AI Coaching Teachers Actually Want


Reading time: 4
Happy Teacher

@TeacherToolkit

Ross Morrison McGill founded @TeacherToolkit in 2007 and is widely recognised as one of the leading influencers in education in the UK and across the world. In 2015, he was named among The Sunday Times/Debrett’s 500 Most Influential People in Britain for his impact on...
Read more about @TeacherToolkit

Can AI coaching for teachers improve classroom practice without becoming another surveillance tool?

Teachers are right to be sceptical about AI. The question is not what the technology can do, but whether teachers would actually trust it.

I got lucky when I trained to teach in the 1990s. My mentor placed a video camera at the back of the classroom and filmed me teaching. This soon became a habit, and it was the best professional development I received. However, I did not have a weekly coach, a video platform, or an AI assistant to analyse what was happening.

Every new platform claims to save time, improve feedback and transform professional development; however,  teachers have heard this promise before. Too often, it becomes another login, another dashboard, another thing teachers are expected to use.

AI Coaching for Teachers

So, when we talk about AI coaching for teachers, the first question should not be, “What can the technology do?” The better question is: Would teachers use it, trust it and return to it?

Starlight is a UK-built AI teaching coach designed to support private, whole-lesson reflection. It’s a tool I would have loved to have used as a classroom teacher!

3 Starlight LogoA teacher records the audio of a lesson, uploads it, and receives a structured coaching report within minutes.

The platform report draws on the transcript, identifies strengths, suggests a development focus and offers practical actions the teacher can keep, adapt or ignore.

That last part matters.

Good coaching does not remove teacher agency. It sharpens it. That is why some schools are now exploring AI coaching for teachers.

Does AI feedback feel generic?

This is the first big objection I would ask. Teachers do not need bland feedback that could apply to any classroom, in any school, on any day. We sometimes receive generic feedback even from people who know them well and observe their teaching regularly. How can we ensure that AI doesn’t make that problem worse?

Starlight’s answer is whole-lesson coaching analysis. It works from the actual lesson transcript, not a short clip, a selected performance moment, or a generic prompt. The richer the input, the more useful the reflection can become. That means teachers can begin to explore the details that often disappear after a busy lesson, and analyse the classroom data we often miss as humans. Questions a teacher could ask include:

  1. How much did I talk?
  2. How long did I wait after asking a question?
  3. Where did pupils contribute?
  4. Where did my explanation drift?
  5. What could I change in the next lesson?

You can try Starlight for free this term!

5 Audio Upload Interface
6 Create Personalised Action Feature Screenshot
7 Create Personalised Action Feature Screenshot 2
8 Actions Feature Screenshot
19 Constellation Dashboard Screenshot 2

Is this coaching or compliance?

The second objection, and most important, is trust.

Could lesson audio be used for performance management, accountability or inspection? That concern is legitimate. In too many schools, “developmental” systems have slowly drifted into monitoring.

Starlight deliberately sits outside performance management. Reports are private to the teacher. Senior leaders see anonymised, aggregated trends only. The leadership view does not name individual teachers, and teachers can delete their data at any time.

This is not a small technical detail. It is the difference between coaching and compliance.

Does Starlight reduce workload?

The third objection is workload. Teachers already carried enough workload; therefore, another platform will fail if it creates more admin. Starlight’s promise is that uploading takes under a minute, reports arrive within minutes, and the Lesson Toolkit turns the transcript into useful resources. This includes recap prompts, retrieval questions, key vocabulary, EAL support and follow-up materials teachers would otherwise create from scratch!

That is where Starlight becomes more than reflective CPD. It starts to touch on the daily workload problem too.

11 Report Ready Email Alert
12 Teacher To Teacher Observation Feature
13 Lesson Toolkit Retrieval Questions
14 Lesson Toolkit Key Terms And Definitions
15 Lesson Toolkit Recognition Suggestions
16 Lesson Toolkit Email For Missing Students
17 Lesson Toolkit Homework Instructions
18 Constellation Dashboard Screenshot

What does the evidence say?

Instructional coaching has a strong research base, but scaling it is difficult. I have written before about instructional coaching, and the central tension remains the same: coaching works best when it is specific, regular and rooted in evidence. The difficulty is finding enough time and expertise to make that happen for every teacher.

Plus, there is also growing interest in AI-assisted coaching. For example, Harvard’s Centre for Education Policy Research has explored AI-supported teacher coaching in mathematics through its M-Powering Teachers project.

AI coaching that teachers want

Importantly, Starlight does not try to replace human coaching. At its best, it could make coaching more frequent, more specific and more sustainable. It supports all the philosophies I advocate in teaching. Coaching, not compliance. Insight, not surveillance. Growth, not grading.

And that is the version of AI coaching for teachers that many teachers actually want.

What about ethics, GDPR and consent?

Any school considering AI feedback for teachers should ask serious questions before a pilot begins. With lesson audio, trust cannot be vague. It has to be designed into the system.

Crucially, Starlight gives school leaders clear answers. The school remains the Data Controller, and Starlight hosts data in the EU (read more). Teachers can delete their data at any time. Individual coaching reports remain private to the teacher, while school leaders only see anonymised, aggregated trends. Starlight also has a published DPIA.

This matters because teacher trust is not a feature you can bolt on later. If teachers suspect that private reflection may drift into performance management, the culture will collapse before the software has any chance to help.

Before adopting any teacher coaching platform, school leaders should still ask:

  1. Will teachers choose to use it?
  2. Is individual feedback genuinely private?
  3. Does it reduce workload or add to it?
  4. Does it support professional judgement?
  5. What evidence will show that practice changes over time?

You can try Starlight for free this term.

The promise of AI in schools will not be found in shiny dashboards. It will be found in tools that help teachers think more clearly, work more efficiently and retain control over their professional growth. If I were still in the classroom, I would be using this without hesitation.

This is a promotion. 

Share?