Using retrieval-augmented generation, a new category of chatbots, consisting of smarter and more autonomous systems, is emerging in retail, bringing numerous benefits. What can we expect?

Artificial intelligence technologies are rapidly gaining sophistication. Generative AI now allows for very quick responses to any question, while machine learning and big data can process ever-larger amounts of information more relevantly and exponentially.

As consumers increasingly use messaging to communicate with brands, and chatbots have become a crucial element in customer relations, the French startup Partoo, a leader in the presence management market in France, Spain, Italy, the Middle East, and Latin America, has decided to push the AI envelope further with a new tool that could revolutionize the relationship between brands and consumers. Explanations with Thibault Renouf, co-founder and CEO of this innovative young company.

What is Partoo’s value proposition?

We support 300,000 points of sale worldwide, representing about 1,200 brands, including companies like Carrefour, Toyota, Leroy Merlin, LCL, and Crédit Agricole. We enable our clients to update information related to all their points of sale on Google Maps, Google Business Profile, Facebook, Instagram, Waze, TomTom, Apple Maps, and even the online directory of La Poste. It’s about online visibility. We help points of sale be found by customers.

We have a second offering focused on reputation through a platform that centralizes reviews left by customers on social networks, tracks their evolution, and identifies why some customers are happy and others are not. The goal is for points of sale to improve their ratings. After being found, we help them be chosen.

How do you leverage AI?

We have a service that connects messaging platforms like Instagram, Facebook, WhatsApp, and chats, and answers recurring customer questions, whether basic or more complex, automatically. Thanks to machine learning, this AI functionality generates responses based on the text written in a review.

We also use generative AI tools. Personalizing the response saves a lot of time. Recall that the non-response rate to customer service phone calls is around 40% in some sectors. Additionally, a point of sale typically takes between three and four minutes to respond to an online message. We reduce this time to 30 seconds. Given the number of messages we process, about 500,000, the time saved for our clients is considerable.

So, what does the new tool you’ve developed consist of?

We are adding an additional layer of automation based on retrieval-augmented generation (RAG), a new data processing method that is more efficient and comprehensive. It allows for fetching information from any source, including websites, social networks, and chats. Thus, for retailers, we have developed JIM, a virtual agent connected to the databases of points of sale and with all the information about a product. Unlike a chatbot, it does not only communicate with customers but also with a company’s information systems, allowing it to perform certain actions autonomously, such as ordering a burger on WhatsApp. The customer talks to JIM, confirms the order, and a coded message is automatically sent to the restaurant’s IT systems to validate the operation. It’s a virtual reservation agent capable of placing any order and scheduling any appointment without human intervention.

The goal is to exploit all the information available about a company, its stock, and its products more finely, intelligently, and proactively?

Exactly. The machine allows the customer to have real-time information on the availability of any product in any store and to order it instantly, without doing anything. The machine also helps them easily choose a vacation destination that meets all their criteria and books it for them. With this technology, the chatbot is set to become an automated personal assistant. It’s a real acceleration.

What other innovations can we expect based on this?

The development of virtual agents with long-term memories, capable of analyzing over time the arguments that can persuade a customer to buy. This new AI will find what perfectly matches their expectations over time, not just at a specific moment, constantly updating the information concerning them. This is a particularly promising tool for customer loyalty. This technology goes beyond short-term memory AI, which can only conduct a conversation without building a long-term relationship.

Is this trend global? Is it set to become indispensable?

Yes, it is. The clients most advanced in these areas are international. We work with players who have a global reach in terms of messaging. It is indeed a global trend.

Will this type of solution widen the gap between retailers with these new tools and those with a less advanced AI approach?

Potentially, yes. However, it depends on how AI is developed internally. Some companies prefer to implement these technologies in-house, while others choose to work with partners. We believe the latter will make the difference as they will have access to solutions that cost less, are easier to implement, and drive change more effectively within the company.