Instructional Coaching Meets AI

The benefits of instructional coaching for teachers and their students are undeniable, and indeed not really a debate anymore. For what it is worth, Hattie, in his Visible Learning For Teachers, reported an effect size of .51 for the practice, well beyond the targeted effectiveness threshold. What does remain problematic, however, are 3 outstanding issues. Fortunately, recent advances in artificial intelligence provide a simple solution. This is why I took advantage of the simple interface provided by ​Poe.com to design an AI-powered instructional coach: Ed.

---

The 3 outstanding issues are: 

1) How can we create the conditions needed for teachers to feel psychologically and socially safe working with a coach?

2) How can we help coaches reach all of the teachers that can use their assistance (which means all of them)?

3) How can we build school cultures where coaching happens beyond formal coaching cycles and becomes embedded in all professional conversations?

Ed is not meant to replace instructional coaches (and to put me out of a job!). On the contrary, its goal is to create favorable conditions for coaches by addressing the 3 issues above, thus helping them recruit teachers, focus on the services that they alone can provide (from in-depth work to classroom visits, expert recommendations, and training), and disseminate their approach throughout the school community.

Lowering Entry Barriers to Coaching

​Some teachers feel a degree of unease or apprehension about traditional, human coaching. While the practice is by definition non-judgmental and respectful of professional ability, teachers might not be fully convinced that it is so until they try it. Ed solves this Catch-22 by providing an anonymous, artificial, and psychologically safe environment where teachers can give coaching a try​.

Scaling Coaching

Up until now, human resources and scheduling/workload constraints have limited the reach and scope of instructional coaching, reducing its potential impact. Chatbots, however, have unlimited availability. Ed can coach any number of teachers, at any time. And it is free. It is certainly not able to provide the full experience and benefits of instructional coaching, but coaching has always involved a range of services, from informal conversations to year-long PD implementation plans. What Ed can provide, however, is a highly flexible coaching "chat" that teachers can start with, or go back to, when and where they want. Ed is thus able to assist you as you problem-solve on the go or reflect at the end of a unit.

Creating a Coaching Culture

Just like they sometimes fear being judged by a coach, teachers can also be worried about their colleagues' perception. While the former is completely misplaced when working with a qualified coach, the second really depends on each school's culture, level of collegiality, and familiarity with coaching. Here again, chatbots can help. Indeed, disseminating coaching also means normalizing the practice, and thus creating both new individual habits and new social norms. Once they become routine, the spirit of coaching becomes embedded in everyday practice and interactions, enriching conversations and reverberating throughout the educational community.

Building-In Best AI Practice

Favorable conditions make coaches' job easier, but the main beneficiaries of an AI instructional coach are of course teachers and their students. Relying on sophisticated instructions, Ed makes it possible for teachers to leverage the potential of AI and to receive high-quality answers (or rather, questions) without any need to develop new technical expertise such as prompt engineering. Likewise, being a chat bot, it ensures that users will employ recursive prompting (feeding back its products to AI for further processing). As a matter of fact, Ed should even invite you to come back with your observations for a follow-up conversation--a practice I call "long-term prompting". This does not only replicate the ongoing assessment and adaptation that would happen with traditional coaching, but also emulates the continuous dialogue and partnership a teacher and coach would have.

Promoting Self-Coaching

But isn't it true that the power of coaching precisely relies on its human dimension? To an extent, yes, and Ed is to an actual coach what AI is to human intelligence: an artificial mimicry. This means that it is not as good, but this does not mean that it is no good at all. What is more, it is also true that the goal of coaching is, ultimately, self-coaching. Perfect self-direction being an aspiration and a "guiding ideal", as Kant would say, rather than an achievable objective, dialogue and an interlocutor are always needed in practice. But Ed, just like a coach, can still help people develop their capacity to reflect independently on their own practice. By doing so, it makes coaching more accessible (for some teachers, the human dimension acts as a barrier) and will hopefully drive more teachers to work with actual coaches.

Limitations

Ed has all the limitations of generative AI. Users should be mindful of what they share, and how, to maintain privacy and confidentiality. It is also not 100% reliable, and should be used with proper critical thinking. This might not be as much as an issue here as it is for other AI uses, as Ed is mostly here to ask questions. Still, you can ask it for suggestions, explanations, etc. In a way, this is still for the better: as a coaching bot, Ed should only assist, and certainly not replace, human reflection and professional judgment. Our goal, after all, is to promote self-coaching. What is more, putting teachers in a situation where they are to learn with and evaluate AI is a great way to give them a sense of what this new reality means for their students, and how they should approach it.

In conclusion, I should make it absolutely clear that Ed is an experiment—and should be used and treated as such. Experimenting involves taking risks, which are necessary to innovate and make progress, but should be kept in mind and minimized. Thus, instructional coaches should trial Ed first before proposing it to colleagues. When they do so, they might still want to schedule a few conversations to check in with teachers willing to give it a try. With that being said, I hope that Ed will help promote coaching by lowering entry barriers, enabling deployment at scale, and facilitating the emergence of receptive cultures--in addition to helping teachers, and their students, reap the full benefits of AI and understand its limitations and appropriate uses.
​
Ed is accessible here (free, without limits) if you want to give it a try and see what it can do for you and for your school.