Expert shares practical tips for human-friendly AI adoption, upskilling

As artificial intelligence remains a hot-button topic in insurance, sparking intense debate and ubiquitous uncertainty, Rima Safari, principal, PwC, is sharing insights on best practices for American insurers on how to adopt and scale AI while incorporating existing human expertise.

“Given the pace at which this technology is evolving, it’s better to start with at least some sort of centralized structure like an AI innovation hub or a center of excellence that can stay on top of latest technology developments, share best practices, drive reusability of common patterns in terms of the use cases that are being built out, manage vendor integrations and help track ROI,” Safari said in an interview with InsuranceNewsNet.

From there, she recommended companies:

• Avoid leveraging AI for narrow use cases
• Ensure everyday users are engaged early on
• Focus on trust and governance when introducing AI into existing workflows.

Safari noted that human expertise is still invaluable in insurance underwriting, claims and customer service. However, when it comes to upskilling the workforce, she recommended insurers adopt AI across all levels of the organization and handle change management with care.

Tips for adopting AI

According to Safari, one of the biggest challenges to introducing AI is leveraging it for “very narrow or isolated use cases.” She explained that while it can be helpful and drive some productivity, it ultimately does not drive significant savings because of the limited extent of usage.

“It’s really important to use AI to truly reimagine processes end-to-end across the value chain, and that’s where we’ve seen significant productivity impact. We’ve seen 30% or more operational efficiency in that space as well,” Safari said.

Additionally, she emphasized that human oversight is the most critical part of AI adoption. As such, it’s key to engage business and operational users early on and not just treat AI as a technology-only solution.

“So, as an example from operations, say you’re building a contact center AI bot for insurance. It’s important to engage the contact center specialists that are feeling the pain today. If you engage them early on, then they are part of that design process and they are also validating the solution to make sure it meets their needs,” she said.

The final peg in the wheel boils down to building trust and establishing governance, which includes setting up all possible guardrails to account for risks like AI hallucination.

“Being able to show that there’s auditability across the process, being able to answer customer questions that come in on, ‘Would you be able to turn this off if I don’t want to use AI?’ or ‘How easy or difficult is it to do this? How was this decision made?’ and setting up the right guardrails around that has been key,” Safari said.

Human-AI collaboration

While human expertise is still needed in key areas within insurance, Safari noted that collaboration with the right AI tools can help drive productivity.

For example, she said AI can be “tremendously helpful” in underwriting, where underwriters currently spend “a ton of time” analyzing data to produce accurate, dynamic risk assessments. This is also the case for group insurance claims, which can take weeks to be processed because of the triage required. Similarly, AI agents can help with basic customer service requests, transcribing calls or drafting letters while insurance advisors tackle more complex cases.

“At the same time, you do want the humans to be verifying at each step of the process. So, for example, if a claim is going to be denied, AI can create the rationale for the denial based on all the information that was shared. Humans should still validate it, make sure there’s empathy in that denial letter, make sure you’re adjusting it to the nuances and personalizing it for the customer before sending it out,” Safari said.

This is where closing the AI fluency gap in insurance is crucial to ensure agents are well-equipped for a technology-driven workforce.

Closing the AI literacy gap

Safari suggested that AI be adopted straight across the board, from frontline staff to managers and C-suite executives, to most effectively upskill the entire workforce.

“We started at every level and we made sure that it wasn’t just our centralized teams, like our AI factory team, that’s focused on it — every single person, from the people in consulting to the operational managers to frontline users. That is what we want to see more and more with insurance carriers as well,” she said.

However, she acknowledged that change management is a crucial first step in this effort.

“You have to have a thorough change management process because this is really not about the tech; it’s about humans adopting it… The only thing that will be the biggest obstacle is if we don’t do the right change management to drive adoption from a broader team and community standpoint,” Safari said.

© Entire contents copyright 2025 by InsuranceNewsNet.com Inc. All rights reserved. No part of this article may be reprinted without the expressed written consent from InsuranceNewsNet.com.

Rayne Morgan is a journalist, copywriter, and editor with over 10 years' combined experience in digital content and print media. You can reach her at [email protected].