Build something bigger than the first working demo.
The first version is exciting. Keeping it understandable, safe to change and useful to the people who depend on it is the next challenge. Learn the method in Engineering With AI, put it to work with the coding harness, and bring in specialist personas when you need a different perspective.
Install the harness for free. Buy books and premium personas online.
The method and harness helped us build Backstory, a platform with over 1.3 million lines of code.
The first version works. Now you need to keep building.
You return to a project and spend half the session explaining decisions you’ve already made. A small change breaks something elsewhere. The code looks convincing, but you’re not sure what’s actually been checked.
We’ve been there. Building Backstory taught us that AI needs more than a good prompt: it needs the project’s context, a clear piece of work and a way to check the result.
That’s what we’re sharing here. The book explains the approach. The harness makes it part of your project. The personas bring other people’s perspectives into your decisions.
Start with what you need today.
Want to understand the method, try it in a project, or get a second perspective on a difficult decision? Start here.
Your project shouldn’t start from scratch in every AI session.
The Engineering With AI harness gives people and AI the same project context: what you’re building, what mustn’t change, what you’ve decided and what the tests actually prove. It supports existing systems too, including projects whose documentation has gone missing.
Explore the harness and its capabilitiesMove fast without making a mess.
Engineering With AI explains how to turn an idea into software you can review, test and maintain. Learn how to give AI useful context, decide when a person needs to step in, and check more than whether the code runs.
Read a sample from the bookWho would spot the thing you’ve missed?
Ask a security persona to challenge a design, an accessibility persona to examine a journey, or a finance persona to question a business case. Explore 294 personas across 33 collections, then choose the perspectives your work needs. You still make the decision.
Meet the personas and see annual access
Read the method behind the work.
They aren’t three versions of the same AI book. They deal with three different responsibilities that increasingly meet in the same piece of work. Engineering With AI is available now; Thinking With AI and Leading IT With AI are due later in 2026.
Partnered Intelligence
Frame the real question, share the context, invite challenge and use the right reasoning mode before you commit to an answer.
Conversational Coding
Turn intent into a reviewable build, keep engineering standards close, and make evidence part of delivery rather than an afterthought.
Unified Workforce
Give people and AI clear responsibilities, manage the hand-offs, and use the capacity you gain to spend more time leading humans well.
Work out what “right” looks like before the code arrives.
“Add an export button” sounds simple. Which records can a customer download? What stays private? The harness helps you work through those questions and keep the answers with the project, so the agent and the reviewer know what they’re working towards.
- Define the outcome and constraints before AI generates code.
- Reconstruct missing project knowledge before changing inherited systems.
- Keep human Ready and Done decisions explicit.
- Retain tests, review evidence and reusable learning beside the work.
Bring in the specialist you wish was in the room.
These are detailed AI personas with responsibilities, methods, questions and boundaries. Think of them as candidate profiles for a digital team: transparent about what they can contribute, and equally clear about what still needs human judgement.
Code Review Specialist
Call them in when reviews are inconsistent, the queue is stuck, or feedback is technically right but landing badly with the team.
View candidate profile →Grumpy User
Call them in before launch to find the needless steps, confusing language and small frustrations your polite reviewers keep overlooking.
View candidate profile →CFO Perspective
Call them in when the proposal sounds exciting but nobody has tested the cash, assumptions, opportunity cost or route to a return.
View candidate profile →Difficult Conversation Facilitator
Call them in when the message matters, the relationship matters, and saying the wrong thing could make a hard situation worse.
View candidate profile →294 personas. 33 collections. One complete pack.
Browse every specialist and collection before you buy. Individual personas aren’t sold separately, so every profile leads back to the complete provider-neutral Premium Persona Pack.
We built these tools to help us build Backstory.
Building Backstory meant making AI useful beyond the first working demo. We needed it to understand the project, work within clear boundaries and produce changes we could review and trust.
We’re sharing what we’ve learned through our books, the specialist perspectives in our personas, and the engineering harness we use to put the method into practice.
This isn’t a promise that AI makes expertise unnecessary. It’s a practical way to give you and your team more support with difficult work.
Start with the work that’s in front of you.
Read the method, put it into an engineering project, or bring a specialist perspective into the conversation. You can begin with one and find the others when you need them.