This Indian startup just raised $20 million to create new opportunities for human workers to do the tedious but important work that makes AI possible

• Data-labelling company iMerit raised $20 million from the UK government-owned development finance institution, CDC Group, the company announced last week.
• The company is a major player in the AI supply chain, helping machine-learning algorithms and computer vision software to interpret data for industries like self-driving cars, medical technology and e-commerce.
• iMerit relies on an army of humans to help make the world more understandable to machines: Its thousands of employees might be asked to identify all the cars in a video recording, or to check a software-created audio transcript for errors.
• The work may seem tedious, but the CDC Group said it invested the money because of the economic opportunity it was creating, by hiring men and women from lower-income areas to staff its centers.
• The company plans to grow its workforce from 3,000 to 10,000 with the funding, and expand its presence beyond its current centers in India, Bhutan, and the US. CEO Radha Basu says its revenue is growing fast and that it's cash-flow positive.
• Visit Business Insider's homepage for more stories.

Artificial intelligence may be driving some jobs to the brink of obsolescence - but it's also creating some new ones.

iMerit, led by Radha Basu, used that pitch in February to raise $20 million from the CDC Group, a development finance institution owned by the British government. With the latest round of funding, it plans to scale its workforce from 3,000 to 10,000. It also plans to grow its global presence beyond its current bases in India, Bhutan, and the US, Basu said.

The company inhabits a small but vital niche in the AI supply chain, that specializes in creating machine-learning algorithms and computer vision solutions. In order to transform oceans of data into images, speech, or performed tasks, the algorithms first need to learn how to interpret it. That's where human labor comes into the equation.

Complimentary Tech Event

Transform talent with learning that works

Capability development is critical for businesses who want to push the envelope of innovation.Discover how business leaders are strategizing around building talent capabilities and empowering employee transformation.Know More

For the past three years, iMerit has been doubling its revenue, Basu told Business Insider. It has also been cash-positive for the past two years, she added.

"The company keeps learning. It's one of the things that keeps our retention rates higher than pretty much anybody in the industry," Basu told Business Insider. "There's so much new stuff to learn, and I think that's what makes this company so exciting."

Tedious but important

Thousands of iMerit employees label the data that trains algorithms to interpret images, transcribe speech, speak conversationally, and perform tasks. In other words, they might be asked to watch a video and note every house, car, and human they see, or to rate the accuracy of a computer-generated audio transcript.

The work may initially sound unexciting, but it lays the foundations for AI-powered algorithms to potentially revolutionize medicine, manufacturing, self-driving cars, e-commerce, and social media.

This makes iMerit's work significantly different from traditional enterprise software outfits, as iMerit CEO Radha Basu points out. "Advances in technology make things change very rapidly," Basu said. Unlike enterprise software, the industry's rapid pace of change requires iMerit stay "nimble and agile," with a "maniacal focus on clients," Basu added.

Last year, iMerit revealed that one of its partners was Amazon's AI service, SageMaker Ground Truth. But Amazon is not the only big client that relies on third-party data-labelling companies. The market to train algorithms was estimated to be worth $1.7 billion in 2019, according to a January report from analyst firm Cognilytic. The report estimates that the market will grow to be worth over $4 billion by 2024.

And data-labelling companies like iMerit and its competitors like Samasource (founded by Leila Janah, who recently died at the age of 37) and Figure Eight (previously known as CrowdFlower) have been snapping up business.

Creating job opportunities

Labelling and feeding data into algorithms may sound like tedious work, but it also appears to be creating opportunities for its workers, who are based in India, Bhutan, and the US.

Basu, who founded the company based off the conviction that her company could both "be inclusive" in its hiring and be a leader in the AI industry, says that she's proud of the social impact of iMerit.

iMerit's workforce is evenly split between men and women, and the average employee is 24 years old, Basu told Business Insider. And the majority of its workforce was recruited from lower-income neighborhoods and less urban areas - a dynamic that helped to convince its latest investors at the CDC Group to put their money into the company.

"We started thinking about the future of work about a year ago," Rob Jordan, an investment manager at the CDC Group said, describing the circumstances that led the institution to start looking at the AI industry. iMerit was the institution's first investment in the AI industry, after "it quickly became apparent that something special happening here," Jordan said, estimating that about 80% of iMerit's workforce came from lower-income areas.