How AI is reshaping the insurance workforce

Safari: AI is empowering people across the insurance industry to rethink how work gets done. In underwriting, we’re seeing a shift from manual risk evaluation to what we call risk orchestration.

AI agents can rapidly process structured and unstructured data across thousands of data points, from health records to driving behavior and making preliminary assessments. The role of the underwriter will evolve to validate AI generated recommendations, orchestrate complex risk and handle nuanced cases.

AI agents can also handle first notice of loss, fraud checks and data extraction from documents and images. That changes the role of a claims adjuster from repetitive task executor to a claims strategist. They’ll increasingly be responsible for exceptions, customer experience and being there in the moments that matter most. AI tools simplify back-office operations by intelligently routing requests and streamlining management, which accelerates claim resolution and reduces processing errors.

Customer service, too, is being reimagined. Contact centers are moving toward hybrid models where multilingual, always-on AI agents handle routine interactions, and humans step in when empathy or creative problem-solving is needed. The goal isn’t just faster service; it’s smarter, more responsive engagement facilitated by these AI-augmented advisors.

PropertyCasualty360: What are some practical tactics for introducing AI into existing workflows, and where is human expertise still required?

Safari: Introducing AI into an organization starts with reimagining the process end-to-end. For example, instead of dropping AI into one step of underwriting, you ask: “How do we reimagine this workflow entirely with AI and human collaboration in mind?”

That shift in mindset is critical. Practical tactics include reimagining entire workflows through an AI-first lens — rethinking how outcomes can be achieved in faster, more tailored and intelligent ways. Rather than automating siloed tasks, the focus should be on end-to-end transformation where AI is embedded at the core of the process, not just layered on.

This can include using confidence codes to communicate the reliability of AI-generated data, helping users know when to trust the output and when to escalate. Additionally, establishing third-party risk protocols for vetting AI tools from vendors and creating internal champions — people who experience workflow issues firsthand and are motivated to test and promote better AI-enabled approaches.

It is important to also delineate between tasks that can be eliminated or transformed with the use of AI versus ones that still need human oversight. Human input is essential to guiding AI development in compliance-heavy environments and determining where AI adds the most value, then adapting workflows accordingly.

Humans excel at creativity and innovation, while also bringing the empathy and interpersonal skills needed for nuanced decision-making, personalized customer experience and strategic planning. AI can generate a summary or prediction, but it takes a trained underwriter, claims analyst or service rep to contextualize that insight and make important decisions, especially in complex or emotionally sensitive situations.

PropertyCasualty360: Can organizational models help teams adopt and scale AI effectively?

Safari: Absolutely. The right organizational model is essential for scaling AI in a way that delivers long-term business value. For instance, pairing a centralized AI Center of Excellence to oversee governance, data quality and ethical implementation with embedded AI champions in business units who drive adoption from the ground up. This structure allows for consistency in approach while empowering teams to tailor solutions to their needs.

Equally important is investing in strategic platforms and partnerships. Organizations that rely solely on single-point solutions risk fragmentation and limited scalability. Prioritize platform-based models that enable AI agents to operate across systems, share context and collaborate with human teams. PwC’s own Agent OS approach is one example, which supports cross-functional orchestration between AI and people, making enterprise-wide transformation more achievable.

At its core, adopting AI is a business strategy, not just a technology decision. The most effective organizational models recognize that and are designed to support innovation, accountability and collaboration at every level.

PropertyCasualty360: Are there working strategies for upskilling the workforce and closing the AI fluency gap?

Safari: Yes, and they need to be embedded in a company’s broader business and workforce strategy, not treated as one-off training initiatives. In insurance, effective upskilling starts with educating people at every level.

We recommend taking a persona-based approach to upskilling the organization, tailoring training to each persona’s individual needs. That means giving C-suite leaders, managers, AI builders and AI users a clear understanding of what AI can and can’t do. Key focus areas include prompt engineering, data fluency, validation techniques and responsible AI usage.

Organizations need to use workforce data to assess where skills gaps exist and where the business is most exposed to disruption. That kind of data-driven approach helps support internal mobility and guide investment in the right reskilling pathways. The AI fluency gap won’t close on its own. It takes intention, structure and leadership to move the needle.

PropertyCasualty360: What human roles will be reduced or removed by AI?

Safari: AI will undoubtedly shift the nature of many roles, especially those centered on repetitive, rules-based tasks such as data entry, form processing and first-line claims triage. These functions are ripe for automation and augmentation.

However, AI is not replacing people; it’s redefining what they do. In insurance, we’re seeing roles evolve, freeing up human capacity to focus on higher-value work such as complex decision-making, customer empathy, risk judgment and innovation. What’s changing is the task mix, not the importance of the role. AI is absorbing repetitive activities and freeing people up to focus on what humans do best: applying judgment, building relationships and navigating ambiguity.

The opportunity for insurers is to redesign roles with this partnership in mind. Invest in upskilling so that people can work confidently alongside AI and take on higher-value, more fulfilling work. As someone who’s led large-scale transformations across the insurance sector, I’ve seen that when organizations approach AI as a co-pilot, rather than a replacement, they unlock both business value and human potential.

PropertyCasualty360: What human roles will remain? Why?

Safari: Human roles aren’t going away, they’re becoming even more critical. In insurance, roles that rely on empathy, interpersonal skills and nuanced decision-making will continue to be essential. Think about claims adjusters navigating sensitive situations after a loss, or agents helping customers make coverage decisions that affect their families or businesses. These interactions require trust, emotional intelligence and context, which are things AI simply can’t replicate.

Even as AI agents become more capable, they still need human oversight and expertise. Roles in compliance, risk and governance will remain vital to defining how AI is used responsibly and in line with regulatory expectations. And strategic roles, such as those focused on innovation, customer experience and business model evolution, will grow in importance as companies rethink how they deliver value.

Ultimately, it’s not about human versus machine. It’s about designing the right collaboration between AI and people, where AI handles the repetitive and analytical tasks, while humans bring the judgment, empathy and creativity that drive meaningful outcomes.

PropertyCasualty360: Describe an insurance company of the future.

Safari: The insurance company of the future is digitally orchestrated and human-led. AI agents will operate across the enterprise, connecting systems, data and decisions in real time. Underwriting decisions that once took days will happen in minutes. Claims will be adjudicated proactively with minimal customer intervention. And, customer engagement will be personalized, contextual and seamless.

But technology is only half the story. These companies of the future will be built on trust and transparency, with clear governance over AI use, ethical frameworks in place and the ability to explain decisions to customers and regulators alike.

Employees will no longer spend time on rote tasks. Instead, they’ll be strategic advisors, risk specialists and insight navigators. They’ll work alongside AI co-pilots. This will unlock new value, enable faster innovation and ultimately deliver a more resilient, personalized and responsive insurance experience for customers.

See also:

• Most and least expensive states for business insurance
• Can tort reform control insurance costs in the U.S.?
• Insurance customers aren’t seeing the benefits of AI