Introducing Kaijo: AI functions that just work
For months, I have wrestled with a problem that has consumed my thoughts and challenged everything I know about software development.
This week I wrote about building the future with AI agents. One of the key areas for me is moving beyond prompt engineering to something more reliable.
I have spent decades learning how to craft reliable software. Now I want to bring that reliability to AI development.
Today I am ready to share what I have been building in the background.
It started with a game. It ended with something that could change how we build AI applications forever.
The Breaking Point
About a year ago, I began working on a new role playing game for fun (my background is in game development working on two different titles, Ealdorlight and Sol Trader). This time, I wanted to use AI to generate dynamic storylines and characters.
The initial progress was exhilarating. The AI created rich, compelling background stories for characters that I could never have written alone.
Then reality set in.
I found myself spending hours not building game features, but tweaking prompts. Desperately trying to maintain consistency in character responses.
Each small change to the game required hours of prompt engineering. The creative joy of game development gave way to an endless cycle of trial and error. I was not making a game anymore. I was becoming some kind of prompt babysitter.
A Familiar Pattern
All my software development experience screamed at me that something was wrong. We solved these problems in traditional software development decades ago. We have unit tests. Integration tests. Continuous integration. Why were we wrestling with prompts instead of writing tests?
The existing solutions felt incomplete. It feels like we are in the punchcard era of AI development. There are some early solutions: DSPy offers fascinating academic insights but requires significant expertise to implement. LangSmith and other services still leave the hardest parts of prompt engineering to developers. We have made AI accessible, but not truly usable for developers.
What if we could automate the entire prompt optimisation process? What if developers could write simple functions and let the AI handle the complexity of prompt engineering?
With that idea, Kaijo was born.
What Is Kaijo?
Kaijo makes AI development feel like normal software development. Write a function (just an api call). Add some tests (bring some examples, or generate them with AI). Let Kaijo handle the rest.
Behind the scenes, Kaijo continuously evaluates and optimises your AI functions. It figures out from your examples what good looks like, and what the best prompt is to get that result. It can do this using cheaper models or different models seamlessly, and can test your prompts in parallel to find the combination of the cheapest fastest and best model to get what you need.
The result? AI functions that just work.
Kaijo Enables The “12 Factor Agents” Approach
The industry is beginning to recognise that successful AI applications share common principles. Gone are the days when you use a big prompt, a loop and a bag of tools and hope for the best.
One set of guidelines is the “12 Factor Agents” way of building agents. Kaijo plays very nicely with this approach:
-
Natural Language to Tool Calls: Instead of wrestling with raw prompts, Kaijo transforms natural language into structured tool calls, making AI interactions predictable and testable.
-
Structured Outputs: Every AI interaction in Kaijo produces structured, predictable outputs that integrate seamlessly with your existing codebase.
-
State Management: You call Kaijo at any point within your business logic, making AI functions behave like any other part of your application. You manage state, workflow and RAG as before.
These principles ensure that AI development with Kaijo feels familiar to any software developer, while handling the unique challenges of AI applications.
See It In Action
On Friday 2nd May 2pm to 6pm UK time, I will attempt to build an entire AI application live on stream. We are going to be building a cheatsheet generator that creates personalised study guides from any text. Try as I might I haven’t found anything on the internet that does this yet, and using Kaijo, this will be much easier to build.
The stream will demonstrate how Kaijo transforms AI development. You will see how the hardest part of AI development becomes the easiest. No prompt engineering required.
AI Development For The Rest Of Us
Kaijo represents more than just a tool. It represents a future where developers can focus on building applications, not wrestling with prompts. Where AI is just another reliable component in our software stack.
By embracing software engineering principles that have stood the test of time and adapting them for the AI era, we are creating a foundation for the next generation of AI applications.
Early access to Kaijo opens very soon. You can sign up for the waitlist at kaijo.ai, and join my newsletter below for more notes about the journey.
The future of AI development should not belong to AI experts. It should belong to regular developers who want to build amazing things. Let us make that future together.
More articles
Building AI Cheatsheet Generator Live: Lessons from a Four-Hour Stream
I built an entire AI-powered app live, in front of an audience, in just four hours. Did I finish it? Not quite. Did I learn a huge amount? Absolutely. Here is what happened, what I learned, and why I will do it again.
The challenge was simple: could I build and launch a working AI cheatsheet generator, live on stream, using AI first coding and Kaijo1 as my main tool?
Answer: almost! By the end of the session, the app could create editable AI cheatsheets, but it was not yet deployed. A few minutes of post-stream fixes later, it was live for everyone to try. (Next time, I will check deployment on every commit!)
Try the app here: aicheatsheetgenerator.com
AI: The New Dawn of Software Craft
AI is not the death knell for the software crafting movement. With the right architectural constraints, it might just be the catalyst for its rebirth.
The idea that AI could enable a new era of software quality and pride in craft is not as far-fetched as it sounds. I have seen the debate shift from fear of replacement to excitement about new possibilities. The industry is at a crossroads, and the choices we make now will define the next generation of software.
But there is a real danger: most AI coding assistants today do not embody the best practices of our craft. They generate code at speed, but almost never write tests unless explicitly told to. This is not a minor oversight. It is a fundamental flaw that risks undermining the very quality and maintainability we seek. If we do not demand better, we risk letting AI amplify our worst habits rather than our best.
This is the moment to ask whether AI will force us to rediscover what software crafting1 truly means in the AI age.
-
I use the term “software craft” to refer to the software craftsmanship movement that emerged from the Agile Manifesto and was formalised in the Software Craftsmanship Manifesto of 2009. The movement emphasises well-crafted software, steady value delivery, professional community, and productive partnerships. I prefer the terms “crafting” and “craft” to avoid gender assumptions. ↩
Why Graph RAG is the Future
Standard RAG is like reading a book one sentence at a time, out of order. We need something new.
When you read a book, you do not jump randomly between paragraphs, hoping to piece together the story. Yet that is exactly what traditional Retrieval-Augmented Generation (RAG) systems do with your data. This approach is fundamentally broken if you care about real understanding.
Most RAG systems take your documents and chop them into tiny, isolated chunks. Each chunk lives in its own bubble. When you ask a question, the system retrieves a handful of these fragments and expects the AI to make sense of them. The result is a disconnected, context-poor answer that often misses the bigger picture.
This is like trying to understand a novel by reading a few random sentences from different chapters. You might get a sense of the topic, but you will never grasp the full story or the relationships between ideas.
Real understanding requires more than just finding relevant information. It demands context and the ability to see how pieces of knowledge relate to each other. This is where standard RAG falls short. It treats knowledge as a stack of random pages, not as a coherent whole.
Time for a totally new approach.
Read moreBuilding the Future
Something has been on my mind for months. The rapid evolution of AI agents has opened up possibilities I cannot ignore.
We are witnessing the emergence of semi autonomous agents that will fundamentally reshape how we work and communicate. The opportunities in this space are extraordinary. I am diving deeper into this world of AI agent development and product creation.
My newsletter is evolving. Instead of dispensing tips from a position of authority, I invite you on a journey of discovery. I will document my experiences building with AI, how to apply my tech experience in a new world, and navigating the inevitable struggles and setbacks.
Read on for several key areas I am exploring.
Read moreThe Reality of AI Power Usage
AI power usage generates significant controversy. Headlines paint it as an environmental catastrophe waiting to happen. The reality proves more nuanced and potentially more optimistic than these dire warnings suggest.
A ChatGPT query uses 10 times more energy than a Google search. This sounds alarming until one realises it equates to running your hairdryer for six seconds. The entire data centre industry, including all AI operations, accounts for just 1.5% of global electricity consumption.
Here is a rundown of the more pressing issues with AI power usage.
Read more