Bringing your support team through an AI rollout

Be transparent about what’s changing, and why

The teams that handle AI transitions well explain a new way of working to their people. That means being clear about what’s changing (the nature of support work, the shape of roles, how performance gets measured), why it’s happening, and what success looks like for the team, not just the organization.

If people find out what AI means for their role secondhand, or piece it together from ticket volume dropping, you’ve already lost trust. Get ahead of it. Describe what’s happening honestly: AI will handle more of the work human team members have traditionally done, which means your team’s focus moves toward more strategic, and improvement-oriented tasks.

Reset expectations on time allocation explicitly

One of the quickest paths to failure is expecting people to take on AI-related responsibilities on top of their existing workload. From monitoring resolution trends to improving knowledge content and reviewing handoff patterns, this work is important, but gets dropped the moment the queue fills up if you haven’t explicitly made time for it.

Reset the expectation clearly: not all of a rep’s time will go to direct customer interactions anymore. Build in dedicated “out of the inbox” time for the work that makes your Agent better. Without this, the feedback loops that drive improvement never get built.

Make AI performance a shared goal, not a leadership metric

If performance tracking is something only leadership cares about, the team loses the connection between their work and outcomes. Bring resolution rate, handoff patterns, and knowledge gaps into team discussions. Make it visible and something everyone has a stake in improving.

This changes the dynamic from “AI is a thing being done to us” to “AI is a system we’re collectively responsible for,” which is a much more productive approach in the long-run.

Empower your team to improve the system

Your reps are closest to the customer. They know where the Agent is struggling, where handoffs feel clunky, and where content is missing or outdated. Ensure there’s a channel for that knowledge to flow into the system, whether that’s a regular review of top handoff topics, a simple way to flag content gaps, or a standing agenda item in team meetings.

At Fin, when rolling out our AI Agent with our own support team, we highlighted the value of knowledge quality explicitly and encouraged every team member to identify content improvements. We also tracked resolution rate as a shared team metric, giving everyone a concrete goal to work toward together. The result was everyone feeling like they were actively contributing to AI performance.

Give people a visible path forward

As roles evolve, people need to know where their careers are headed. The new responsibilities that come with AI (knowledge management, conversation analysis, AI operations) should map to clear progression opportunities, not just be absorbed into existing job descriptions without acknowledgment.

Recognize and reward the strategic work, not just throughput. If the only metrics you track are average handle time (AHT), cases handled, and first contact resolution (FCR), you’re measuring the old model. Build in ways to surface and celebrate the higher-leverage work your team is doing as AI takes on more of the queue.

Start with an honest conversation

This week, schedule a conversation with your team about what’s changing.

Ask them:

• What do you understand about how our Agent is performing?
• What do you wish you had more visibility into?
• What would make it easier for you to help it improve?

The answers will tell you exactly where your change management approach has gaps. Start there.