I studied in Sciences Po & HEC with a gap year in Berkley. I started working on my project while studying thanks to observations I made on the American campus of Berkley. Once I graduated, Loïc and I decided to found Jam. His profile is more technical than mine. We were incubated in Sciences Po, then in Numa in 2015. Jam was officially launched in April 2015.
Jam is an artificial intelligence service, supervised by humans, which allows students to ask for any information simply via a text message. You can ask for the closest sushi shop or for a nice bar to bring a date. Jam gives you a immediate answer and makes you win a precious time.
When I was in Berkley, I quickly realized how hard it was to find accommodation on campus, a carlift to get to San Francisco, or even people to teach French lessons to… None of it was centralized. I wasted a lot of time looking for all this information. I started thinking about a service that would make it easier for students. At first, we wanted to create a platform on campus that would let students share their good plans. After a few thoughts and tests, it turned out to be too complex to set up. After many discussions with students, we went for a simpler solution: sending the information they need by SMS.
The hardest part was the “scalability” of our service. It was quite easy to give information to 10 students at the same time, but much harder for 100. Today we send between 4000 and 6000 text messages each day, we had to find a way to meet the growing demand.
We had the chance to be supported by Business angels who believed in us, especially ISAI. It is essential for us to be financially supported, this way we can work serenely. We also benefited from infrastructures in the Sciences po and Numa incubators; it really helped us! We also have a great team, that is able to adapt to a rapidly evolving ecosystem, this helped us to adjust our service.
Our platform is developed using the Ruby language. We are linked to Yelp, Foursquare, Airbnb, Youshould or even Seloger by API. We automate answers thanks to previous requests from students. When a student asks for something completely new, unknown from our systems, we treat it manually. This means a member of our team searches for the answer himself and checks if the information is correct. We then save it as a template to automate the answer.
Today, none of our competitors has the same model. We also have a completely different positioning, by targeting students only. With Jam, we really want to make the interface between men and services disappear. We have created a service that voluntarily moves away from traditional maintenance. Students want a quick and clear answer when they address us. Allowing a choice is a waste of time. We give personalized advices.
You can ask any service via text or Facebook Messenger on Jam since April 2015. We count more than 30.000 users.
I’ve been following Airbnb since their beginning, it’s really been inspiring to me on many aspects. I particularly admire their ability to treat a localized problem and to solve it in a simple and iterative way. That’s what we tried to apply with Jam.