Having entered the business and technological mainstream two years ago, Generative AI is now helping to boost operations and cost efficiency across the insurance landscape. Among GenAI’s cutting-edge developments is the introduction of cognitive AI assistants that can learn, adapt, and collaborate to execute knowledge-based insurance tasks and processes independently.

Autonomous AI assistants are role-based digital coworkers. They can be trained and self-correct their output to improve their performance continuously. According to BCG, experts estimate that autonomous AI assistants will be mainstream within three to five years.

In commercial underwriting, AI assistants can streamline core processes and boost the efficiency and accuracy of underwriting teams. They can ingest underwriting guidelines and diverse data inputs to manage tasks such as appetite determination and risk analysis. AI assistants can make inferences and decisions, such as flagging submissions with contradictory information or applications that do not meet appetite eligibility standards.

They operate as well-prepared and consistent knowledge workers. In many ways, the contributions of role-based AI assistants in boosting underwriting proficiency are limitless, as they can always learn and advance their capabilities.

Addressing Underwriting Challenges with AI Assistants

Insurers and MGAs face various challenges as they navigate changing market conditions and customer expectations. For example, manually reviewing commercial applications to determine whether risks are in or out of appetite is time-consuming but also a strategic imperative. AI assistants can quickly ingest underwriting guidelines and determine if a risk falls within appetite. Then, underwriters can manually review just the risks that fall into gray areas.

To adjust appetite parameters, the assistant can quickly learn new guidelines and apply them to incoming submissions.

Insurance organizations can leverage AI assistants to help bridge the talent gap, as experienced insurance professionals can be hard to find and recruit. For underwriting, the hiring challenge can be acute, considering the specialized skills and time it takes to manage and master the art and science of risk evaluation. AI assistants can be seamlessly integrated and taught to undertake a range of risk assessment tasks, freeing underwriting teams to focus on the more complex functions and decisions. The underwriters, not IT staff, can train these digital co-workers to absorb nuances and complete tasks as per standard procedures.

AI assistants can help improve communications and information collection between underwriters, agents, and brokers, enabling policyholders to get more complete and precisely priced policies faster. For instance, they can instantaneously flag missing information in complex commercial applications, so the agent or broker can obtain details from the business before the submission goes to the underwriter.

Insurers and MGAs can also set up AI assistants to scan and organize structured and unstructured data sources in real time to gather risk-quality insights about businesses, considerably reducing work by agents or brokers to collect policyholder exposure details.

Deploying AI Assistants in Commercial Insurance

There are some practical ways insurers and MGAs can use AI assistants to automate and expedite commercial underwriting workflows.

• Insurers often rely on agents to answer risk questions, increasing potential subjectivity and dependence on the insured’s understanding or responses about exposures. This can lead to inaccurate risk assessment and potentially insufficient coverage. AI assistants can consistently identify and capture risk signals associated with specific business classes and support a more accurate risk evaluation process.
• When a new business—such as a daycare center, cannabis shop, or sports bar—moves into a policyholder’s building or next to a policyholder’s business, the insurer or MGA may not know all the exposures associated with new tenant occupancy. A risk appetite AI assistant with access to Lessor’s Risk Only (LRO) exposure data can provide real-time property and occupancy insights. This capability helps underwriters determine the full impact of tenant exposure on their policyholders and enables more refined decisions about an overall property/LRO portfolio.
• Underwriters can utilize AI assistants to help with workers compensation premium audits. These annual audits require many manual steps to ensure premiums are accurate and reflect the risk exposures of individual businesses. An AI assistant can support automated data collection and classification, resulting in faster, more efficient audits. It can accelerate field or phone audits by addressing incomplete information and refining business classifications. Furthermore, the AI assistant can incorporate state-specific regulations, ensuring full compliance and continuous monitoring for ongoing policy risk assessment.

Integrating AI Assistants into Existing Underwriting Workflows

Insurers and MGAs can best start integrating digital co-workers into their underwriting processes by recognizing them as knowledgeable and capable interns. Within their personas and roles, the AI assistants can be trained and supervised by the underwriting team, and their training can be tweaked as needed to complete assigned tasks reliably.

As an example, the AI assistant can be assigned to help with portfolio management. Typically, underwriters would perform this ongoing task manually, looking at a select sample of accounts to ensure they are within appetite. The AI assistant can be trained to do the same. However, with their processing speed, an AI assistant can swiftly review the entire portfolio instead of reviewing a select sample.

As the AI assistant digests underwriting guidelines, task-oriented nuances can be expanded in the workflow. For example, an insurer or MGA may want to consider restaurant policies only when alcohol sales are less than 30 percent of revenue. On an application, a restaurant might indicate 25 percent of revenue from alcohol sales, but it is open until 2 a.m. For an underwriter, this could raise a flag to investigate further, as alcohol sales may seem underreported, given the hours of operation. The underwriter can then teach a specialized risk appetite AI assistant to flag similar discrepancies among exposure factors.

Digital co-workers can complete distinct processes within commercial underwriting workflows, enabling underwriters to focus on more strategic responsibilities. The partnership between underwriters and their autonomous AI assistants is increasing workforce capacity, strengthening operations, and helping to grow books of business.