Chisel AI is a software firm that uses artificial intelligence (AI) to improve sales, processes and efficiency for commercial insurance brokers and carriers.
The technology behind it, which founder Ron Glozman dreamed up in a dorm room at the University of Waterloo in Ontario, Canada, extracts, interprets and analyzes data in policies and other digital documents hundreds of times faster than a human. This, says Glozman, is how AI is transforming commercial insurance.
Q
How has artificial intelligence changed in recent years? What trends are you seeing?
A
AI is being implemented across the insurance value chain to solve problems that have plagued the industry for decades. It’s giving brokers new capabilities to automate cumbersome manual processes and remove friction from the commercial insurance buying process.
We’re seeing more production deployments with plans underway to scale these initiatives. It’s still early days, but I’d say we’re entering the early majority stage of AI adoption. There’s growing industry maturity and confidence in AI. Brokers and carriers are more AI literate, and innovation is less siloed. Business and operations leaders are now heavily involved in, if not driving, AI projects with well defined business cases. I think we’re at an inflection point in the evolution of AI where, if we fast-forward 10 years, those brokers and carriers that didn’t make strategic investments in AI early on will be struggling to compete and stay relevant—if they are still in business at all.
Q
How can brokers harness the power of AI to deliver a better policyholder experience and differentiate in today’s hard market?
A
Being a trusted advisor is at the heart of insurance brokering. Today, with policyholders needing more guidance on new coverages like cyber, business interruption, and renewals where premium rates have increased, brokers really need to focus 100% of their time on their customers’ needs and placing business faster. One way to achieve this is to spend less time on manual, paper-laden tasks that can be automated and streamlined, like checking policies for errors and omissions.
Let’s face it: manually reviewing lengthy commercial insurance policies prior to issuance is a thankless task. Policyholders expect high-quality, error-free policies, but errors and omissions—like missing endorsements, incorrect limits, and premium shortfalls—are all too common. As with any mind-numbing repetitive task, there is a high probability of human error. After page 55, tired eyes prevail, and it’s easy to miss things. By using AI, brokers can automate and standardize the policy-checking process to reduce turnaround times, mitigate E&O risk, free up staff to focus on nurturing customer relationships, and place business faster.
Q
What is natural language processing, and how does it differ from RPA and regular expression solutions?
A
Natural language processing [NLP] is a fancy term for teaching computers to read. What makes NLP solutions so powerful is that they can be trained to read and understand insurance-specific language and jargon just like a human, only hundreds of times faster. When a well-trained NLP solution reads an insurance document, it can recognize and contextually understand data points such as policy number, producer code, named insured, street address, policy term, limits, premiums, deductibles and more. It can also make sense of key relationships, such as who is the underwriter, who is the broker, who is the insured, what is the premium, and what is the limit?
Unlike pattern-matching tools like robotic process automation [RPA] and regular expression [Regex], only NLP continues to learn and discover through data. The more data you feed the model, the more accurate it gets over time. Each time a human knowledge worker corrects a mistake or confirms a prediction, this information is recorded and used in the next training cycle. This is often referred to as “a human in the loop” and is part of the AI continuous learning cycle that ensures the system gets smarter and more accurate as it goes along. NLP systems can capture feedback, learn, iterate and improve, whereas RPA and Regex are business-rules driven. Over time, NLP systems get better, while RPA and Regex systems suffer from increased maintenance and support issues as rules become stale.
Q
What are we talking about in terms of time saved in a given day for the average insurance professional?
A
In a recent poll, brokers indicated that it can take up to seven hours to manually check a policy for errors and omissions prior to issuance. More complex policies can take days to manually review and validate. With an AI-powered policy-checking solution, brokers can check policies in minutes, digitally comparing documents such as a new policy, quote, binder and existing policy to identify errors and omissions, such as missing endorsements, incorrect limits, address errors, or premium shortfalls. With AI, brokerage staff can free up hours each day to focus on what they do best—providing customer advice and growing their book of business.
Q
How will brokers know if their organization is ready for AI?
A
Start with the business problem. Define the business case first and include business and operations leaders in all the discussions. I like to think of AI as one tool in the toolbox. Based on the problem and scale, a different solution may be required. Before engaging AI vendors, insurers need to identify the best alignment with their overall corporate and business objectives first. Also, define the success criteria so it’s clear how you will evaluate, select and roll out an AI solution. I’ve seen projects fail because there were no measurable milestones defined.
Q
Can you address some of the common misconceptions around AI?
A
The most damaging misconception is that AI is coming to replace insurance workers. The reality is that most effective AI and machine-learning processes today are supervised by humans. While AI can help enormously by automating many workflows and mind-numbing tasks, one thing it will never replace is the personal touch. Insurance is a human business. We will always need people exercising their knowledge, expertise and judgment. A successful AI-powered strategy integrates tried and true customer relationship-building strategies with new data-driven digital experiences.
Q
How can AI play a role in future business disruptions like what we’ve just been through with COVID-19? Can you give us a hypothetical case study?
A
One lesson we’ve learned from the pandemic is that automation and resiliency are intertwined. Throughout the COVID-19 crisis, insufficient digital enablement and a lack of workflow automation created capacity shortages and operational issues. Remote work also exposed long-standing issues with manual, paper-based processes. Striking the right balance between digital and human capacity will give brokers the flexibility to continue to do business during difficult times with minimal to no impact.
Another lesson is that digital capabilities are habit forming. Customer expectations around the services they receive from their broker have been radically reset. The pandemic has brought us to the tipping point of digital transformation, and brokers can no longer afford to sit on the sidelines.
