Oyster is a leading insurance platform for commerce that provides tailored business insurance for modern companies. We believe that AI presents the biggest opportunity ahead to drive innovation in our industry after serving hundreds of fast-growing companies and scaling insurance-specific data volume. As we accelerate the delivery of innovative products to our merchants, we are leveraging AI and LLMs to transform underwriting efficiency and deliver best-in-class customer service with quantifiable impact across our business.
Today, the way that businesses buy and manage insurance is completely different from how individuals do so. The legal language is more complicated. The risks are more heterogeneous and ever-changing. The entire process can take weeks of back-and-forth, with phone calls and emails, long documents, PDF forms, questions and answers, and more. As a result, getting effective commercial insurance coverage has been a relationship-driven “I know a guy who knows a guy” game, bottlenecked by human bandwidth.
However, with recent advances in generative AI and large language models (LLMs), this all stands to change. Over the last few months, Oyster has been reimagining the entire commercial insurance experience, from customer to carrier. This work helps Oyster live into our values of making insurance accessible for all and innovating and simplifying the process of obtaining and maintaining coverage.
Oyster is deploying the first AI commercial insurance agent – an AI teammate that helps our insurance experts move faster, more efficiently, and deliver the best commercial insurance experience ever.
Oyster AI is trained specifically to understand commercial insurance documents, coverages, and agency operations. It’s been taught to be particularly good at the specific verticals that Oyster primarily serves. Oyster AI is more than just automation – it can reason, react, seek input, and autonomously adjust its plan as it hits roadblocks. It results in a more streamlined, data-infused underwriting process with a human-in-the-loop approach to ensure agents have the final judgment.
Oyster AI is the teammate insurance agents always wished they had. Oyster agents can assign Oyster AI a ticket with a task – such as “Follow up with the customer to request their loss history” or “Generate and submit this customer’s application to the carrier” – and Oyster AI will respond to commandeer the ticket and post its plan, status, and log of actions as context for an agent to review.
Agents can talk to Oyster AI on Slack and ask questions about the current status of an insurance application or quickly summarize certain information about a customer’s policy. Agents spend less time reading, parsing, and switching contexts and more time delivering value to our customers.
Oyster AI never sleeps. It can react to events in real-time, such as new customer conversations or receiving a completed quote from the carrier via email, and ensure that someone is always available to help our customers.
Oyster AI is particularly adept at handling mundane insurance operations that take a lot of time and require linear reasoning. Currently, Oyster AI can autonomously handle about 30% of the tasks that agencies typically do (by time spent). That frees up agents, underwriters, and CSRs to focus on higher ROI tasks, ones that are best suited for humans. Our team is already working on Oyster AI v2, which aims to autonomously handle tasks that take up almost 80% of an agent’s time.
Here’s a sample of what Oyster AI is currently capable of doing:
When an application is submitted, Oyster AI can automatically read the submitted information and documents and determine whether any additional information is needed. Oyster AI can instantly surface this for an agent to act on, or automatically follow up with the customer requesting the missing information.
Most carriers that insure Oyster’s customers still operate via PDF forms and email, which are tedious and time consuming for humans to fill out. Oyster AI can be asked to reference a customer’s application and fill out these forms instead, intelligently matching the right information, and follow up with a customer or carrier with the completed forms.
Carriers often send quote documents in PDF form, and each one has its own bespoke format. When a quote is ready, Oyster AI reads the email and documents, prepares them to be signed by detecting the coverage information, premium, and signature and date fields, creates a financing program, and drafts an email to start the binding process with the customer.
Oyster AI can manage communications over multiple days and parties, facilitating information exchanges that typically cost agents a lot of time because of context switching.
Oyster AI understands each and every policy for Oyster’s customers. As a result, agents and customers can ask Oyster AI pointed questions about their policies, such as “how do I file a claim” or “when is my next payment,” and Oyster AI can instantly respond with the information or even draft up a response directly to the customer.
When a customer submits an application, Oyster AI automatically begins to evaluate it by collecting information from the website, the internet, and other third-party data sources. It organizes over 80 data points, provides an initial assessment of the risk, and presents the information for an underwriter to review. Oyster AI condenses research that would take a human hours down to minutes.
Oyster AI enables Oyster insurance agents to spend less time on repetitive operations and more time providing direct value to our customers and booking more premium. In fact, Oyster agents currently write about 3.5x more premium per agent than industry average.
We benchmarked the typical time it takes an agent to complete some common, well-defined tasks against 4 different scenarios:
As Oyster AI gains more capabilities, human agents can spend more time on work that better and more quickly serves our customers, and customers won’t be stuck waiting for days or weeks to have their insurance needs taken care of.
Premium efficiency gains improve at the start because we can automate the most time-consuming tasks, and again at the end because the agent moves more into the sales space, allowing us to first achieve time efficiency and then achieve scale efficiency.
This is just the beginning. Insurance is a massive data and service industry, making it prime for disruption by AI innovation. With Oyster’s all-in-one insurance platform, we believe that our team is uniquely positioned to deploy zero-latency AI systems to modernize high-fidelity underwriting and create faster and better customer services for businesses.
With a developed foundation of Oyster AI, we have an ambitious roadmap to integrate AI and LLM across our platform:
We’ve demonstrated Oyster AI’s efficacy based on an initial set of capabilities. The next step is going to be to continue to train the system and enable it to handle more complex agency tasks.
Oyster AI is currently being tested internally at Oyster. We’d like to make it generally available to our customers as well, to directly use it to manage their coverage, request changes and information, and make applying for insurance a breeze.
Customers will soon be able to use Oyster AI to proactively manage and mitigate risk using the brain that underlies its risk intelligence and underwriting capabilities.
Request a product demo with Oyster.
