The conversation about agentic AI in insurance has been dominated by process design and governance. Carrier executives are asking the right questions about control, audit trails, and escalation paths. But there is a second-order problem that is not getting enough attention: the people who actually run claims are afraid of the technology. And that fear is not a soft cultural issue. It is a hard operational constraint.
A recent Glassdoor analysis found that claims adjusters report higher levels of AI-related fear and dislike than almost any other job category in insurance. Meanwhile, a separate report shows overall employee confidence in the industry is tanking on AI concerns. These are not isolated anecdotes. They are signals that the workforce responsible for executing claims is bracing for displacement, not embracing augmentation.
Here is the pattern I am watching: the carriers that are moving fastest on agentic AI are the ones that treat adjuster trust as a design requirement, not an afterthought. They are not just building agents that work. They are building agents that explain themselves, that hand off cleanly, and that make the adjuster look good in front of the policyholder. The ones that skip this step are going to hit a wall, not because the technology fails, but because the humans who are supposed to supervise it will quietly resist, slow down, or find workarounds.
The Fear Is Rational, and That Changes the Design
It is easy to dismiss adjuster fear as resistance to change. But the fear is rational. The claims process has been commoditized and squeezed for years. Adjusters have watched their roles get narrowed by automation, outsourcing, and tighter cycle time targets. When an AI agent shows up, it is not a stretch for them to assume they are next.
That rational fear has a direct impact on agentic AI deployment. An agent that requires human review but is met with a skeptical, disengaged reviewer will not perform well. The human will rubber-stamp outputs they do not trust, or they will overrule the agent at every turn, destroying the efficiency gains. Either way, the agent fails.
The carriers that understand this are designing agents with the adjuster in mind. They are building in transparency features that show the agent's reasoning, not just its conclusion. They are creating exception queues that feel like a tool, not a surveillance system. They are measuring success
Governance Is Not Enough
There is a lot of good thinking right now about agentic AI governance. The Carrier Management pieces on redesigning insurance processes for agentic AI are asking the right questions about control frameworks, audit trails, and human-in-the-loop requirements. That is necessary work. But governance is a top-down solution. It tells the adjuster what the agent is allowed to do. It does not tell the adjuster why they should trust it.
Trust is built bottom-up. It comes from the adjuster seeing the agent handle a routine claim correctly, then a complex claim correctly, then a weird edge case correctly. It comes from the agent being able to explain, in plain language, why it made a particular recommendation. It comes from the adjuster knowing they can override the agent without political consequences.
That is a different design problem than governance. Governance is about control. Trust is about confidence. You need both, but they are not the same thing.
What This Means for Operators
If you are running claims operations at a carrier or an IA firm, the takeaway is simple: your agentic AI project will live or die on adjuster buy-in, not on model accuracy. You can have the best agent in the market, but if your adjusters do not trust it, they will not use it effectively.
That means you need to invest in change management as much as you invest in technology. You need to bring adjusters into the design process early. You need to show them how the agent makes their job better, not just faster. You need to give them a clear path for escalation and override, and you need to make sure that path is actually used and respected.
This is also where the restoration and contractor side comes in. The same trust dynamic applies. Contractors are skeptical of AI that tries to replace their judgment on a roof or a wall. They will adopt tools that make their estimates more defensible, but they will reject tools that feel like a black box. The companies that win in this space will be the ones that build agents that are transparent, explainable, and respectful of the human's expertise.
The Opportunity for Built by SMR
We are seeing this pattern play out in real time with the work we do. The integrations and AI agents we build are not just about automating a step. They are about giving the adjuster or contractor a tool that they can trust, that they can verify, and that they can explain to a policyholder or a client.
That is the difference between an agent that gets deployed and an agent that gets adopted. And adoption is where the real ROI lives.
This is the kind of trust-driven agentic AI problem Built by SMR is being asked to solve for carriers, IA firms, and restoration companies right now.