•  Kili Technology, a French enterprise AI startup, has closed a $25 million Series A funding round.
  • Rather than tuning up algorithms, Kili helps companies prepare messy data for use in AI. 
  • See the pitch deck that lays out Kili’s approach to “focusing on the data first.”
  • See more stories on Insider’s business page.

Just six months after closing a seed round, the French artificial intelligence startup Kili Technology is announcing a $25 million Series A on Tuesday as it pursues a new approach to dramatically speed up enterprise AI

Kili (which derives its name from Mount Kilimanjaro) focuses on preparing data to be analyzed, rather than tuning up the AI algorithms which ultimately process it. The company’s software takes “raw” data — that is, data pulled straight from disparate sources, including contracts, images, recordings, and more — and helps users annotate and “clean” it to use in AI systems. Pouring messy data into AI uses cases slows processes way down so even though cleaning the data adds a step, Kili’s product ultimately shaves down time it takes for customers to go from idea to insights. 

François-Xavier Leduc, CEO of Kili Technology, says the data is like the gas you put into the engine of the algorithm that processes it. For a decade companies have been tuning up engines, and then dumping in gas without thinking much about it. But that means putting in funky gas and getting a clunky performance. 

“Developing AI by focusing on the data first is a totally new paradigm,” Leduc told Insider. “It helps to select the right data in the first place and filters out bias that can be the worst result of an AI process.”

Biased data might incorrectly skew results. For example, if the data used to train an AI system has signals that leads it to assume that data from the US is more important than data from Europe, that will affect the kind of insights it spits out.  

Tuesday’s Series A round pulls together some of the biggest names in Europe’s AI startups sector. The lead investor, Balderton Capital, previously backed London AI powerhouse DarkTrace’s recent IPO, and continuing investor Serena Capital backed $1.4 billion Dataiku, which has offices in NYC and Amsterdam. Dimitri Sirota, founder of BigID, and Chris Schagen, founder of Contentful, are advisors. 

“Enterprise AI really started in 2010 with a huge amount of data, and companies like DataRobot and Dataiku were very successful. They were super good for structured data. But today unstructured data is 80% of enterprise data,” lead investor Maxime Le Dantec, principal at Balderton Capital, told Insider. Le Dantec said his firm was impressed by how much the bootstrapped Kili was able to accomplished with a small staff. 

The company now has about a dozen employees. Prior to Tuesday’s round, the company had raised a total of $6.85 million in funding at a valuation of $25 million, in PitchBook’s December estimate. The company declined to disclose new valuation or revenue. 

Kili says the new funds will be used to accelerate growth through recruitment and tech investment and to move quickly into US markets. 

Here is the pitchdeck Kili Technology used to raise its new $25 million Series A round. 

Source: Business Insider

Leave a Reply

Your email address will not be published. Required fields are marked *

You May Also Like

EXCLUSIVE: Suzanne Morphew’s body was found in Colorado dumping ground known as ‘The Boneyard’ where missing man James Montoya’s body was found during search for Edna Quintana

Suzanne Morphew’s remains were found in a Colorado body dumping ground known…

Illinois supreme court halts end to cash bail – HotAir

On New Year’s Day, a dangerous but hilariously misnamed new law was…

Kinder, gentler Taliban told to reverse ban on women aid workers in Afghanistan – HotAir

Last month the Taliban in Afghanistan insisted that humanitarian organizations fire all…

Tech companies ramping up AI efforts while cutting ethics teams ‘might not be a good look,’ feds warn

The Federal Trade Commission continued its crusade against deceptive uses of artificial…