Michael Krikheli
Co-Founder & CTO of Five Sigma
Michael leads the development of advanced claims management products powered by data, analytics, and AI. With over a decade of experience in software engineering and team leadership, Michael excels in designing, building, and scaling innovative technologies that solve complex, real-world challenges.
15 MINUTE READ
Executive Summary
Claims management stands at a pivotal juncture, with AI actively reshaping how insurers handle claims across the entire lifecycle. This article explores the transformative potential of AI in the insurance industry, envisioning a future where highly intelligent machines lead the claims process.
The article delves into how AI can function as a superhuman claims professional, combining deep industry expertise with unparalleled efficiency. It outlines the potential for AI to analyze policies, assess liability, detect fraud, and orchestrate every stage of the claims lifecycle with precision.
The article examines the implications of this AI revolution, addressing crucial questions about the future roles of adjusters, supervisors, and chief claims officers in the insurance industry. It also confronts the challenges and risks associated with AI integration, including the need for explainable decisions, unbiased assessments, and seamless integration with existing operations. By breaking down the key processes in claims management, from FNOL and triage to litigation management and compliance, the article illustrates how AI agents can enhance each step, potentially transforming the entire claims ecosystem.
The author argues that while AI will lead claims processing, human oversight remains crucial for complex decision-making and ensuring fair and accurate claim resolutions. The article ultimately paints a picture of a future where AI and human expertise combine to create a claims experience that is faster, smarter, and more customer-centric than ever before.
Introduction
Reading the piece of beauty that is the essay by Dario Amodei (Co-Founder and CEO of Anthropic) – Machines of loving grace, has sparked a thought: What would an AI revolution in claims management look like? More importantly, why should we pursue it, and what happens if we get it absolutely right? This essay explores these questions, imagining an industry where AI reshapes the claims process for the better.
The Importance of Insurance Claims
There’s a saying that no one dreams of becoming a claims handler—you just find your way into it. The same can be said for claims system innovators. But insurance itself is a fundamental pillar of modern society—it’s what enables people to take risks. Without it, most of us wouldn’t buy a home, drive a car, or start a business.
Claims management is at the heart of insurance. It’s about getting people their lives back when things go wrong. Every claim represents a disruption—whether it’s the hassle of a fender bender disrupting your day or a severe injury with uncertain medical expenses.
The role of claims professionals is to guide people through these moments of uncertainty and restore their stability. That’s what makes this work important. It’s a noble profession—one worth optimizing, improving, and investing in.
AI Improves Claims Handling with Superhuman Precision
What if AI could function as a superhuman claims professional, seamlessly combining deep industry expertise with automation? A system that understands regulations and best practices at a human level while analyzing individual claims with superhuman precision—assessing coverage, liability, and necessary actions with speed and accuracy. Wouldn’t that be something?
Beyond single claims, AI can detect patterns across cohorts, identifying fraud risks, refining reserving strategies, and optimizing customer service. It doesn’t just process data—it actively drives claims forward, generating documents, sending emails, and even providing status updates or making calls, ensuring a seamless and highly efficient claims experience.
AI Redefines Claims Professionals Roles
If AI can handle claims with superhuman efficiency, it raises fundamental questions about the industry’s future:
- What will be the role of adjusters, supervisors and chief claims officers?
- What will be the role of the system of records (Claims Management Systems)?
- How will AI impact customer experience—enhancing speed and transparency or diminishing the human touch?
- How will AI impact claim cycle times?
- How will AI reshape the structure and role of insurance claims organizations?
Can an AI-led Claims Management Reality Exist?
Short answer: Yes.
Long answer –Let’s break down the key processes a claims organization follows to handle a claim and explore how AI can enhance each step.
- FNOL & Triage – The First Notice of Loss (FNOL) is the first step in the claims process, usually arriving through phone calls, emails, or digital platforms. AI can extract and interpret unstructured data, triage the claim, and identify fraud indicators. AI products like Five Sigma’s Clive™, the AI Claims Adjuster, already ingest FNOLs, while speech-to-text technologies enable conversational FNOL experiences. AI streamlines this process by ensuring immediate next steps are triggered.
- Coverage Verification – AI analyzes policy documents, claim data, and past interactions to determine coverage eligibility. By cross-referencing policy terms with FNOL data, AI can flag exclusions, calculate exposure, and instantly assess whether the claim is valid. This automation ensures faster decisions and minimizes manual interpretation errors.
- Information Gathering – AI agents proactively reach out to claimants, third parties, and repair shops for missing details, streamlining documentation retrieval. Whether it’s medical records, police reports, or damage photos, AI agents identify gaps and trigger automated follow-ups, ensuring claims advance without unnecessary delays.
- Damage Analysis – At the core of claims automation, damage analysis engines are being developed to assess losses instantly. These systems combine historical and real-time data—such as car part prices or medical treatment costs in specific locations—with AI-driven logic that matches claim information (e.g., damage photos) to real-world pricing. Many great companies combine both raw data with decisioning logic on top of it to optimize the analysis (CCC is a notable example). Today’s technology enables AI to interpret images, assess damage, and proactively follow up with customers when additional information is needed.
- Repairs Coordination – Managing repairs involves coordinating multiple stakeholders, reviewing estimates, and ensuring cost efficiency. AI agents monitor timelines, flag excessive charges, and detect repair delays. They also assist in negotiations, making sure that repairs align with fair market pricing and avoiding unnecessary costs. Moreover, AI agents ensure the process keeps moving forward, preventing delays caused by bottlenecks or waiting on human intervention, reducing costs across the entire lifecycle.
- Reserving & Payments – Reserving is about predicting a claim’s financial outcome by analyzing historical trends, industry benchmarks, and real-time claim details. AI can assess past claims—both within a company and industry-wide—to determine appropriate reserves, continuously updating them as new information emerges. On the payment side, AI agents can review and match bills to coverage, apply deductibles, co-insurance and more, ensuring correct payouts while flagging discrepancies. With human oversight, adjusters can validate these decisions for accuracy and compliance before gradually shifting toward automated straight-through processing where appropriate, ensuring a balance between efficiency and control.
- Recoveries & Liability Assessment – AI identifies subrogation opportunities by analyzing liability factors and prior claims data. It assesses recorded statements, police reports, and witness accounts to determine fault allocation, ensuring that insurers recover appropriate costs from responsible parties.
- Litigation Management – AI aids in legal case preparation by analyzing demand packages, drafting responses, and summarizing case details for legal teams. By structuring claim data effectively, AI ensures attorneys have clear, well-organized information for negotiations, litigation and settlements.
- Compliance & QA – AI compliance checks prevent regulatory breaches by embedding Standard Operating Procedures (SOPs) and legal requirements into claim workflows. AI ensures decisions follow both internal policies and external regulations, reducing compliance risks. Additionally, AI-based QA continuously audits claims, identifying inconsistencies before they escalate.
- Claim Triage & Assignments – This step determines the type and complexity of a claim, using all available data to classify and direct it appropriately
This list, while detailed, is not exhaustive. But it does capture some of the nuances and intricacies of the process.
An Illustration of AI Agents Architecture in Claims Management
The Process of Claim Handling by AI Agents
A behind-the-scenes look at how AI agents handle claims—where every step is analyzed, assigned, and executed with precision by the claim Orchestrator.
The Incident
An insured is involved in a multi-party collision, sustaining severe leg injuries and requiring emergency care.
FNOL & Initial Assessment
The insured reports the claim to the FNOL agent, who gathers the details, logs the loss, and escalates the case to the Orchestrator.
Triage & Categorization
The Orchestrator assigns the claim to a Triage Agent, which evaluates its complexity and classifies it accordingly. In this case, it recognizes a multi-party collision requiring a structured handling plan.
Claim Planning
The Planning Agent generates a step-by-step workflow, outlining tasks such as:
- Reviewing coverage and policy details.
- Requesting medical records from the ER.
- Assessing liability and gathering recorded statements.
- Estimating vehicle damage and repair costs.
- Ensuring all injured parties are accounted for.
- Updating reserves as new information emerges.
- Managing treatment plans and monitoring recovery progress.
- Handling rental car coverage, if applicable.
- Processing payouts for each exposure before finalizing the claim.
Plan Execution
The Orchestrator assigns tasks to specialized AI agents:
- Coverage Analysis Agent – Reviews policy terms, exclusions, past payments, and loss dates, confirming applicable coverages such as Collision, PD, PIP, and BI.
- Financials Agent – Establishes initial reserves based on the severity of the claim and updates them dynamically as the case progresses.
- Information Gathering Agent – Requests supporting documents such as medical records, police reports, photos, and videos from claimants and third parties. If a police report exists, the agent retrieves it via API integrations (e.g., LexisNexis).
- Liability Assessment Agent – Collects recorded statements from all parties and cross-references them for inconsistencies. It synthesizes a detailed liability analysis and sends it for human adjuster review.
- Damage Analysis Agent – Deploys multiple instances to assess each damaged vehicle and injured individual, determining next steps. The agent may wait for shop estimates, request a virtual assessment, or dispatch an appraiser for on-site inspection.
- Compliance Checker Agent – Runs continuously to validate that all decisions and actions align with policies, regulations, and legal requirements. It flags any non-compliant actions for human review.
This example illustrates how AI agents collaborate under the direction of an Orchestrator, each playing a distinct role. They analyze real-world events, make decisions, communicate with stakeholders, and integrate with third-party data sources. While AI leads claims processing, human oversight remains crucial for complex decision-making, ensuring fair and accurate claim resolutions.
The Role of Humans in an AI-led Claims Operation
What role do humans play in a world where AI leads claims handling? Will their role change?
To answer this, we must answer a more fundamental question: what is the purpose of claims management?
In my view, it has always centered around these key pillars:
- Helping the customer – Whether it’s a minor scratch on a car or a life-altering injury, no one wakes up planning to file a claim. When they do, they are looking for support to get back on track. This is the promise behind every insurance policy, and the core mission of claims handling: to restore normalcy for the policyholder.
- Making financial and legal decisions – A claim is not just about payouts; it involves assessing the financial and legal aspects of the policy. Understanding what happened, how it aligns with policy terms, and ensuring the right financial and legal decisions are made is a crucial part of the process.
- Identifying bad actors – The unfortunate reality is that fraud exists. Claims organizations must distinguish legitimate claims from fraudulent ones, ensuring that resources go to those who genuinely need them while stopping exploitation of the system.
With these goals in mind, and with powerful AI claims agents at our disposal, where do humans fit into this new landscape? Here are some key roles they continue to play:
- Supervision & Governance – Adjusters, supervisors, and chief claims officers must oversee AI processes, refine AI-generated plans, and ensure regulatory compliance at multiple levels; from individual claim reviews to broader operational consistency and system guardrails.
- Empathy & Human Touch – Claims are often stressful and emotional events for policyholders. While AI can streamline processes and reduce friction, human intervention remains critical in providing reassurance, building trust, and offering personalized support when needed.
- Catching Nuance & Complexity – AI excels at automation and pattern recognition, but edge cases and subtle judgment calls still require human expertise. Experienced adjusters and supervisors play a vital role in handling exceptional cases, and ensuring claims decisions account for real-world complexities AI might overlook.
The future of claims isn’t about replacing humans—it’s about elevating them. AI and automation handle repetitive, data-intensive tasks, while human expertise focuses on decision-making, customer care, and strategic oversight. This partnership creates a claims experience that is faster, smarter, and more customer-centric than ever before.
The Role of the System of Record (CMS)
We’ve outlined the roles of AI agents and humans—now, what about the claims management system that has always been there? How does it fit into an AI claims environment?
- Communication Hub – The system of record serves as the bridge between AI agents and humans. AI logs actions and insights, while humans review, provide feedback, and intervene when needed. It also connects with external systems—policy admin, billing, police reports, damage assessment tools, and regulatory bodies—ensuring seamless data exchange.
- Phased AI Deployment – Claims systems are where human adjusters work today. By integrating AI agents into the system of record, insurers can introduce automation step by step, extracting value from each process and each phase before rolling out the next.
- Process Mapping & Structure – While claims handling is often messy and unstructured, the system of record provides a structured framework, ensuring no critical steps are missed. It translates complex SOPs and regulatory requirements into actionable workflows for both AI and human adjusters.
With AI seamlessly integrated, the system of record evolves into more than just a data repository—it becomes a convenient, automated workspace. AI agents handle data extraction, documentation, and insights generation, freeing adjusters from manual work and enabling faster and better claims decisions.
Claims Handling Problems AI Can Solve
Insurance companies and claims teams face mounting challenges. While AI agents won’t solve everything, they can help alleviate some of the industry’s biggest pain points:
- Knowledge Gap – Experienced adjusters are retiring, and there aren’t enough skilled professionals to replace them. This loss of institutional knowledge impacts settlement quality, customer experience, and overall efficiency.
- Lack of New Talent – Fewer young professionals are entering claims roles, making it even harder to close the knowledge gap and staff adjusting teams adequately. New talent and college graduates seek modern work environments where they can use advanced tools and contribute meaningful value.
- Regulatory Complexity – Ever-changing compliance requirements create a major burden on claims teams, requiring constant updates and adjustments to workflows.
- Rising Claims Costs – The cost of claims is increasing globally due to factors like climate change, supply chain disruptions, and inflation. This leads to higher loss ratios, market exits, and affordability issues for policyholders.
AI agents can help mitigate these issues in key ways. They retain and apply institutional knowledge at scale—functioning like expert adjusters who never tire or leave. Their presence also makes claims handling more attractive to new talent, shifting the role from data entry to AI-based customer service. AI ensures compliance by dynamically adjusting workflows as regulations evolve. And by reducing leakage and errors, claims automation allows insurers to operate more efficiently, scaling their business with smaller, more agile teams.
The Risks – Where AI in Claims Can Go Wrong & How to Ensure Success
While the picture painted above looks great and rosy, the new technologies do not come free of risks. Claims management impacts both human lives and major financial decisions, making it critical to identify and mitigate potential pitfalls before deploying AI agents at scale.
Key Risks of Deploying AI in Claims Management:
- Bias – AI must remain objective. Claims decisions cannot favor certain groups over others. Bias can stem from training data, model alignment, and real-world inputs, requiring continuous monitoring and correction.
- Incorrect AI Decisions (Hallucinations) – AI can sound confident while being totally wrong. If errors cascade across automated decisions, the impact could be severe. AI must involve human oversight where uncertainty exists.
- Transparency – AI decisions must be explainable to adjusters, policyholders, and regulators. AI must justify recommendations with clear reasoning and evidence. It’s not enough to just blindly go along with the suggestions.
- Privacy & Data Security – Protecting customer data has always been critical in claims management. AI must never expose customer PII. Claims data must be masked, protected, and never used in AI model training to ensure compliance and maintain trust.
How to Mitigate These Risks:
- Recognize & Address Risks Early – Companies must assess risks in AI-driven claims solutions and implement safeguards before full deployment.
- Implement AI Guardrails – AI agents should detect and correct each other’s errors, while strict code-based constraints prevent unintended outcomes.
- Enhance Human Oversight – The system of record must provide a UX that allows adjusters to monitor and intervene in AI-driven decisions when needed.
- Improve AI Models & Training – AI models are improving rapidly. Fine-tuning them for specific claims use cases will reduce risks over time.
- Secure Customer Data – AI must never store or process sensitive data improperly. Strong anonymization and compliance frameworks are non-negotiable.
Conclusion
The companies that act now—embracing AI while managing its risks responsibly—will define the next era of claims management, setting the new standard for efficiency, accuracy, and customer experience. As the industry evolves, those who hesitate risk falling behind, constrained by outdated processes and inefficiencies. AI is no longer a distant concept—it’s already here, actively reshaping claims handling from FNOL to settlement.
Adoption must be strategic. AI needs to be explainable, compliant, and secure, integrating seamlessly into existing claims operations rather than disrupting them. The goal isn’t just automation—it’s intelligent augmentation, where AI streamlines workflows and enhances human expertise, ensuring claims are handled faster, more accurately, and more fairly than ever before.
The opportunity is real. The technology is here. The transformation is already underway.
It’s time to build.
