In this interview conducted on September 12, 2024, we had the opportunity to chat with Emmanuel Rilhac. The conversation helped us understand the evolution of the marketing automation platform NP6, after its acquisition by ChapsVision in 2021. You'll soon discover that what drives them is the integration of geodata (geographical data available in open source) within organizations' RCU (Référentiel Client Unique) data models.
ChapsVision website: https://www.chapsvision.fr/
Emmanuel Rilhac's LinkedIn profile : https://www.linkedin.com/in/emmanuel-rilhac-a40922/
Four years after the acquisition of NP6 by ChapsVision, how would you describe NP6 and your market positioning?
Emmanuel Rilhac: Several factors have changed. When we acquired NP6, ChapsVision had sales of €20 million. Today, we're up to 200 million euros, which is a clear sign of our success. solid financial stability.
Secondly, our approach to marketing has changed profoundly. Originally, we were an e-mail router, a simple activation engine. Today, we offer an extremely wide range of services: CRM, workflows, RCU (Référentiel Client Unique), customer service, sales force, and so on. We can now intervene across the whole spectrum of unified commerce, from digital in-store and clienteling, to payment collection and loyalty. We have evolved from a highly verticalized player in marketing automation to a a complete transformation offer for our customers. In four years, we have gone from 100 to over 120 customers, continuing our organic growth of around NP6.
What really sets you apart?
Emmanuel Rilhac: In terms of marketing scope, I think we go further than most of our competitors, especially when it comes to combining CRM/RCU and marketing automation. Where we stand out is in our approach to unified commerce and digital in-store. Few players go as far as we do in customer relations, which makes us unique. But if we focus on marketing automation, two aspects set us apart.
Firstly, our data model. ChapsVision's DNA is data. We have built a platform capable of processing very large quantities of heterogeneous data. Over the last 18 months, we've been developing NP6 to make this data model more open and flexible. We want to give it an ultra-mega-open dimension. Any customer can implement his data model with all the specificities he wants. They can create as many dimensions, as many tables and as many criteria as they like, and we'll link them together as they see fit. Why is this fundamental? Because we offer our customers the ability to integrate external data into all their tools, particularly marketing automation. I'm convinced that the relevance or performance of a marketing automation program will depend on the knowledge linked to the customer's context. It's a basic thing to say, everyone knows it. But, in reality, the knowledge we have of our customers comes from sales exchanges, marketing reactivity and service exchanges. But if we really want to understand our customers better, we need to understand their context. In France, we are fortunate to have a huge number of data sources available. Some are paid for, but there is a huge amount that is free. We integrate this data natively into our customers' data models. So when a company goes to talk to its customer, not only does it have its conversational history, but it also has contextual data that it didn't have before.
Are you talking about open source data?
Emmanuel Rilhac: Yes, that's it. It comes from open source data that we geolocalize. A few months ago, we acquired Articque, France's number 1 in geo-decisional solutions.
For example, for an insurance company, we're able to define the level of risk to drought, soil swelling, flooding, bad weather, rainfall, for each of its database contacts. So, you can go to them and say: "I warn you. Today, you are not insured for these types of losses, even though you are in a flood zone".. So you can create upsell quite easily, but an upsell that's ultra-targeted. We're not talking about a "I'm selling you additional flood insurance for the Aquitaine region, whereas the person actually lives in the Aquitaine region, but she's on a plateau, so it's not her problem."
I'm using the example of an insurance company, but you can imagine a whole host of scenarios for a social landlord, for example. It could be extremely interesting to help your tenants settle in, not just in their apartment, but in their neighborhood. So you can tell them where the schools are, where the nearest baker is, and so on. In other words, to provide them with a complementary service linked to their own context. At the end of the day, the addition of this contextual data makes our communication much more relevant. This is the first dimension on which we consider ourselves unique, and we'll probably be copied one day.
The second thing we're not unique in, but we've come a long way, is on the hyper-customization engine. We didn't design it just for content, but for hyper-customization of processes and workflows. It's the ability to have extremely advanced rules engines. To take my example from earlier: on the one hand, you have a person in a flood zone with a flood warning, and on the other, another person who is in the same zone, but for the moment, the flood isn't going to reach his home. What's needed is for the automation to be triggered differently in these two cases. There's one person who's going to be exposed to the risk, and you're going to want to warn him with preventive messages such as: "...".Climb to the second floor, save your most precious belongings, etc.." So your automation chain is going to be specific to one individual versus another. You'll have a rules engine that tells you when to communicate and via which channel.
You say you don't specialize in any particular sector, but you offer a complete range of products for the retail sector, and you're giving me some examples of insurance. Are there any sectors in which you're more comfortable?
Emmanuel Rilhac: We don't want to specialize in any particular sector. But there are sectors in which we have made an extra effort to verticalize. When we want to bring more value, we need to be finer and more precise, and to have a whole set of data, additional data that is extremely relevant to businesses such as insurance, mutual insurance, social housing, retail or banking.
In a previous exchange, I expressed my doubts about the harmonious integration of different tools within the same group. You mentioned that you were developing your own range of technologies to solve this problem. Can you tell us more about this?
Emmanuel Rilhac: In fact, when we develop a component for our marketing automation solution, it's also useful for our other CRM, customer service, unified commerce solutions, etc... We won't create a software suite where all our solutions are interwoven. Each tool remains what it is. We want to keep this best-of-breed concept, but enrich it with components that can speak to all the tools. For example, enriching the data model with external data also makes a lot of sense for our customer service or call center solution. So the brick we're going to develop to source the data will be created in a common way, compatible with all our tools. So we're not in favor of merging our applications, but of our tools talking to each other and exchanging data. An NP6 user will see very few changes in the years to come.
Surprisingly, the two most talked-about topics in marketing in 2024 are AI and CSR, two terms that may seem antithetical. How, do you integrate CSR into your own practices, and how do you help your customers progress in their commitments?
Emmanuel Rilhac: The ChapsVision group was created by Béatrice & Olivier Dellenbach. They created a foundation called Happy Capwhich fights against mental handicap. This foundation owns more than half of ChapsVision. ChapsVision has an economic vocation to grow and develop, but for the benefit of a foundation that fights for the disabled. The CSR dimension is extremely important to us. There are a lot of in-house activities to raise awareness among our employees and our ecosystem.
Need help?
Reading content isn't everything. The best way is to talk to us.
Do you integrate AI into your solutions? What's your AI roadmap?
Emmanuel Rilhac: For the past year, we have been deploying energy-saving strategy. There are teams working on engines and product development. And in parallel, we have another team looking at how to make version upgrades more frugal: once we've released version 1, make version 2 more frugal.
On the ethicsFor the past year and a half, we've been trying to implement the directives of the future European AI regulation (AI Act). Our first task was to map all the uses we make of AI. The second was to set up compliance analysis processes for all the new AI we create. It's an analysis of compliance with what we think is going to be the AI act. These are the two things we're doing now.
We are not only a player in the corporate world, but also a major player in the sovereign world. We understand that the State and the major government ministries impose relatively strict constraints on us in terms of sovereignty, ethics and CSR. We apply these constraints and objectives to the Group's activities.
What does AI in your solutions mean in concrete terms?
Emmanuel Rilhac: What we're interested in and will soon be implementing is thePredictive AI. Predictive tagging. This means being able to identify the signals that someone is about to churn (leave the company), so as to anticipate churn. For example, if someone asks about the non-renewal or termination clause in their contract. This is a sign of potential departure. We're currently developing a lot of predictive technology.
The other AI topic we've been working on is the translation. Not all our customers' clients are 100% French speakers. So we want to be able to personalize the message in their mother tongue, without obliging the company's employees to speak or have their message translated into Portuguese, Russian, Spanish or any other language.
We want to work on theAI for data production. We're going to use AI to extract data, format it, structure it and inject it into our models.
On the dimension Generative AI such as content generation and image generation. We're also working on it, but for the time being, we're not going to implement it, because with all the feedback we've had, the ROI and relevance are absolutely unproven. General models like ChatGPT require very specific training depending on the use case you're going to have. You're going to train an AI differently depending on the insurance, banking or retail sector. You also need to train it to avoid hallucination. ChatGPT never answers: "I don't know". If he doesn't know, he'll make something up. That's something we can't accept. We're working on it, but we're not convinced yet.
What are your next product developments?
Joker! I promise I'll reveal them to you as soon as they're released for public viewing.
A final word?
What makes ChapsVision so special is its relevance of our solutions. Each time, we ask ourselves the question of relevance, to avoid creating gadgets. The solutions we've been developing for some time now have a much higher level of relevance. It's hard to appreciate, but that's really what drives us.
The other thing is our French Tech dimension but not in the buzzword sense. We're a technology company, we're French, we're very proud of that, and we're close to our customers. And that's really at the heart of the project too.
