Rethinking AI in the Classroom with Kami: Key Insights from the CAS Community Meeting

If you were unable to join us for the Kami online community meeting, don't worry! You can catch up on all the content and a recording of the session below

Key Takeaways

• Kami’s approach to AI focuses on supporting student understanding rather than generating teacher resources.
• The platform combines annotation, accessibility, and collaboration tools within existing workflows like Google Classroom and Microsoft Teams.
• AI features such as translation, explanation, and guided support are designed to scaffold learning without simply providing answers.
• Teachers can use tools like split and merge, text recognition, and live class monitoring to adapt and manage resources more efficiently.
• The session prompted wider discussion about how AI can be used safely, thoughtfully, and inclusively in the classroom.

A practical look at AI-supported learning

During this CAS online community meeting, Wendy Peskett introduced attendees to Kami and explored how the platform is being used to support teaching, accessibility, and student independence in the classroom. The session focused less on AI as a content-generation tool for teachers and more on how carefully designed AI features can support learners directly.

Wendy began by explaining that Kami operates on a freemium model, with a range of annotation and document-management tools available for free, alongside premium AI-supported features. She also highlighted the company’s emphasis on data privacy and security, something that many schools continue to scrutinise carefully when considering AI tools.

A significant theme throughout the session was reducing friction in teachers’ existing workflows. Rather than requiring teachers to learn an entirely new system, Kami integrates directly with platforms many schools already use, including Google Classroom and Microsoft Teams.

Supporting teachers with existing resources

One of the early demonstrations focused on opening and adapting existing teaching materials. Wendy showed how PDFs, Word documents, slide decks, and web-based worksheets could all be opened directly within Kami for annotation and editing.

Particular attention was given to the “split and merge” tool, which allows teachers to separate or combine pages from PDFs. Wendy demonstrated how this could be used to remove answer pages from worksheets or create custom revision packs by combining questions from different exam papers.

For many attendees, this practical focus on adapting existing materials rather than creating entirely new ones appeared to resonate strongly. The discussion repeatedly returned to the idea of saving teacher time while still maintaining control over classroom resources.

Accessibility and inclusive practice

Much of the session explored how accessibility features can help students engage more independently with learning materials. Wendy demonstrated annotation tools including highlighting, text boxes, drawing tools, and collaborative commenting, all of which can be used directly on digital worksheets without printing.

One particularly useful feature was text recognition. Wendy demonstrated how scanned PDFs or image-based worksheets could be converted into selectable text, allowing them to work with read-aloud tools.

The conversation then shifted towards AI-supported accessibility tools. These included:

• Picture dictionaries for vocabulary support
• Translation into nearly 200 languages
• Read-aloud support in different languages and dialects
• Simplified explanations of concepts and questions
• Speech-to-text functionality for student responses

 

Importantly, Wendy emphasised that these tools are intentionally limited in scope. Students are not interacting with open-ended AI chatbots or prompts; instead, the AI features are constrained to carefully designed support functions.

This prompted useful reflection around how schools are currently approaching AI access. Some participants noted that AI tools are blocked entirely for pupils in their schools, while others discussed the challenge of balancing safeguarding with opportunities for accessibility and personalised support.

AI as scaffolding, not shortcutting

One of the most interesting aspects of the session was the distinction between “giving answers” and “supporting understanding”.

The “Explain” tool was demonstrated as a kind of contextual learning assistant. Rather than solving questions outright, it guides students towards the key features they should notice or the steps they should consider. Wendy described it as “a little teacher on your shoulder”.

Examples included:

• guiding students through mathematical methods
• helping distinguish between insects and spiders
• simplifying explanations for different reading levels
• translating explanations into other languages

 

This approach sparked broader reflection on what effective AI support in education might actually look like. Rather than replacing thinking, the tools demonstrated during the session aimed to scaffold understanding while keeping students focused on the original task and language of instruction.

Live classroom support and feedback

Towards the end of the session, Wendy demonstrated some of the premium classroom-management features available through LMS integrations.

The “Class View” tool allows teachers to view multiple students’ work simultaneously, provide live feedback, and monitor progress in real time.

This led into a discussion about supporting students who may struggle to participate verbally in lessons. Wendy shared examples of using private written prompts and feedback to support anxious learners or school refusers without drawing attention to them publicly.

The session also covered AI-assisted feedback generation. Teachers can generate draft comments based on student work, then edit or approve them before students see anything. Wendy stressed that the AI is intended to reduce workload rather than replace professional judgement.

The final demonstrations explored self-marking questions and analytics dashboards, enabling teachers to identify misconceptions across a class quickly and adapt teaching accordingly.

Next Steps: Questions to reflect on in your own practice

As with many CAS community meetings, the most valuable part of the session was not necessarily the tool itself, but the wider conversations it prompted about teaching, accessibility, and the role of AI in learning.

You might want to consider:

• How accessible are your current digital resources for students with different learning needs?
• Are there areas where students could be given more independence through scaffolding tools?
• How do you currently balance AI restrictions with accessibility and inclusion?
• Which classroom tasks genuinely benefit from AI support — and which do not?
• Could live digital feedback change how you support quieter or less confident learners?

Possible classroom activities to explore

• Ask students to compare AI-generated explanations with teacher explanations and identify differences.
• Trial annotation-based revision activities using PDFs rather than printed worksheets.
• Explore multilingual support tools with EAL learners while maintaining English-language instruction.
• Use collaborative documents for peer review or debugging exercises in computing lessons.
• Investigate how live feedback tools could support formative assessment during programming tasks.

Further Resources

Kami official website

Kami Teacher Resources Library

Kami for Google Classroom

Event slides