As Insurtech Sector Embraces AI, Startup Zelros Attracts Another $11M in Funding
Among the financial sectors embracing AI technology is the insurance industry, which is looking for more personal and efficient ways to promote its products and services while streamlining underwriting and other data-heavy functions.
Investors sense an opportunity in leveraging AI to finetune the insurance business through the use of technology in what is being called the insurtech sector. This week, for example, insurance tech startup Zelros closed an $11 million Series A funding round led by Silicon Valley-based BGV (Benhamou Global Ventures), the fund founded by former 3Comm and Palm CEO Eric Benhamou.
New investors included ISAI Cap Venture along with Plug and Play, which specializes in fintech and insurtech investments.
Paris-based Zelros said the latest investment brings its total funding to $16.5 million.
Its AI business platform learns and adapts in real time to help insurers achieve a measure of automation in matching potential customers with products and services as well as underwriting policies.
The four-plus-year-old startup said this week it would use the proceeds of its latest funding round to expand its footprint in Europe and North America, including a branch office in Montreal scheduled to open later this year.
The company’s AI platform reviews insurance claims, quotes and underwriting data, then makes recommendations. Last year, it reviewed more than 20 million insurance policies that generated more than two million personalized recommendations.
Training data is gathered from customer interactions along with automated policy reviews. That information is used to update databases, enabling adjusters, for example, to retrieve fresh customer data when claims are filed.
The platform also can be used to automate underwriting, a business process that industry analysts note is suited to machine learning.
Along with large European insurers, Zelros also works with industry regulators on transparent AI applications. To that end, it has published an open standard for the ethical use of AI. The open spec anticipates thousands of machine learning algorithms handling business processes ranging from sales and customer support to hiring and purchasing.
Earlier this month, Zelros released the latest version of its “responsible AI” standard.
“Process automation will be a key trend going forward for insurers to gain trust from customers, by providing immediate answers while ideally positioning them on new distribution platforms,” Paul-Henri Cahbrol, Zelros’s director of pre-sales, customer success and partnerships, noted in a blog post.
The startup promotes its AI-based software as able to ingest a variety of data, including structured customer information, claims and quotes, along with unstructured data such as voice and email communications. A "Voice2Insights" feature, for example, analyzes conversations between policyholders and the advisor in real-time, “allowing the insurance agent to instantly adapt recommendations to meet the specific needs of its customers in a personalized way,” the startup noted in an email exchange.
The platform also uses machine vision tools “to automatically read and analyze documents to detect potential fraud while maintaining regulatory compliance,” the startup added.
Market analysts predict the insurtech sector will increasingly leverage deep learning technologies like convolutional neural networks to process huge volumes of unstructured data generated by customer interactions, including voice and text.
Among the problems AI-based insurtech could solve is the wrangling and analysis of mountains of data generated by different insurance products. The streams for some policies run into the hundreds of data elements. “It’s a lot of data,” said Doug McElhaney, a partner in McKinsey & Co.’s insurance practice.
Companies like Zelros are offering frameworks for ingesting and analyzing those data sets, then generating recommendations. “That’s a flavor of [insurtech] feature engineering that would be help,” McElhaney said.
A McKinsey report concluded that manual tasks like underwriting would “cease to exist” by 2030. McElhaney amended that projection during an interview, saying underwriting would cease to exist “in its current form” by the end of the decade as the balance between human and machine learning shifts decidedly toward the latter.
Overall, McKinsey estimates the “untapped value” of AI applications for insurtech could be has high as $1 trillion.
Meanwhile, emerging insurtech platforms are forecast to generate more than $556 billion in new premiums by 2025, according to market tracker Juniper Research. The total represents a 123-percent growth rate.
--Editor's note: This story has been updated.