Author transcript
Sean Nolan: When online shopping first became available, the customer journey changed forever. Now we're gearing up for what could be the next revolution in retail, that being AI and agentic commerce. Here to discuss how AI is changing customer experience in retail, I'm delighted to be joined by Nicolas Salat-Baroux, CRO and co-founder of Joko. Hi Nicolas, thank you so much for joining me today. For those of you who aren't aware, could you start by giving us a bit of background on Joko and how the platform works?
Nicolas Salat-Baroux: Yes, of course, and thank you very much for welcoming me. It's a pleasure to answer your questions today. Basically, Joko is an AI shopping assistant, helping consumers find the best products for them and get them at the best price.
You can just download Joko, search for anything, and we will give you the best product recommendations based on the catalog of everything that exists online. So we built a universal product catalog with more than 100 million products. We know our users very well, thanks to our access to their shopping data, and we can recommend the best product for them at the best price.
And the best price, because not only can you find the best product for you based on your taste, the brands you shop at, and your size and so on, but also the best price thanks to features like cash back and price comparison. We also automatically apply coupons for you. You can buy now, pay later through Joko. So basically, it's the best product you can imagine to find the right product for you and get it at the best price.
Sean Nolan: Yeah, absolutely. Thank you so much. And you mentioned there about AI playing an important role in Joko. Could you tell us a bit more about how AI is integrated into the platform and how it helps serve the customer?
Nicolas Salat-Baroux: Yes, definitely. Maybe I can give you some perspective about what we are building and how AI is helping us. Basically, the ambition we have is to build a new front door for shopping, so that anyone having to search for something opens Joko first and searches for the product.
And why we can do it is because we have three things that only Joko has. The first one is what I mentioned about the universal product catalog. We have a catalog of products that is basically the catalog of the entire e-commerce. So we have everything that is available on Amazon, on Walmart, on Nike, Sephora, like everything.
The second thing we have is proprietary data on user spending, because when users download Joko, they link it with their bank accounts, giving us access to all the spending online and in store. They also download the extension, giving us access to all the products they are browsing online, buying, adding to their baskets, that they bought. So thanks to this data, we have a very, very good understanding of our users. And then we have trust, because they trust us on offering them the best price, and they trust us linking their bank to Joko.
When you look at the ecosystem, Amazon for instance, it's very good to search on Amazon, but you are only limited to the inventory of products that is available on Amazon. And in a lot of categories like fashion, home and decoration, and so on, the inventory is not that great, and they don't always have the best price. You can search on Google, but on Google, very quickly, you end up opening thousands of tabs, comparing prices on the different products, searching for the different reviews. And then if you want a promo code, opening a new website to try promo codes that are not working. So the experience is not great. And then if you look at one retailer specifically, the inventory is very limited.
So we have a larger inventory than Amazon, because we have access to everything. A much better experience than on Google, because you can just search on Joko, and we find you the best product based on everything without having to open tens of tabs. So not only do we have the best catalog of products, but we also have the best place to personalize and find the right products for you based on your taste and on your shopping habits.
That's how we can offer the best experience to users, because ultimately what you want as a consumer is to just find the right thing for you at the best price, and you want one single entry point. And AI is helping us with everything. AI is helping us build this product catalog, because it's a real technical challenge to have access to fresh and accurate information on price, availability, colors, and so on, delivery information, reviews, on the whole internet catalog of products we have. It's also a challenge, and AI helps a lot, understanding users based on their shopping history, to make sure they can find in this catalog exactly the right products for you and suggest what is good for you. And then AI helps us because as a user, you search in natural language on Joko, saying, "I'm looking for this kind of product, can you help me search for it?" And basically understanding the query, interacting with the users as you can interact on products like ChatGPT or the others, requires AI. So AI is everywhere.
Sean Nolan: I wanted to ask about what you mentioned there at the end about natural language and how users can search for products. If I'm not wrong, for example, if I wanted to find a particular type of jumper, I could say, "Find me a blue jumper at this price, made of this material," this kind of natural language search. Why do you think that's an important value add for customers? Why was that quite an important part of what you were building, to have this natural language front door for customers?
Nicolas Salat-Baroux: Yeah, I'd say it's very important, because exactly as you said, when you're searching for something, let's say I'm searching for a new pair of running shoes for a beginner training for a marathon, and under like $100, things like that, it's much easier to ask it in natural language than going on a website and clicking on all the different filters, because the filters are not matching necessarily exactly what you are looking for and are not available. So it's much easier to do it in natural language.
Another thing that is very important is that a lot of the valuable information is in the product descriptions. On the product descriptions of a product, there could be information like "this product is much better for extreme temperature if you are going to the mountain" and so on, but you don't have any filters like that. So AI is powerful to scan all this product information, all these reviews, and to answer much more precise questions that you wouldn't be able to answer in a world of pre-natural language.
I also think that when you look at the future of shopping, a lot of shopping behavior will be done through agents and using agents. Basically you have two kinds of products. You have the ones you used to shop that are low price, where you can 100% trust the agent to do everything for you. For instance, "order for me the same salad I ordered yesterday on Uber Eats," or "buy again these black T-shirts I used to buy from H&M," or "buy more coffee pods from Amazon because I'm out of stock." All these things, it's much easier to do in natural language because it's the natural way to do it. It saves you time. You can talk to the agent to do it and not click on different tabs.
And then you have also the products which are very high value, like you need to buy a new TV. So maybe you will not trust the agent for the whole process of searching for the right product up to the checkout and paying for you for a thousand-dollar product. But in any case, the agent and AI in general will be part of the journey. Because even if you do the whole search yourself and you end up finding the right TV, you will tell the agent, "Okay, this is the right TV, but can you check if it's not available somewhere else at a lower price?" Or "Can you make sure there are no coupons available somewhere that I can apply?" Or maybe if you're looking for, I don't know, a high-end designer bag, "Can you check if it's not secondhand, very high quality, on the secondhand websites?" and so on.
So at some point, AI and your agents will help you during this journey, and it's much easier to do it through natural language. It's the same way when you're in store and you talk to someone to help you find the right product for you. It's natural, and I think it makes a big difference.
Sean Nolan: Yeah, definitely. I think what I've noticed around a lot of the features that you've mentioned is that so many things that AI can do are things that humans could do. We could filter by different products, we could read every product description of every item, but the thing is we just don't want to, because it takes so much time and so much effort. So actually what AI is doing in one of the features you mentioned is really cutting down on the inconvenient sides of shopping and letting people focus more on this kind of faster customer journey. I think overall it's probably a higher satisfaction with the customer journey that they're on.
Nicolas Salat-Baroux: Definitely, I totally agree with you. I feel AI, in the context of shopping but everywhere, is giving power back to consumers and helping them really save time and save money. Because basically when you buy something online, you don't want to spend 10 hours comparing prices everywhere, making sure the coupon works, waiting for the price to drop, because there will be a price drop with a high probability the week after, and so on. You don't want to trade this time to get the best price. But in a world of AI and agents doing the whole work for you, you will be sure that you will get the best price and save money very efficiently.
Also, today, because you don't want to spend this time, you end up buying one of the first-result products recommended by Google on shopping or by Amazon, but very often these are the products that are more expensive than on other websites, because brands are paying to be highlighted in the first results. So I feel AI will give power back to consumers, and it will ultimately unlock a massive purchasing power increase for consumers thanks to it.
You can think of adjacent topics to shopping. For instance, today there is a huge amount of money spent on unused subscriptions by consumers. It will not be the case in an agent world, where you can just ask the agents to check all the subscriptions and automatically cancel the ones you don't want. Or you don't want to spend time calling your energy provider to renegotiate your bill, but if AI can do it for you, you will get the best price. And I think if you project, if I go back to this TV example, you can ask your agents to go and say, "I'm on the market to buy this TV, who is the retailer who can offer it to me at the best price?" So I really feel that AI is giving power back to consumers, and it will unlock a very important purchasing power increase and reduce all the money spent that could not be spent, because there is a kind of information asymmetry between consumers and retailers.
Sean Nolan: Definitely. I think it's a major benefit for businesses as well, because if you're a business offering one product at the lowest price, and your competitor is offering the same product but for more expensive and maybe putting more spend into ads, and therefore they're getting more business than you, actually with agents, the business that's offering the lower price will be rewarded for that, and the consumers will go there rather than going to their competitor. So it actually evens out the playing field a little bit for retail businesses as well.
Nicolas Salat-Baroux: Definitely. And it's not only a matter of price, it's about the right product and the product quality. Maybe you have two players on the market. One is very big, with very large marketing budgets, able to basically take the market on a specific category of product. At the same time you can have products, maybe at the same price, but with a much better quality, much more advanced features, I don't know, that are suiting you more. And today you can't find this product if you go with the one spending the most budget on marketing. In an agent world, where you will have access to the time to find the right product, I think ultimately it will probably reward the best product. So there will be an incentive in the market, and around competition, on building the best product. Because if you build the best product and it's known, you will have many more sales than what you have today in a world where it's 100% dependent on the marketing spend you put, or very dependent.
Sean Nolan: Yeah, absolutely. That's a really good point, thank you. I wanted to ask a few questions based on your experience with AI and building Joko. Do you have any advice for retail businesses on how they can deploy AI? Because I know one really common challenge for a lot of businesses is how do you deploy AI that actually creates value, and it doesn't just create noise and it's not just a fancy toy to play with that doesn't actually add much ROI. So do you have any advice for deploying AI in this particular sector?
Nicolas Salat-Baroux: Yeah, so for retail in particular, I think there will be a very big impact on customer experience if AI is applied as much as possible on retailers' websites or services. For instance, in terms of product recommendation, we discussed it before, but as a retailer you can give much better product recommendations thanks to AI, with natural language, better understanding of your own products, screening all the reviews and descriptions, and understanding the needs of users. So I think this should be implemented. There are lots of features like virtual try-on for products, or visualizing a new sofa in your living room very efficiently. These kinds of features I think create a lot of value for users.
Of course, one obvious one is around customer care, because you can now have 24/7 excellent customer care thanks to AI for all your customers. That was not really possible in the past, pre-AI. So I think all these blocks are kind of obvious. Of course, retailers should not try putting AI everywhere when it's just noise and not useful. But on these specific examples, and I think you can have lots of other examples, there is huge value created.
And then what retailers should do in the world of AI agents and shopping agents is make sure their catalog is accurate, the catalog of products, with the right product information, price, availability, inventory, and so on. Because whether it is on your website directly, or when asking for a product sold on your website through an agent, the agent should be able to have access to the information with the right quality, because it could be very deceptive for consumers if there is wrong price information, stock information, availability and so on. So I think there is a huge amount of work currently to be done on the retailer side to make sure their inventory of products is easily accessible and accurate.
Sean Nolan: Yeah, no, that's absolutely a really good point. I think how businesses list items, how they create their descriptions, the kind of information they include, all this is going to have to keep up and adapt with shopping agents as they become more common. And it's really important, as you say, to empower consumers with that information so they can make the best decision for themselves. So yeah, thank you, that's a really good point and something really important.
I wanted to ask you about the development of Joko as a business, and your experience as a co-founder. If you could go back in time to when you first started with the business and first started harnessing AI, is there anything you would do differently? Is there any advice you'd give to other business leaders, with your experience with AI and what you've learned?
Nicolas Salat-Baroux: Differently, I think I would start earlier, definitely. We started, for instance, allocating more than half of our engineering resources to AI and our AI shopping assistant two years ago. I think we should have started earlier. We also completely transformed the company in the past year to operate as an AI-native company, as a company founded a year ago would operate in the time of AI. I think this as well we could have started earlier, because the productivity improvements when operating this way are basically massive.
To give you one example, we launched in the company what we call "anyone can ship." Basically, anyone in the company can ship product improvements and new features, people from the marketing, care, and commercial teams with no technical background. 100% of the employees, and we are more than 130 now in the company, shipped something in production, a feature improvement and so on, and it had a massive impact on product velocity. So I think it's great that it's moving in that direction. We should have started earlier, and it's a good learning for the future. I think we will now try, as early as possible, adapting to the new ways of working, shipping, developing the product. And I think in this world, if you want to be relevant, it's probably worth being earlier than the opposite.
Sean Nolan: Yeah, no, thank you. It's really fascinating to hear about that policy within the company, "anyone can ship." I think it shows that when different people within the workforce are empowered, they can help improve efficiency. And I assume that maybe the IT team is not going to be overworked, because everyone has the ability to chip in and help with that workload that traditionally the IT team would be forced to deal with. So I can imagine it helps a lot with the efficiency overall.
Nicolas Salat-Baroux: Definitely. And yes, when you improve the product thanks to AI and improve how the company operates thanks to AI, you can also have massive productivity gains and output from the team. And Joko has been growing very fast in the past years thanks to AI. To give you a few figures, we have a bit more than 7 million users now. We generated more than $100 million in purchasing power to users thanks to the savings features. More than $50 million in revenue in the past year. More than a billion in sales for retailers on the platform last year. And it's really accelerating, so growth is accelerating, and it's AI and the way we implemented it in the company and in everything we do that really had a massive impact on this growth trajectory, and it's accelerating.
Sean Nolan: Yeah, absolutely. Thank you so much. I think it's really fascinating to learn about Joko's progression, and I think you're in a really good position to give us a prediction about how businesses in the future will deploy AI to create better customer experiences. Do you have any insights into how you see AI being used in the future? Maybe something that's not possible right now but something that might be available in the future.
Nicolas Salat-Baroux: Yes. I think we talked about a few things that businesses, retailers in particular, are doing with AI to improve the customer experience, things like virtual try-on, customer care, better personalization. All these things will have a positive impact on consumers. But what's really exciting at the moment, with everything happening around AI agents and personal AI agents, is that B2C is back and is booming again.
Because when you think of it, ChatGPT was built in the first place as a B2C product to create value for consumers, improve the consumer experience, and it unlocked a huge value for consumers. Then Claude arrived, and basically the past two years for OpenAI and ChatGPT in particular was everything about catching up on Claude on the B2B side, because of course businesses are able to pay more than consumers for AI. Revenue growth accelerated so much at Anthropic. Then OpenAI put all the resources into their B2B business, at the expense of the B2C one, and without really improving the B2C product. And when you look at all the Y Combinator companies, for instance, in the past years, 95% of them were B2B companies helping businesses improve thanks to AI.
But now, with all these AI personal agents like Muse, Instinct and all the others, it's booming again on the consumer side, and on AI for consumers, and it will be to the benefit of consumers ultimately. Because when you talk about how AI will improve customer experience, I think there will be a massive improvement in people's lives thanks to AI. So it's very good news that consumer AI is sexy again and growing again, because it will be to the benefit of consumers.
Sean Nolan: Yeah, absolutely. I think one thing that we're seeing on our side is that personal AI agents are going to be really, really important in the customer journey moving forward. And I think agentic commerce is something that we're going to be following very, very closely. So maybe we'll have to invite you back soon to talk more about that and hear more of your insights. But yeah, thank you so much for your time today, Nicolas. It's been great speaking with you. I really appreciate it.
Nicolas Salat-Baroux: Thank you so much for the invitation, and talk to you soon.
Sean Nolan: Thank you so much for watching. If you want to stay up to date on the future of customer experience and AI, create your account on CX Today now.
