The Top-10 French Artificial Intelligence Startups
Table of contents
As France’s youngest president at 40 years old, Emmanuel Macron is known for his strong handshake, boyish good looks, and his controversial method of selecting a mate. Recently, he set his sights on artificial intelligence, announcing a government program to invest $1.8 billion into AI over the next four years. The cornerstone of all this spending will be in the healthcare sector, where the aim is to make predictive and personalized care a reality using AI and big data. This prompted us to take a look at the ten most-funded French AI startups according to Crunchbase.
Founded in 2014, Shift Technology has raised $40 million to develop an AI-based insurance fraud detection service. Their algorithm sorts through huge data sets to flag fraudulent claims that cost $80 billion a year globally. Since we first came across this startup in our article on “10 Artificial Intelligence Startups in Insurance”, they’ve secured another $28 million in funding which is being used to expand coverage to new types of insurance, open offices in New York and Tokyo, and develop an automated claims handling service debuting later this year that looks something like this:
Pilot testing of the new service on a sample of one million claims documents resulted in a 97% accuracy ratio, which is better than human claims handlers. Shift Technology estimates it will save $50 million annually in compensation costs and $20 million in fraud leakage for every million claims they process.
Founded in 2014, Prophesee has raised $37 million so far to develop a “neuromorphic vision system” that mimics the human eye and brain. Covered in our recent article “Watch for These 8 AI Startups Doing Computer Vision”, Prophesee’s technology provides an output equivalent to more than 10,000 frames per second, even in extreme lighting conditions, and manages to do this with high power efficiency and less data processing. That’s 10-100X better than the human eye, depending on who you ask. This efficiency gain is due to how their AI processes the images it sees. It looks only at changes, as opposed to taking thousands of pictures and comparing each one. Use cases include the following:
Using this technology, product designers can add multiple sensors without losing battery life and computational power, making Prophesee an ideal platform for connected, autonomous, and mobile applications in fields like automotive, robotics or industrial automation. Prophesee is currently evaluating the idea of implementing the same event-driven approach to radar and LiDAR outputs as well.
Founded in 2010, Tinyclues has raised $25.4 million to develop an AI-driven campaign marketing solution that helps Business-to-Consumer (B2C) marketers drive revenue and engagement from customers. Their service unleashes machine learning on your marketing database and draws conclusions based on seemingly unrelated “tiny clues” that your information provides. Algorithms combine facts like purchase history, browsing methods, address, and newsletter click through to assess customers’ likelihood of engagement and purchase.
Tinyclues removes the need for arbitrary (and usually biased) manual profiling of customers like “female, housewife, aged 30-40, newsletter subscriber, spends $200 on average purchase, prefers electronics”. Operating in 10 countries now, the platform has recently expanded to North America. In sifting through their list of success stories from companies you’ve largely never heard of, it looks like the end result is a 49% increase in marketing campaign revenues for the average user.
Founded in 2006, Linkfluence has raised around $24 million to develop a social media intelligence tool that scrapes through all the drivel people post every day. Employing machine learning, deep profiling, and visual recognition, their Radarly service captures and analyzes 150 million sources including social media, blogs, forums and websites across 190 countries every day, they claim. Users can search and identify keywords associated with their brands, competitors and sectors, profile their client base, and improve social media presence.
The resulting insights help drive company strategy and product development. Linkfluence already boasts an impressive client base that includes companies like Coca-Cola, McDonald’s, Ford and LVMH (the people that sell those ridiculously overpriced handbags that your significant other buys). They’re not alone in this game however, as a few years back we highlighted at least 4 other companies building social media listening tools, every one of which has raised more money than Linkfluence.
Founded in 2013, Snips has raised $21 million to develop a context-aware AI-powered virtual assistant. The company’s value proposition is to avoid cumbersome app interfaces and use voice recognition technology instead. Does this idea sound familiar, Alexa? Since our first article on Snips back in 2016, they’ve shifted from marketing a virtual assistant to rewiring and embedding their voice platform into household products. Snips received an additional $13 million funding to market this platform, becoming a rival to Google’s Home and Amazon’s Alexa in the process.
The embedded voice assistant is private and secure, operating on the device rather than connecting to the cloud, unlike competitors’ offerings. It will be offered as a component for a fixed price rather than priced based on queries, and will be available to consumers as well to adapt for themselves using a Raspberry Pi. Snips has recognized that limited language understanding is enough to control speakers or vacuum cleaners, and their platform is custom taught for the product it is fitted into. Because the AI learns flexibly, customers can also opt to teach their assistant themselves, and use it across different platforms.
Also founded in 2013, Saagie has raised $12.4 million in funding to develop a plug-and-play big data platform for corporates. “Saagie” is also how you pronounce the Japanese word for heron, a migratory bird living and hunting near lakes. Similarly to the heron plunging its head into the water to grab its prey, clients of Saagie can use the platform to plunge into their vast pools of unstructured data (or data lakes), and find the right pieces of information for predictive analytics:
Moving away from the migratory bird analogies for a second, the platform stores, logs, and organizes company data, making it straightforward to run AI algorithms against it. It even monitors user consent and anonymizes sensitive data. Use cases include customer relationship management, sales, risk management, and process automation. Saagie has a number of big-name clients like Johnson & Johnson, and has partnered with the aforementioned Shift Technology among others.
Founded in 2011, Allo-Media has raised $10.8 million in funding to develop vocal cookies that use speech recognition and semantic analysis to transcribe customer calls and flag important triggers throughout these calls. According to the company, telephone is a more efficient channel than the web, generating stronger leads and more sales. Yet, it remains underutilized because it is handled by humans who have to pay attention to the caller and cannot efficiently transcribe and flag all the important details. Enter Allo-Media:
Their vocal cookies use machine learning with human feedback on semantic algorithms to automate call qualification and follow-up. The team has more than 500 clients in 15 different industries and is in hiring mode, so it must be working well. The value proposition is in automating a significant part of pre-sales, sales support, and customer support jobs, which will, of course, free up “John in Mumbai” for “more value-added activities”.
Founded in 2016, Kayrros has raised $10 million from Index Ventures to develop a big data analytics platform for the energy industry. Led by senior energy consultants, mathematicians, and economists, their team aggregates huge amounts of data globally on oil, gas, and energy infrastructure and turns this into actionable insights using machine learning.
Analysis output ranges from crude oil inventory estimates to individual company performance forecasts based on satellite imagery, road traffic, customs data, and public reports. Clients of Kayrros’ “weather forecast for oil & gas” include governments, NGOs, financial institutions, and energy companies. The company has just had its second anniversary and its team of 80 is reportedly busy developing a new product “focused on the balance between energy sources and the environment”.
Update 09/20/2018: Kayrros has taken in an additional $24 million from a Series B Round led by Cathay Innovation to bring new solutions across energy markets and expand their offering towards petroleum products, power generation, natural gas, and renewables. This brings the company’s total funding to $34.4 million so far.
Founded in 2008, Streamdata.io has raised $9.4 million to develop an automated data streaming service that enables live updates for small and large consumers of data. Traditional models for websites and apps update when the data consumer requests it, much like when you refresh your browser manually. With Streamdata.io, all changes are pushed to the data consumer automatically creating a real-time flow.
This is particularly useful for data scientists developing and feeding their machine learning models in real-time, replacing cumbersome and inefficient batch processing where large amounts of data are uploaded from time to time and the system is idle between uploads. (Just how do users go about troubleshooting data problems in their models when the datasets are constantly changing?) Pricing for the service is scaled based on the data flow which means even small businesses can use it.
Founded in 2016, Foxintelligence has raised $9 million to develop a secure, anonymized e-commerce intelligence platform. Data aggregation and analysis is available for brick-and-mortar retail segments through companies like Nielsen, but this kind of market intelligence does not cover e-commerce:
Foxintelligence aims to change this by collecting information from the consumers. The company has released a newsletter unsubscribe app called Cleanfox that lets users efficiently clean up their inbox. In return, they are asked to share their online purchase receipts anonymously. (Thanks, but no thanks.) The data is sorted, aggregated, and analyzed by AI algorithms. The output shows what consumers buy, where, and for what price, highlights market trends, top players and their turnover, lists best-sellers, and the list goes on. No launch date has been set for their intelligence platform yet.
You may have noticed something odd about the above list. None of these startups focus on healthcare, which is where the $1.8 billion in investment money from the French government is expected to be allocated. $1.8 billion is a big number, more than ten times the total funding these startups have taken in so far. As it turns out, there are quite a few AI startups in France, and some entrepreneurial Frenchman put together an initiative called “France is AI” which contains more than 280 French AI startups. Of course, that list doesn’t match the Crunchbase list which brings us to our final thought.
If we missed your cool French AI startup that should have made this list, please go update Crunchbase. That’s where we pulled our list from. If you have founders with funny French names but you list your primary location in Crunchbase as San Francisco, then you wouldn’t have made it on the list (we’re looking at you, Algolia). What we will be doing next is pulling another list of healthcare-related AI startups from the “France is AI” initiative and doing a future article on that topic, so stay tuned.
Become a premium member and get access to hundreds of premium articles, reports and additional content.
Nanalyze Premium is your comprehensive guide to investing in disruptive technologies. Read by the top investment banks, management consultancies, VCs, and research houses. Trusted by over 100,000 institutional and retail investors. Covering disruptive technologies for nearly two decades.