Best of Show winner CrowdFlower is leveraging Microsoft’s Cortana Intelligence Suite to bring new AI functionality to its human-in-the-loop platform for data scientists. With CrowdFlower AI, the company will apply artificial intelligence to a range of business processes including support ticket classification and sentiment analysis, accessing Microsoft’s technology via its Azure Machine Learning cloud service.
“With the Azure Machine Learning offering, the CrowdFlower platform now allows data-science teams to train and deploy machine-learning models with humans-in-the-loop,” says CrowdFlower CEO Robin Bordoli. “This is an exciting step toward making AI a business reality rather than a science project.” Writing at the CrowdFlower blog, Bordoli described CrowdFlower AI as the product of a “shared vision” between his team and the team from Microsoft Azure Machine Learning to bring machine learning to the enterprise that is a year in the making.
Bordoli, who took over the helm as CEO in February 2015, also emphasized that the combination of human and machine intelligence was a critical factor in what makes CrowdFlower’s solution unique. He added that it was also a better description of the likely future of AI compared to the way AI is portrayed in the media and even the way it is used by the “technology elite” like Apple, Facebook, Google, and Uber. “Machine learning without human-in-the-loop leads to bad outcomes,” Bordoli wrote, referencing the incident this summer when Facebook fully automated its Trending feature only to have its algorithms publish fake news stories. “AI is machines augmenting humans, not replacing humans,” the CrowdFlower CEO explained. “Machine and human have complimentary strengths. AI is about the art and science of combining the strengths of machine learning and human intelligence.”
A data-enrichment platform for data scientists, CrowdFlower specializes in managing data from an on-demand workforce of more than five million contributors. The company’s platform automates management tasks like data categorization, image annotation, and transcription that require human intelligence, and has been used, for example, to help automate data collection, learning and labeling from financial transaction data. CrowdFlower’s customers include Flickr, Bloomberg, and Home Depot.
Founded in 2009 and headquartered in San Francisco, California, CrowdFlower demonstrated its platform at FinovateFall 2014, earning a Best of Show award. The company has raised a total of $38 million in funding; its most recent investment was a $10 million Series D completed in June.