AI and machine learning form the backbone of Lemonade’s products, which are designed to help customers purchase policies and file claims efficiently, For Dickman, machine learning is a particular favorite because it is novel to insurance.
“Machine learning has brought a set of capabilities that developed additional options [for] the world that didn’t exist before,” Dickman said. “The challenge the insurance industry has faced until today was not collecting or storing necessary data overall but analyzing and producing insight that will enable companies to give their customers a much more effective and better service.”
Lemonade’s strategy has carried it far, from the raise of nearly $500 million in venture capital financing in multiple rounds to a 2020 IPO that pulled in $308 million, and beyond. But like many insurtechs and overall technology companies, its stock price has plunged since then as investors became skittish about their prospects. Lemonade saw a big jump in losses for 2021 due to catastrophes including Winter Storm Uri and California’s wildfires, though it has generated growing premium, and its customer base reached 1.4 million last year – a 400,000 person increase. At the same time, Lemonade’s annual premium giveback to policyholders’ preferred non-profits dropped this year, based in part on prior year claims performance.
Growth and changing the game
Dickman joined Lemonade in February 2019, helping to build operations as the company generated significant growth, through its launch in the public markets, and its debut of pet, life and insurance car in the US to complement its renters’ and homeowners’ insurance products already in play. Over that period, Lemonade has also expanded into Germany, France and the Netherlands, which has given Dickman and his team ample opportunities to guide technology growth.
“This was incredibly exciting for the engineering team, as we were able to scale from two product-specific teams – or squads as we call them – to more than 200 engineers globally,” Dickman said.
Lemonade, which has been “cloud-ative” since its launch relies on a number of programming types and languages to support platform and services.
Initially, Dickman said, Lemonade launched its debut product on what is known as a Ruby monolith – a single-tiered software application where the user interface and data access code are combined a single program from a single platform. The emergence of car and pet insurance at Lemonade has accelerated its transition to microservices written in Node and Typescript, though Dickman noted Lemonade is in an ongoing process of rewriting old technologies to a uniform one.
The company also relies on Lokalise – a cloud-based localization and transition management system that can offer services in multiple languages, a technology that Dickman spoke highly of.
“While using Lokalise, our development process is much more friendly and flexible, which gives our developers a unique experience,” Dickman said.
In addition, Lemonade relies on branching and APIs, which lets the company streamline its text content within the development process in an easy to control manner, he noted.
Internal tools developed in-house also help support Lemonade’s scale and propel benefits.
“We built an internal automation brain that handles and delivers tasks to production and helps developers manage their workflows,” he said.
Use this, not that
Data and machine learning are extremely important to Lemonade and insurance in general, Dickman said.
“At Lemonade specifically, data and machine learning are deeply integrated into our product and internal processes,” he said. “Interactions with our customers across our platform generate a trove of data, which in turns interactions with our customers across our platform.”
Some of those innovations include Lemonade’s AI Maya bot focused on “playful onboarding and customer experience.” It sells Lemonade policies and personalizes coverage, generates quotes and facilitates payment. There’s AI Jim, a claims bot that handles the “first of notice loss” for most Lemonade claims and often manages an entire claim through resolution without any human involvement.
Insurers should always stay up to date with new technologies that can enhance their products and services, Dickman said, but they should be thoughtful about the process.
“Any technology can be scalable, but the question is at what price,” Dickman said. “To replace tech you should do it mindfully in a gradual process.”
That’s what Lemonade did when it rolled out pet insurance cover.
“When we made the decision to launch pet insurance, we chose to take the opportunity and build it in a new stack and paradigm, and it worked amazingly,” Dickman said. “We also know that it could have failed, but it was a chance we took.”
When it comes to technology that the insurance industry needs and isn’t widely used yet, Dickman places his vote behind a couple of the technologies in play at Lemonade.
“Combining automation and AI with fast and efficient analysis of large quantities of data will lead to a huge improvement in the tech process of the insurance industry,” he said.
Dickman added that doing such a combo will help insurers better met regulatory obligations and standards, among other benefits.