To successfully pilot AI coaching in an insurance business, you need to start with a single measurable department, map the specific work personalities of that team, and frame the technology as confidential career support rather than performance surveillance.
Key takeaways
- Starting small with a specific unit like claims or underwriting provides clear baseline metrics for your pilot.
- Understanding the work personalities of your team ensures the coaching adapts to how they actually process information and handle stress.
- Positioning the tool as private development rather than management oversight drives higher engagement from skeptical staff.
- Measuring success requires looking at behavioural shifts in conflict resolution and decision-making rather than just login frequency.
Insurance businesses face a unique set of pressures. Staff are caught between strict regulatory compliance and highly emotional customer interactions. A claims handler might spend their morning explaining a denied policy to a distressed family and their afternoon navigating complex internal approval systems. The cognitive load is massive.
When these employees burn out, they leave, taking years of technical knowledge with them. Replacing an experienced underwriter or broker costs the business significantly in recruitment fees and lost productivity.
When you try to support these teams with generic training modules, the advice falls flat because it ignores how different people actually work. An underwriter who naturally defaults to cautious, detail-oriented work needs an entirely different coaching approach than a broker who thrives on networking and big-picture thinking. Throwing a standard leadership course at both of them wastes time and money.
This is why many insurance firms are looking at AI coaching. They want scalable support that adapts to the individual. Rolling out new technology across a heavily regulated, risk-averse organisation requires a careful approach. People are naturally suspicious of anything that looks like monitoring.
The biggest mistake companies make with new technology is going too broad too fast. A company-wide launch dilutes your focus. You end up with scattered data and no clear narrative about whether the software actually works.
Isolate a specific group for your pilot. A claims department is often the best place to start. These teams face high daily stress, frequent conflict, and strict performance metrics. If you can move the needle on a claims team's resilience and communication, you have a proven use case for the rest of the business.
Alternatively, an underwriting team offers a different test environment. Here, the stress comes from high-stakes decision-making and balancing risk against revenue targets. Choose one group of 20–50 people. This size is large enough to provide meaningful data but small enough that you can personally check in with the participants.
You cannot effectively coach someone if you do not understand how their brain works. Generic coaching programmes assume everyone learns and processes feedback in the exact same way.
At Compono, our research into organisational psychology shows that teams are made up of distinct work personalities. Consider the difference between an actuary and a sales broker. The actuary might align with 'The Auditor' personality – they are methodical, reserved, and need time to process information before making a decision. If a coach pushes them for immediate answers, they will disengage.
The broker might be 'The Campaigner' – energetic, future-focused, and eager to talk through big ideas. They need a coaching style that matches their pace. An effective AI pilot must account for these differences. The system needs to adapt its tone and pacing to the individual.
If you want to see how this adaptation works in practice, you can explore personality-adaptive coaching to see the methodology in action.
The insurance sector operates under intense scrutiny. Employees are accustomed to having their calls recorded and their processing times ranked against their peers. When HR announces a new AI tool, the default assumption is that management has found a new way to monitor them.
You must dismantle this fear immediately. The success of your pilot depends entirely on psychological safety. If staff believe their manager gets a weekly report of their coaching conversations, they will only discuss safe, superficial topics. They will never admit that they are struggling to handle an aggressive client or that they feel overwhelmed by a new compliance framework.
Position the coaching strictly as a confidential career development tool. Be explicitly clear about data privacy. Show them exactly what aggregate data management will see and assure them that individual conversations remain private.
Many businesses evaluate software pilots by looking at login frequency. They assume that if a majority of the team opened the app twice a week, the pilot was a success. Building an app-usage habit is meaningless. The goal is building better employees.
Before the pilot begins, sit down with the managers of the test group and establish behavioural baselines. Ask them what specific actions they want to see change. For a claims team, success might look like a reduction in calls escalated to a supervisor. It might look like staff recovering faster after a hostile customer interaction.
For an underwriting team, success could be clearer communication of risk factors to the sales department without causing internal friction. Write these desired behaviours down. They become the true measure of your pilot's impact.
When staff start receiving personalised coaching, their behaviour will change. They might start setting firmer boundaries, communicating more directly, or asking more strategic questions. Managers need to be ready for this shift.
If a team member learns a new conflict resolution technique from their coach, but their manager shuts them down when they try to use it, the coaching is wasted. The manager becomes a bottleneck for growth.
During the pilot phase, keep the managers closely involved. They do not need to see the confidential coaching transcripts, but they should understand the concepts their team is learning. Educate the managers on the different work personalities in their team. When a manager understands that their direct report is an 'Evaluator' who needs logical, data-driven feedback, they can reinforce the lessons the AI coach is providing. This alignment creates a compounding effect on team performance.
A pilot is only useful if you learn from it. At the end of the 90-day period, gather qualitative feedback to sit alongside your behavioural data. Do not just send out a generic survey. Have honest, direct conversations with the participants.
Ask them if the coaching advice felt relevant to the specific pressures of the insurance industry. Did the AI understand their natural communication style? You will likely find that different work personalities engaged with the tool in completely different ways. 'The Doer' might have loved the practical action plans, while 'The Advisor' might have appreciated the open-ended reflective questions.
Document these insights. When you take the business case to the executive team for a wider rollout, this combination of behavioural data and personal feedback is incredibly compelling. You are presenting a proven method for supporting the specific people in your business. If you are ready to test this approach with your own teams, Hey Compono provides the framework you need to get started.
Key insights
Starting your pilot with a small, measurable department like claims or underwriting establishes clear baseline metrics for success. Mapping the work personalities of your team beforehand ensures the coaching adapts to how they naturally process information. Communicating clearly that the AI coach is a confidential development tool prevents staff from viewing it as a performance monitoring system. Measuring success requires looking for observable behavioural changes in the workplace rather than just tracking software login rates.
Ready to give your insurance team support that actually fits how they work? Start mapping their work personalities today and see the difference adaptive coaching makes.
Hey Compono helps teams give and receive feedback that actually moves the needle. Start free and see how it fits your workflow.
A pilot should typically run for 90 days. This gives staff enough time to build trust with the tool, apply the coaching advice to real scenarios, and demonstrate measurable behavioural changes in their daily work.
High-stress, highly measurable teams make the best pilot groups. In an insurance business, claims departments or underwriting teams are ideal because they face clear daily challenges and have established metrics you can track for improvement.
Yes, staff in highly regulated industries usually assume new technology is for surveillance. You must explicitly frame the tool as confidential support and be completely transparent about what data management can and cannot see.
Different work personalities need different communication styles. A detail-oriented person requires structured, methodical feedback, while a big-picture thinker needs open-ended, creative discussions. Coaching that fails to adapt to these preferences is usually ignored.
Look beyond software login rates. Real success is measured through behavioural shifts, such as reduced call escalations, better conflict resolution between departments, and improved clarity in team communication.