Ki: Reinventing commercial insurance with algorithmic underwriting
About Ki Insurance
Ki was launched with the aim of helping redefine the commercial insurance market for the digital era. As the first fully digital and algorithmically-driven Lloyd's of London syndicate, Ki provides a completely different experience for the Lloyd's of London broker, enabling them to obtain a quote in seconds rather than days.
Tell us your challenge. We're here to help.
Contact usKi leverages Google Kubernetes Engine and BigQuery as the foundation of its fully-digital insurance platform.
Google Cloud results
- Supports Ki to build a ground-breaking and fully digital commercial insurance platform in six months
- Reduces resources spent on model training and retraining by up to 90%, enabling more iterations over a shorter period
- Facilitates improved governance with an audit trail that helps Ki to trace it’s dataflows
Six months from conception to release
Lloyd's of London has been a fixture in the heart of London for more than 300 years. The world's oldest insurance market, which began as a coffee house in 1688, is one of the last financial markets where business is done face-to-face. The company is one of the world's leading marketplaces for commercial, corporate, and specialty risk solutions, from satellites to ships. In its iconic underwriting room, brokers and underwriters connect in person, negotiating policies like they've done for centuries.
Although work is well underway on Lloyd's of London's flagship Blueprint Two programme to digitalize the market, Ki is the first fully digital Lloyd's of London syndicate, taking a data-led, and unique algorithmic approach to underwriting. It's a first-of-its kind digital insurer for the market that specializes in complex and unusual insurance risks, from offshore wind farms to athletes' limbs.
"Businesses go to Lloyd's of London to insure assets that are incredibly complex," explains Richard Hodgson, Chief Technology Officer, Ki. "This complexity is one of the many reasons the market has been more difficult to transform. Ki combines the know-how of insuring these specialized assets with the speed and agility of a tech startup, all within a highly regulated environment. And we're doing it with Google Cloud."
Ki and Google Cloud: Sharing the desire to innovate
When the concept for Ki was being formed, the team were looking for a partner to collaborate with on the project, and quickly found common cause with Google Cloud. "This had much to do with the innovative mindset of the people and the engineering culture at Google Cloud, whilst clearly having a track record of delivering new technology into traditional industries," says James Birch, Development Director, Ki. "We're out to create a completely different experience of trading in the market by building the first fully digital business in specialty insurance, and Google Cloud was undoubtedly the best partner to do it with."
In less than six months, Ki went from zero to launch. By working closely with the Office of the CTO of Google Cloud (OCTO), a global team of senior Google technology experts and former enterprise CTOs, Ki was able to quickly move from ideation to creation. At every stage of design, the needs of the users – namely brokers in the Lloyd's of London market – were front and center. Today, these brokers can use the Ki platform to get quotes almost instantaneously. The solution is accessible anywhere, at any time, and offers instant capacity – a true game changer for the market and its participants.
"We're out to create a completely different experience of trading in the market by building the first fully digital business in specialty insurance, and Google Cloud was undoubtedly the best partner to do it with."
—James Birch, Development Director, Ki InsuranceThe speedy release of the platform, which started trading in 2021, was the result of seamless collaboration between three parties: Ki and its insurance expertise, University College London's (UCL) leading academic knowledge of algorithmic & quantitative trading, and the technology foundation of Google Cloud. "It didn't feel like three different businesses - we fully embedded the Google Cloud and UCL teams into ours," says Birch. "We operated as one, collaborating in a cross-functional setting, which enabled us to go to market quickly."
Powering unprecedented growth with a small team
Today, Google Kubernetes Engine is the foundation of Ki's most important asset: its algorithms, which use all available data to price risk and generate quotes instantly for users on the Ki Platform. The solution is connected to a data lake built with BigQuery, which feeds the algorithm with data and powers commercial decisions at Ki.
"With our BigQuery-powered data lake, we get incredible insights into our own portfolio, enabling a better understanding of our portfolio on a real-time basis," says Hodgson. "It also enables our Portfolio Underwriting team to really target the best quality risks and make true data-driven decisions on how to balance the portfolio. Having those data-driven insights is how Google Cloud really gives us an edge as we scale."
"We're the largest first-time insurance business to enter Lloyd's of London in its 300-year history. Ki is experiencing amazing growth, which is accelerated by the Google Cloud-built technology platform. More traditional insurance businesses would really struggle to achieve similar growth."
—Richard Hodgson, Chief Technology Officer, Ki InsuranceThis edge is evident in the numbers: "We're the largest first-time insurance business to enter Lloyd's of London in its 300-year history," says Hodgson. "Ki is experiencing amazing growth, which is accelerated by the Google Cloud-built technology platform. More traditional insurance businesses would really struggle to achieve similar growth."
Scaling faster with 90% more cost effective model training
Ki's growth depends on the scope and quality of its algorithms. That's why the company has continuously iterated the original algorithm and its individual components, enabling the company to expand to new classes of businesses. However, retraining the model costs money, and Ki's algorithm was originally delivered in a more expensive machine learning environment.
"By having a Google Cloud-based setup, we managed to lower the price of retraining the algorithm by about 90%," says Hodgson. "This enables us to do tens of thousands of model trainings and retraining in a cost-effective way. Our ability to train and iterate models with Google Cloud powers the growth of Ki and unlocks the scalability we need."
Reinventing insurance for the digital future
Looking ahead, Ki's algorithms are only going to get better and more sophisticated. With algorithmic underwriting, Ki is helping open a new chapter for Lloyd's of London, embracing the broker experience for the digital age. The team is now working with Google Cloud to further develop Ki's data platform, seeking to harness additional datasets to better understand risk and improve performance.
"There's enormous potential for change in the insurance market, especially a market such as Lloyd's of London that's still undergoing its digital transformation. With support from Google Cloud, we've already established Ki as an example of what can be achieved and we're excited about driving digitization in the industry together."
—James Birch, Development Director, Ki Insurance"There's enormous potential for change in the insurance market, especially a market such as Lloyd's of London that's still undergoing its digital transformation," says Birch. "With support from Google Cloud, we've already established Ki as an example of what can be achieved and we're excited about driving digitization in the industry together."
Tell us your challenge. We're here to help.
Contact us
About Ki Insurance
Ki was launched with the aim of helping redefine the commercial insurance market for the digital era. As the first fully digital and algorithmically-driven Lloyd's of London syndicate, Ki provides a completely different experience for the Lloyd's of London broker, enabling them to obtain a quote in seconds rather than days.