming eco-citizen relay points that are a great service!

We are probably at point zero of customer process automation and new conversations enabled by artificial intelligence. Yet things are accelerating. With the arrival of OpenAI's ChatGPT last November and the planned arrival this year of Sparrow, designed by Google Deepmind, the most advanced and the most ancient practices should quickly coexist better and better at the point of sale and in e-commerce. To be continued.

5 views0 comments


bottom of page