AI products and workflows

Context Engineering: The Next Evolution Beyond Prompt Engineering

by

Part of: Sovereign AI for Education Systems: A Practical Guide for Ministries, a guide to the decisions, infrastructure options and roadmap for ministries.

If you’ve been using AI tools like ChatGPT or Claude, you’ve probably experienced this frustration: every time you start a new conversation, you have to re-explain who you are, what you need, and the context for your work. It’s like introducing yourself to the same colleague every single day.

This is the fundamental limitation of prompt engineering – and it’s why the field is rapidly evolving toward something more powerful: context engineering.

The Problem with Prompt Engineering Alone

Prompt engineering has been the dominant approach to working with AI: craft the perfect instructions, get better results. But here’s the challenge:

Every conversation starts from zero. You must re-specify your context, preferences, constraints, and requirements with every single prompt.

If you’re a teacher in South Africa working with CAPS curriculum, this might look like:

  • “You are a South African teacher”
  • “Use CAPS curriculum standards”
  • “Create materials for Grade 7”
  • “Consider multilingual classrooms”
  • “Use South African examples and context”
  • “Align with DBE guidelines”

Research shows that 80% of prompt engineering effort goes into re-establishing basic context, not describing the actual task. This is inefficient, inconsistent, and doesn’t scale.

Read more

The Complete Guide to n8n Workflow Automation for Education and Training

by

This is the hub for n8n in education and training: what it is, where it fits, and the main use cases. The pipeline write-up is automating online learning creation. The accuracy write-up is RAG in n8n.

Last reviewed: 31 August 2026.

The education and training landscape is transforming. From K-12 schools managing thousands of student communications to corporate L&D departments coordinating complex learning pathways, to business schools delivering world-class MBA programs, organizations face an overwhelming operational challenge: managing increasingly complex workflows across disconnected technology systems while maintaining focus on what matters most – effective learning and development.

Enter n8n, an open-source workflow automation platform that’s quietly revolutionizing how educational institutions operate. Unlike expensive enterprise solutions that lock organizations into rigid ecosystems, n8n offers a flexible, cost-effective approach to automation that adapts to your unique educational context.

This comprehensive guide explores how education and training organizations are using n8n to automate repetitive tasks, integrate disparate systems, enhance content quality, and create seamless learner experiences – all while working within tight budget constraints.

Read more

Implementing RAG in n8n for Educational Content Workflows

by

This page is the RAG and vector-database sibling: how to ground generated educational content in your own curriculum sources. Start with the n8n hub if you need the wider picture.

Last reviewed: 31 August 2026.

Read more

Fundamentals of Prompting

by

Lesson 1, Prompt Engineering Module

In this 30-minute session, we explored the foundations of effective prompt engineering using our 10/10/10 format: instruction, demonstration, and practice. The core focus was the CRAFT framework, a systematic approach to writing prompts that consistently generate high-quality AI outputs. We also examined common pitfalls that prevent people from getting the results they need, and introduced practical techniques for immediate improvement.

This session is part of our “Your Journey from AI Curious to AI Confident” programme, designed to equip Cambridge staff with practical AI skills they can apply immediately in their daily work.

What is the CRAFT framework?

CRAFT is a five-component framework that transforms vague requests into precise, effective prompts. Each letter represents a critical element:

Context provides the background and situation. Rather than asking “Write an email,” you might specify “I need to inform our department about a policy change affecting annual leave.”

Role defines who the AI should act as. This might be “Act as an experienced HR communications specialist” or “Respond as a technical documentation writer.”

Action states the specific task to complete. Be explicit: “Draft a 200-word email announcement” rather than “Help me with this.”

Format specifies the structure of the output you need. Do you want bullet points, a formal letter, a table, or JSON data? State this clearly.

Tone sets the voice and style. Should the output be formal, conversational, technical, or empathetic? This prevents mismatches between your needs and the AI’s response.

When you combine these five elements, you create prompts that consistently deliver relevant, appropriately structured results. The framework is flexible—not every prompt requires all five elements, but considering each one significantly improves your outcomes.

When the output is a learning task or a question, name the cognitive demand in the Action. “Write some questions” is ambiguous. “Write five Remember stems, then five Evaluate stems, using this Bloom’s Taxonomy question-stem bank” is an Action you can check.

Read more

Rapid Prototyping AI Products with Claude Projects

by

From Hypothesis to Working Prototype in Days

Developing AI-powered products traditionally requires months of development and significant technical resources. Using Claude Projects, we transformed this timeline from months to days, testing five critical AI capabilities and building a fully functional automated lesson plan generation system for South African teachers. This article explores how hypothesis-driven prototyping with Claude Projects enables rapid validation of AI product concepts before committing to full-scale development.

Introduction

The gap between AI capability and practical implementation remains a significant barrier in educational technology. While large language models demonstrate impressive potential, translating that potential into reliable, curriculum-aligned tools requires answering fundamental questions: Can AI reliably research curriculum documents? Can it follow complex multi-step processes? Can it integrate with existing systems? Can it produce professionally formatted outputs?

Rather than spending months building infrastructure to test these questions, we used Claude Projects as a rapid prototyping environment. Within one week, we moved from concept to a working system that generates comprehensive, curriculum-aligned lesson plans for the South African CAPS curriculum, complete with professional formatting, WordPress integration, and evaluation capabilities.

Why This Matters:

Educational technology development often fails not from lack of ambition but from building solutions before validating core assumptions. Rapid prototyping allows educators and developers to test AI capabilities quickly, identify limitations early, and make informed decisions about scaling production systems.

Read more

Automating WordPress SEO Optimisation with MCP and Claude Code

by

A Case Study in AI-Powered Site Management

This case study demonstrates how AI automation via the Model Context Protocol and Claude Code transformed a labour-intensive WordPress SEO optimisation project into a streamlined, chat-based workflow. What would typically require hours of manual WordPress administration was completed in two sessions without accessing the admin interface, addressing 11 critical errors, optimising 52 pages, and implementing strategic internal linking across 40 posts.

Note – I am not an SEO-bro or marketing guru, but have been running my own websites for years.

Where we started!

Read more

Why I Built an AI Twin to Run My Life

by

Building NiallOS with Claude Code 

“AI will save you so much time!” Silicon Valley promised. So I tried everything. ChatGPT for writing assistance. Claude for research summaries. Gemini for project management. Copilot for planning. Every morning started the same way: Ten minutes explaining my context to whichever AI I was using that day.

“I’m a Product Lead at Cambridge Education Futures…”

“I’m working on a Master’s…”

“I need this in British English for an internal audience…”

“Keep it practical, I’m building training programmes…”

Then I’d finally get to the actual work. AI was supposed to save time. Instead, I felt busier than ever.

I was drowning in AI tool overload while still doing everything manually.

The working laboratory is now NiallOS. The method behind it is expertise encoding.

Last reviewed: 31 August 2026.

The AI Promise vs. Reality

2023-2024: AI will transform knowledge work! Automate everything! Work 4 hours a week!

That was the promise. My reality as a Product Lead at Cambridge looked quite different. I was designing training programmes for education leaders working with organisations like UNICEF and KFAS, and managing a Master’s research project. Creating workshop materials and training curricula. Publishing reports, papers and research outputs. Tracking professional development and learning goals. Staying current on EdTech and AI in education.

AI was supposed to save time on routine tasks, help me focus on strategic work, make complex work manageable, and give me breathing room.

What actually happened? Every morning became an AI context-switching marathon. ChatGPT for brainstorming workshop activities. Claude for analysing research papers. Back to ChatGPT for LinkedIn post drafts. Gemini for curriculum outline suggestions. Each tool required a fresh start and re-explaining everything from scratch.

I calculated the daily context tax. I wasn’t using AI to work smarter. I was managing a portfolio of AI tools, each one requiring setup, context, and translation. It felt like having five junior assistants who all had amnesia every morning.

Claude Code stepping up to run my life

Read more

Transforming Educational Publishing with AI

by

A Practical Workshop Framework

As organisations across higher education and publishing confront the challenge of integrating artificial intelligence into their operations, the question is no longer whether to adopt AI, but how to do so effectively. At Cambridge, we have been running an internal training programme focused on what we term “AI-ifying the organisation” – helping teams understand, adopt, and implement AI tools in ways that enhance rather than replace human expertise.

This article documents our approach to training publishing professionals in AI capabilities, drawing from a comprehensive workshop delivered on AI and Educational Publishing. Rather than treating AI as a mysterious black box, we focus on practical techniques that publishing teams can implement immediately, whilst also addressing the critical change management and upskilling requirements that successful AI adoption demands.

The framework we developed serves dual purposes: demonstrating what is genuinely possible with current AI technology, and providing a replicable training structure that other organisations can adapt to their own contexts. Throughout this article, we share specific examples, prompting techniques, and workflow implementations that have proven effective in real-world publishing scenarios.

Read more

AI Fundamentals for Educators: AI, Language Models, Capabilities and Ethics

by

Part of: How to Design AI Professional Development That Changes Practice, a guide to AI training that lasts beyond the workshop.

This page brings together the four lessons of the AI Foundations module from the Cambridge AI Lunch and Learn series: what AI is, how large language models work, what they can and cannot do, and how to use them responsibly.

Lesson 1, AI Foundations Module

In our first foundational session, we demystified artificial intelligence by exploring what it actually is—and what it isn’t. The key insight? AI doesn’t follow rules like traditional software; it recognizes patterns learned from vast amounts of data, much like how you learned to recognize cats through examples rather than explicit instructions.

We distinguished between different types of AI (Large Language Models, Computer Vision, Predictive AI, and Generative AI) and discovered that AI is already embedded in our daily lives – from email spam filters to smartphone facial recognition. Through practical exercises, participants experienced how AI can adjust its communication style for different audiences and solve multi-constraint problems like planning office pizza orders with dietary restrictions.

The takeaway: AI is a powerful tool for augmentation, not replacement, and like any tool, it requires understanding its capabilities and limitations before we can use it
effectively.

AI_Lunch_9_OctoberDownload

Read more

10 Simple Prompts for Cambridge Teachers to Use with AI Chatbots

by

Ready-to-use prompts for lesson planning, activities and classroom resources

I’ve created 10 easy prompts that give Cambridge teachers immediate access to syllabus and assessment content, suggestions for effective teaching delivery, ideas for engaging classroom activities, and resources you can use in class or share directly with students. The beauty of these AI chatbots is their accessibility; they work on mobile devices and require no special setup or additional data inputs.

I’ve tested these prompts across ChatGPT, Claude and Gemini, and here’s what I’ve found: they’re remarkably similar in terms of capabilities, so use whichever one you are familiar with or have access to. For the demos, I’ve pretended to be an IGCSE Business Studies teacher, but you can adapt the prompts for your subject and level.

The key insight? If you’re a thoughtful teacher who knows how to check and verify outputs, these tools can genuinely increase your productivity and potentially improve your teaching practice. I’ve included all the prompts in the slides so you can test them out yourself, get ideas for lessons, find suggestions for activities, and create content for both your lessons and students. When the model writes questions, check the cognitive level against a stem bank rather than trusting the verb. I keep Bloom’s Taxonomy question stems beside the output.

Read more

Vibe Coding: How South Africans Can Build Software by Simply Describing What They Want

by

An approach to programming that could enhance digital skills development across the country

I’m a big fan of Cursor for coding with AI, and this movement now has a name – “vibe coding” – which allows people to create programmes by describing what they want in natural language and letting AI handle much of the actual coding. As South Africa grapples with developing digital skills and implementing new coding curricula in schools, vibe coding represents a potentially transformative approach that could democratise software development across the country.

The timing couldn’t be better. With the Department of Basic Education rolling out coding and robotics as part of the new curriculum framework, and the country’s ongoing push to develop digital literacy, vibe coding offers a bridge between traditional programming education and practical software creation. Rather than spending months learning syntax and programming languages, South Africans can now focus on problem-solving and creative thinking while AI handles the technical heavy lifting.

Read more

n8n for Educational Content Automation

by

This page is the multi-agent content pipeline: how n8n can run specialised agents in sequence for educational materials. For the wider map, see n8n for education and training. For source-grounded generation, see RAG in n8n.

Last reviewed: 31 August 2026.

Scalable Workflow Implementation

Read more