Coaching at Scale, Part 2: How to Introduce AI Tools In a Way Teachers Actually Embrace

This is the second in a three-part series, Coaching at Scale. In Part 1, we explored why even well-resourced coaching systems struggle to move practice at scale — and how AI changes that equation. Here, we look at how to introduce these tools in a way that makes the human side of coaching more powerful, not less.

Start with Trust, Not Tools

The biggest mistake districts make is leading with the technology. They announce a new tool, explain the features, and expect teachers to embrace it. Then they're surprised when adoption is sluggish.

Here's what they missed: Teachers don't resist coaching. They resist being evaluated under the guise of coaching.

Before introducing any AI tool, be crystal clear about what this is for and who it's for. The most successful implementations start with three commitments:

The data belongs to the teacher first. Teachers should see their own insights before anyone else — principals, coaches, or district leaders. They decide what to share and when. Coaches can see aggregated, anonymized trends, but individual teacher data remains private unless the teacher chooses to share it.

This single design choice changes everything. It shifts the tool from feeling like surveillance to feeling like support.

This is about coaching, not evaluation. AI-generated insights should never feed into evaluation processes. Not indirectly, not "just to add context," not even if a teacher volunteers it. The moment data has evaluative stakes, teachers will optimize for what the tool measures instead of what actually matters.

Coaches are collaborators, not compliance monitors. Train coaches to lead with curiosity, not prescription. When a teacher shares data showing they're doing most of the talking, the coaching question isn't "How will you reduce your talk time?" It's "What do you notice? What surprised you? What do you want to try differently?"

Celebrate the Wins

One of the most underutilized opportunities in AI-powered coaching is celebration.

When teachers see data showing they're already doing strong instructional work — high student talk time, open-ended questions, equitable participation — that's a moment to recognize and reinforce what's working.

Recognition makes strengths visible. Many teachers don't realize they're doing something well until they see the data. A fourth-grade teacher might not know her wait time averages seven seconds — well above the typical two-second average — until the tool shows her. That's a skill worth naming and celebrating.

Make celebration specific and connected to practice. Not "great job," but "I noticed your students are talking for 60% of class time — what are you doing to make that happen?" That kind of recognition gives teachers language to describe their own expertise.

Build Reflection Into Existing Workflows

Teachers don't have extra time. The districts seeing high adoption integrate AI tools into rhythms teachers already have.

PLC time becomes data-informed collaboration. Instead of planning lessons in isolation, teams look at aggregated classroom data together. A middle school math team might notice they're all asking mostly closed questions during problem-solving. That becomes a shared goal: What would it look like to ask more open questions?

Coaching cycles get faster and deeper. With AI handling data collection, cycles compress. A teacher can record three lessons in a week, reflect on patterns, and bring specific questions to a 15-minute coaching check-in. Instead of treating coaching as a big event, it becomes an ongoing conversation.

Professional learning connects to practice immediately. When districts offer PD on student discourse, teachers can use AI tools to see how those strategies play out in their own classrooms. Teachers try a new approach on Monday, see their data on Tuesday, and adjust on Wednesday. That kind of immediacy changes what professional learning can be.

In Part 3, we'll look at what it actually takes to build the conditions for this to work — the roadmap, the cultural shift, and the leadership questions that determine whether AI coaching becomes transformative or just another layer of compliance theater.

Jamie Poskin is CEO of TeachFX, an AI-powered reflection tool that helps teachers analyze their instructional practice and accelerate their growth.