Artificial intelligence is reshaping education, but not all AI tools serve the same purpose. For UK teachers navigating an increasingly complex technology landscape, understanding when and how to use different AI tools matters as much as understanding which tools to use.

This guide breaks down the most useful AI tools across the three critical phases of teaching: preparation, delivery, and post-lesson reinforcement.

Understanding the three phases

Before diving into specific tools, here's what we mean by each phase:

Phase 1: Preparation – Everything before pupils enter the room: curriculum planning, resource creation, differentiation strategies, assessment materials.

Phase 2: Teaching – Live delivery of instruction. This is where teacher expertise, subject knowledge, and responsiveness to pupils matter most.

Phase 3: Post-lesson (our focus) – When learning needs consolidating, checking, and extending through formative assessment, homework, retrieval practice, and gap identification.

Different AI tools excel in different phases. Here's what UK teachers need to know.

Phase 1: Preparation tools

According to the 2024 Teacher Workload Survey, teachers spend an average of 12 hours per week on planning and preparation. These tools aim to reduce that burden whilst maintaining quality.

Oak National Academy

What it does: Comprehensive curriculum resource with over 40,000 lessons spanning primary and secondary subjects.

Why it's useful: All lessons created by subject specialists and aligned with the National Curriculum. Recent AI features help teachers adapt pre-made lessons to their context.

Best for: Schools working with limited planning time or covering curriculum areas outside staff expertise.

Limitations: Pre-made content may not fit your exact pupil context or school's curriculum sequence.

MagicSchool AI

What it does: Generates lesson plans, differentiated resources, and assessment materials through 60+ education-specific tools.

Why it's useful: Structures inputs around pedagogical needs (learning objectives, pupil age, prior knowledge) rather than requiring perfect prompts.

Best for: Creating first drafts of worksheets, quizzes, and differentiated materials that you then refine.

Limitations: Outputs require professional oversight—AI can produce grammatically correct but pedagogically weak resources.

Curipod

What it does: Creates AI-powered slide-based lessons with embedded formative assessment.

Why it's useful: Saves time on mechanical aspects of resource creation while building in checks for understanding.

Best for: Interactive presentations where you want to poll pupil understanding throughout.

Limitations: Still requires teacher input on learning outcomes and success criteria.

Diffit

What it does: Adapts complex texts to different reading levels whilst maintaining core content.

Why it's useful: Solves the problem of finding age-appropriate texts on specific curriculum topics.

Best for: Primary teachers needing accessible science or history texts; secondary teachers differentiating reading materials.

Limitations: Text-focused only—doesn't help with other resource types.

What preparation-phase AI cannot do:

Understand your specific pupils' prior knowledge or misconceptions
Anticipate classroom dynamics that will shape how lessons unfold
Replace subject-specific pedagogical knowledge
Make professional judgements about curriculum sequencing

Bottom line: Preparation-phase AI is best understood as an assistant that handles routine tasks, allowing teachers to focus expertise on aspects requiring professional judgement.

Phase 2: Teaching (where teacher expertise is irreplaceable)

Despite considerable hype around "AI teachers," the evidence consistently shows that responsive, expert teaching requires human judgement that AI cannot replicate.

Why teachers cannot be replaced

Effective teaching requires:

Real-time responsiveness – Recognising when confusion signals a fundamental misconception versus momentary uncertainty. Adjusting explanations on the fly.

Relational trust – Learning happens in relationships where pupils feel safe to take risks. Teachers build this through consistency and care that no AI can replicate.

Subject expertise and pedagogical content knowledge – Knowing not just what to teach, but how to teach it, including common misconceptions and how to address them.

Classroom management – Creating conditions where learning can happen through behaviour management, routines, discussions, and balancing challenge with support.

Soft Skill Development – Confidences, communication, empathy

Limited AI support during teaching (use cautiously)

Mentimeter / Kahoot What they do: Live polling and formative assessment with AI analysis of responses.

Why they're useful: Surface data faster than manual counting, helping identify misconceptions across a class.

Key point: Tools don't make instructional decisions—teachers do.

Speech-to-text / Translation tools What they do: Support pupils with SEND or EAL needs during whole-class instruction.

Why they're useful: Ensure access to content whilst teacher focuses on delivery.

Key point: Assistive technology, not instructional replacement.

Guiding principle for Phase 2: The teacher is the expert in the room. AI should only ever support—never replace—professional judgement.

Phase 3: Post-lesson tools (where AI shows real promise)

The post-lesson phase is where AI offers the most pedagogically sound opportunity to support both teachers and pupils. Research from the Education Endowment Foundation shows that high-quality formative assessment and timely feedback deliver eight additional months' progress—yet these are among the most time-intensive activities.

This is where AI designed for post-lesson work becomes genuinely valuable.

Willow Learn

What it does: A personalised and safe AI for students, controlled by teachers. Willow creates post-lesson activities that reinforce what's just been taught, working across both classroom and home environments as an extension of teacher-led instruction.

Why it's different: Willow is designed specifically for teachers to use throughout the teaching process—from in-class post-lesson activities to follow-up homework that reinforces curriculum toward exam goals.

The digital footprint advantage: As pupils interact with Willow, it builds a detailed picture of each student's:

Vocabulary range and natural writing style
Common misconceptions and error patterns
Current working level and readiness for more complex content
Progress toward exam specifications

Why this matters for homework accountability: Because Willow knows each pupil's authentic capability from classroom interactions, homework completed at home can be compared against their established baseline. If a pupil suddenly submits work wildly inconsistent with their digital footprint, it's immediately visible—not through unreliable AI detection software, but through genuine knowledge of that pupil's voice and level.

The AI literacy piece: Willow doesn't give pupils answers—it uses Socratic teaching, probes understanding, and adapts support based on where the pupil genuinely is. This teaches pupils that AI can be a learning scaffold rather than a shortcut.

Best for: Schools wanting integrated post-lesson reinforcement safely across subjects, with insights into genuine pupil understanding and accountability for homework.

Key strength: Works in-class (post-lesson activities) and at home (homework follow-up), creating a continuous thread of learning aligned to your curriculum.

Century Tech

What it does: Uses diagnostic assessments to identify knowledge gaps, then provides personalised learning pathways.

Why it's useful: Strong for GCSE and A-level preparation with close mapping to exam specifications.

Best for: Post-lesson consolidation and structured revision programmes.

Limitations: Requires initial setup time and works best when pupils use it consistently.

Educake

What it does: Curriculum-aligned homework quizzes with automatic marking.

Why it's useful: Reduces marking workload whilst providing regular low-stakes retrieval practice.

Best for: Knowledge recall homework across subjects.

Limitations: Doesn't adapt content in real-time or build a qualitative picture of pupil thinking—focused more on recall than understanding.

Sparx Maths

What it does: Algorithm-based maths homework that adapts difficulty based on performance with instant feedback.

Why it's useful: Substantial time savings on marking; pupils get immediate support when stuck.

Best for: Secondary maths homework at scale.

Limitations: Maths-specific only; doesn't translate to other subjects.

What makes post-lesson AI tools effective?

The best post-lesson tools share four characteristics:

Reduce teacher workload by automating routine assessment and feedback
Provide better diagnostic data about pupil understanding than traditional marking alone
Enable consistency in formative assessment that's difficult to achieve manually at scale
Extend learning beyond the classroom without requiring additional teacher time

When implemented well, these tools don't replace teacher expertise—they amplify it.

Choosing tools strategically: questions to ask

With hundreds of AI tools marketing to schools, here are key questions for each phase:

For preparation-phase tools:

✓ Does this save genuine time, or just shift work around?
✓ Does it align with the National Curriculum and our exam boards?
✓ Can teachers adapt outputs to their specific pupils?
✓ Is the quality assurance process transparent?

For teaching-phase tools:

✓ Does this support teacher decision-making, or attempt to replace it?
✓ Does it enhance teacher-pupil relationships, or diminish them?
✓ Can teachers override AI suggestions in real-time?
✓ Is there evidence this improves outcomes beyond good teaching alone?

For post-lesson tools:

✓ Does it integrate with what was taught in the classroom?
✓ Does it provide actionable insights, not just data?
✓ Can it build a meaningful picture of pupil understanding over time?
✓ Does it teach pupils to use AI responsibly, not just rely on it?

Implementing AI tools school-wide

Individual teachers experimenting with AI can be valuable, but meaningful impact requires whole-school adoption—particularly for post-lesson tools where consistency helps pupils develop good habits.

Four keys to successful implementation:

1. Clear pedagogical rationale Staff need to understand why a tool is being adopted, not just how to use it. What problem does it solve?

2. Adequate training and support Not just technical training—pedagogical training. How does this tool fit into effective lesson sequences? How do teachers use the data it generates?

3. Leadership commitment Tools only deliver value with consistent use and time to demonstrate impact. This requires protected time for training, clear expectations, and ongoing monitoring.

4. Pupil education Pupils need to understand the purpose of AI tools, how to use them effectively, and why shortcuts undermine their learning.

What does the evidence say?

The Education Endowment Foundation's 2024 review of AI in education found promising early indicators but cautioned against over-investment in unproven tools.

What research shows:

✓ Adaptive learning platforms show moderate positive effects for homework and revision, particularly in maths (Educake, Sparx, Century)

✓ AI-assisted formative assessment helps teachers identify misconceptions faster (though impact depends on how teachers use insights)

✓ Automated feedback on writing shows promise when it supplements—not replaces—teacher feedback

✓ Preparation tools show teacher time savings, but no direct evidence yet of improved pupil outcomes

Key takeaway: AI tools are most effective when they support proven pedagogical strategies (retrieval practice, formative assessment, immediate feedback) rather than attempting to replace core teaching.

The bottom line: AI as amplifier, not replacement

The question facing UK schools isn't whether to use AI—it's already here. The question is how to use it strategically.

The three-phase framework offers clarity:

Preparation – AI can reduce workload, but requires professional oversight

Teaching – AI should support, never replace, teacher expertise

Post-lesson – AI offers the greatest opportunity to improve consistency, reduce workload, and deepen insights into pupil learning

Used well, AI tools amplify what good teachers already do. Used poorly, they become expensive distractions that add workload without improving outcomes.

The schools that will benefit most are those that approach AI adoption with pedagogical clarity, invest in proper training, and remain focused on the question that matters: is this helping our pupils learn better?

What AI tools is your school using? What's working, and what isn't? We'd welcome your insights - book a call.