When it comes to AI in the insurance industry, 2026 is shaping up to be a defining year. Adoption is accelerating, investment is increasing, and nearly two-thirds of independent agents say they're optimistic that AI can support their work.
But putting generative AI to work inside an independent agency, MGA, or carrier doesn't happen with the flip of a switch. Successful AI implementation takes an operational shift through disciplined execution, clear governance, and a defined path to measurable benefits. Most importantly, it requires investment in your most valuable resource: your people.
That was our guiding principle when we implemented enterprise AI tools at Vertafore. In 2025, we set out to equip ourselves and our teams to use AI in the best way possible. Our goal was to learn how to make ourselves, teams, and products better for the end customer.
To accomplish that, we committed to pausing business as usual for a full week as a company to give our global team time to immerse themselves in AI confidently, responsibly, and creatively. We made a deliberate decision to treat employee confidence as the primary KPI of our AI transformation, and we measured it before and after implementation.
Before expanding access, we concentrated on three priorities: mindset, training, and governance. Here's what happened and takeaways for businesses that want to find success with AI.
Employee confidence sets the foundation
Understandably, some employees fear that companies want to use AI to replace their human workforce. That may be the case in some organizations, but many businesses—including Vertafore—see AI as a tool to help make work more manageable, improve efficiency, and create space for higher-value thinking and better customer service.
The key to managing employee fears is transparency and communication. We addressed employee concerns early and reinforced our intent to use AI to support our human teams, not replace them. That's especially true in a relationship-driven industry like insurance, where judgment and accountability cannot be outsourced to a model.
We reinforced that human oversight remains central, established clear security guardrails, and created structured opportunities for employees to ask hard questions and understand expectations. This included a simple framework: asking "Should I?" instead of "Can I?" Employees were encouraged to consider who benefits, who could be harmed, and whether AI was truly the best fit for a given scenario. The framework reinforced that AI supports professional judgment rather than replacing it.
To measure the impact, we surveyed employees before and after our immersion week. We didn't focus only on productivity metrics or efficiency gains. We measured employee confidence. We positioned AI as a productivity multiplier and asked one core question: How do I make myself, my team, and our products better for the end customer?
The results speak for themselves: confidence rose 13% and those that were "not confident" dropped 11%. Employees who saw AI as a "go-to collaborator" doubled and expected daily or near-daily use rose 27%. Among people managers, confidence using AI rose 23%, with 80% saying they're confident with helping their teams use AI responsibly.
Create space to learn
To help employees build real skills, we made a deliberate choice to set aside dedicated time for learning.
Instead of squeezing experimentation between daily responsibilities, teams were given bandwidth to explore AI. Leaders set up time for employees to test ideas, collaborate across teams, and explore ways to apply AI to their specific roles. And as an organization, we created full-company sessions to dive into AI questions and hear from experts on their lessons learned from putting this technology to work.
Prioritizing learning in this way accelerated adoption. Our teams spent time getting to know the AI tools in the best way possible: by seeking solutions to the challenges and opportunities they encounter every day. Not every idea and trial made it out of the gate. And that's okay. But many of our teams built the foundation for real AI applications.
Internally, our learning and development team created a custom GPT to help employees and managers build individual development plans that support career growth. The tool guides users through identifying where to start, developing skills within their current role, or preparing for a move into another internal role. By following prompts and exploring development ideas, employees can create clear, actionable plans while discovering recommended resources such as LinkedIn Learning and other company tools.
How to measure what matters
If employee confidence with AI increases, it stands to reason that successful adoption will follow. With that in mind, we treated the mindset shift as the first KPI. From there, we narrowed our focus.
Rather than attempting to transform the entire organization at once, we identified one to two areas where return on investment (ROI) could be clearly measured within a defined timeframe. Functions such as customer support often provide the clearest early signals because call times and quality improvements are easier to quantify.
For each use case, we established a clear hypothesis to answer:
- Where will AI create value?
- What outcomes should we see within the first 90 days?
Within customer support, for example, we used AI to refine knowledge base articles so representatives can quickly access the most up-to-date product information during calls. Measurable outcomes include reduced time spent searching for information, faster turnaround between calls, and higher customer satisfaction.
Another area we tracked was product development, with defined objectives around output volume and delivery timelines. We've established targets to measure progress and ensure consistent improvement over time. Peer experts also led office hours that created a safe space to test ideas, ask questions, and build confidence.
By focusing on a specific department, establishing clear goals, and tracking quantifiable metrics, such as time saved, communication volume, and call response time, you can better determine the next steps for scaling AI across your teams.
Lessons learned
Our early efforts drove real engagement, but they also exposed some opportunities. Not every team benefited in the same way, and not every employee felt ready to apply what they'd learned. That tension helped clarify what actually drives progress. Here are the lessons we took forward:
- AI adoption isn't one-size-fits-all. Teams across the organization have very different AI use cases and skill levels. All-company sessions, such as discussions with external business leaders, offered useful perspectives but did not give employees the role-specific guidance they needed to apply AI in their daily work.
- Progress doesn't happen at the same pace. Some individuals jumped in quickly, while others needed more context and support. Moving forward required patience, targeted enablement, and learning experiences tailored to where people actually were rather than where we expected them to be.
- AI success starts with mindset, not mandates. Tools alone don't drive transformation. Confidence does. When employees understand the "why" and feel equipped to experiment, adoption follows. That's where change management becomes a true advantage. Organizations that invest in building understanding and trust are best positioned to turn AI from a tool into a differentiator.
A clear path forward
AI transformation begins with people. We learned that dedicating focused time for learning helps employees step away from daily work and fully engage in building new skills.
Pairing that time with privacy and security training ensured employees explored AI tools while understanding the company's protocols for responsible use. When employees understand the purpose, feel supported as they learn, and trust the guardrails in place, AI stops being an abstract initiative and becomes part of how the business operates. In insurance, where judgment and relationships matter, that human foundation is everything.
Invest in your people first. The technology will follow.
