Not a library
Nothing to install. No dependencies. No vendor lock-in. Works in any language, on any stack, with any AI coding agent — Claude Code, Cursor, Copilot, your custom build.
The Methodology · $99
I’ve consulted on AI projects from $5K to $2 Million.They all start here.
THE METHODOLOGY
The AI Foundation Framework
PUSHBUTTON AI
Old code. New code. Greenfield. Mature. It doesn’t matter. Before we write a single feature, draft a single scope, or have a single architecture conversation, we install this. It is the foundation under every AI project that actually ships.
Less than 1% of developers walk into a project with this foundation in place. After 25 years of consulting, I’ve put it in writing — every prompt, every script, every template. Drop the framework into your AI agent and it installs the foundation itself.
$99Drop-in install · ~10 min · Instant download
02 · Diagnosis
Your AI projects are failing for one reason. It isn’t what you think.
- Six months in, the project is “almost there.” Three months after that, it’s still “almost there.”
- The team is shipping code, but the same kinds of bugs keep coming back.
- The AI agents are doing roughly what was asked — but every output drifts a little, and the drift accumulates.
- New developers spend their first three weeks re-explaining the project to the AI from scratch.
- Decisions get made, then quietly relitigated next month.
- Velocity feels good in week one and terrible by month four.
Most CEOs are told this is a model problem. “We’re waiting on the next generation of agents.” Or a skill problem. “We need someone better at prompting.” Or a tooling problem. “We’re switching frameworks.”
It is none of those things. It is a foundation problem.
When the foundation isn’t there, you get drift. Hallucinations. Errors that compound. Agents that confidently produce code that has to be torn out and rewritten. You get the horror stories.
When the foundation is there, you get builds and scopes that move at lightning speed, and the AI gets better at your project the longer you work on it.
03 · Why nobody talks about this
Less than 1% of developers know about this.
I’m not exaggerating. In 25 years of consulting, I’ve worked with hundreds of developers — junior and senior, in-house and contractor, brilliant and competent. The number who walk into a project with this foundation already in place is fewer than one in a hundred.
Your team isn’t behind. The methodology I’m describing is genuinely new — formalized in academic research published this year out of the University of Edinburgh — and the practitioner adaptation that makes it work for production codebases hasn’t been written down anywhere as a complete system.
Your developers aren’t lazy. They aren’t bad. They aren’t behind the curve.
This is why you keep hearing the horror stories. AI projects that ate $200K and shipped nothing. Agents that confidently produced code that had to be torn out and rewritten. Three-person teams that should have shipped in two months and are now eight months in with no end in sight.
Every one of those projects had the same root cause. The foundation wasn’t there. Nobody knew to put it there. Nobody knew it existed.
04 · What this actually is
Four things it is not.
I’m being deliberately vague about the mechanism on this page. Inside the framework, I name it, document it, and hand you the exact prompts and scripts your AI agent uses to install it. Here’s what I will tell you:
Not software
It’s a structural protocol — a way of organizing what your AI agents read, see, and remember. Implemented correctly, it transforms how every AI session your team runs.
Not a SaaS
Drop it into your AI agent once per project — about 10 minutes, the agent does the work. After that, the foundation is in place permanently and every AI session compounds against it.
It is the single most important thing
your team can do before scoping, building, or hiring for any AI work. That’s how I run my own company. Hundreds of codebases, all running on this foundation.
05 · The paradox
This is designed to slow you down.
While the rest of your industry is racing to ship features as fast as possible, this framework is going to make you stop for 10 minutes — just long enough to drop it into your AI agent — before you build anything else. That feels wrong. That feels like the opposite of what you should be doing in a market where speed matters.
It is exactly what you should be doing.
Skip the foundation
Every AI session your team runs starts from zero. Pasted context, drifting output, accumulating debt. Eventually someone proposes a tech debt sprint to clean it up. Three weeks later, you’re back to shipping features and the cycle repeats.
With the foundation
Every AI session starts with the agent reading the project’s foundational files in under five seconds. It already knows what the project is, what conventions apply, what mistakes have been avoided, what’s being worked on right now. The session goes straight to productive work.
Ten minutes of foundation work prevents months of accumulating debt — and unlocks lightning-speed execution forever after.
The teams that install this in early 2026 will have a permanent structural advantage over teams that install it in late 2027. The advantage compounds quietly, in the form of code that fits, decisions that stay decided, and AI tooling that gets better at your projects over time instead of worse.
06 · What you get
Drop it in. Your AI agent installs in 10 minutes.
Two short tracks for you to read — one for the person signing the checks, one for the person writing the code. Drop the prompts, scripts, and templates into your AI agent and it self-organizes and installs the foundation for you.
The Orientation PDF
The first thing you open. How to use the framework, in what order, with what rules.
The CEO Track
Why this matters at the business level. The methodology in plain English. The five diagnostic questions to ask your developers — and what their answers tell you.
The Coder Track
Full methodology. Audit phase. Apply phase. The session protocol that keeps the foundation current, and the hub-and-spoke pattern I use to scale across hundreds of codebases.
The Prompts
Drop these into your AI agent. Phase 1 audit scans any project. Phase 2 apply installs the foundation. Session starter + checkpoint commands keep it current.
The Scripts
A session wrapper that logs every AI session for later debugging or audit. Agent-agnostic.
The Templates
Every new feature, decision, and gotcha gets shaped the same way every time, so the foundation grows consistently across every project.
Drop it into your AI agent. 10 minutes to install on your most active project. Cost: $99.
Get the Framework $99Instant download · License covers your entire team
07 · Run the math
What this costs to not have.
Run the math against your last AI project.
- How many hours did your senior engineers spend re-explaining the project to AI agents?
- How many bugs got reintroduced because the AI didn’t know they’d already been fixed?
- How much code review time went to pattern drift instead of real review?
- How many decisions got relitigated three months after they were settled?
- How long did it take to onboard the last new developer?
- How much did your last contractor handoff actually cost in synchronous walkthrough time?
Total it up honestly. The number lands in five or six figures — not three. Pure waste, repeating every quarter.
The framework is $99. You earn it back on your team’s next AI session — and every session after that compounds in your favor.
08 · Two tracks
Who this is for.
The framework has two tracks. One for the person signing the checks. One for the person writing the code. Whichever side of the project you sit on, this is built for you.
For CEOs, owners, presidents.
- You manage a software team that uses AI coding tools — Claude Code, Cursor, Copilot, anything similar.
- You are about to start, or are in the middle of, an AI-driven product or feature initiative.
- You have hired or are about to hire a contractor or vendor to build something with AI.
- You manage more than one project and want consistency across them.
- You sit on the board of a company building AI products and aren’t sure whether the foundation is in place.
- You have heard yourself or your team say: “the AI is being weird,” “the agent forgot what we decided,” or “we keep re-introducing the same bug.”
For coders, developers, AI agents.
- You use Claude Code, Cursor, Copilot, Aider, or any AI coding agent every day.
- You’re tired of re-explaining the project to the AI on every session.
- You’ve felt the drift — bugs reintroduced, decisions relitigated, context lost between sessions.
- You’ve tried rules files or agents.md and want a complete system instead of half-measures.
- You manage multiple repos and want a consistent foundation across all of them.
- You want a session protocol your team will actually follow — not another doc in /docs.
09 · After you buy
You see the difference on your next AI session.
- Minute 1
You drop the framework into your AI agent — Claude Code, Cursor, Copilot, anything similar.
- Minute 10
Your AI agent has self-organized, audited the project, and installed the foundation. You commit to Git.
- Next session
Your AI agent reads the foundation in five seconds and goes straight to productive work — no re-explaining, no drift, no relitigating decisions from last month.
By the end of today, you’ll have a working foundation, a session protocol, and the start of a quiet but compounding advantage.
10 · What buyers say
Drop it in. Tell us what changed.
Fifteen developers, CTOs, and technical founders on the first thing they noticed after installing the framework into their AI tool of choice.
“Dropped the framework into Cursor and the hallucinations stopped within an hour. It's the difference between an agent that guesses and an agent that asks.”
Marcus K.
Staff Engineer, Series B SaaS
“We were two sprints from shipping when I installed this. It immediately flagged three architectural decisions our team had been letting the AI slide on for weeks. Saved us a rewrite.”
Priya N.
CTO, Healthtech Startup
“I onboarded a mid-level dev to our AI-assisted workflow in a single afternoon using the framework. Used to take two weeks of pairing.”
Daniel R.
Engineering Manager, B2B SaaS
“Claude Code was making changes that broke our test suite weekly. After installing the framework, the agent reads the context first — like it suddenly knows the codebase.”
Aiden T.
Indie SaaS Founder
“$99 saved me a $40K rebuild. The 'before we write a single feature' part is real — we caught two scope creeps in the foundation phase that would've blown up by week six.”
Elena V.
Solo Technical Founder
“My juniors are now allowed to use Cursor in production without supervision. The framework forces the agent to flag risky changes before it commits them.”
Wesley B.
CTO, Fintech Series A
“Used it on a 12-year-old Rails app we're modernizing. The agent stopped trying to rewrite everything in Next.js and started suggesting incremental refactors. That alone was worth ten times the price.”
Sanjana R.
Staff Engineer, Established SaaS
“We didn't realize until week three that the framework was generating documentation as a byproduct of how it structures the AI's work. Our docs site got a year's worth of updates in a month.”
Quentin M.
Tech Lead, DevTool Startup
“I run six client projects in parallel. This is the only thing that's let me let the AI write production code without me babysitting every diff. ROI in two days.”
Felix O.
Independent Software Consultant
“Replaced three weeks of architecture review meetings with one prompt cycle. The framework forces the AI to surface the decisions our team needs to make — before code gets written.”
Iris L.
VP Engineering, Mid-size SaaS
“I've been reviewing AI-generated PRs for a year. Half of what I caught is now caught by the framework's checks before the PR even opens. My review time is down 70%.”
Theo P.
Senior Engineer, B2C SaaS
“Our stack is Python / Postgres / FastAPI / React. The framework picked up our conventions in the first install pass and started enforcing them. No more agents 'helpfully' importing Lodash into Python.”
Anya S.
ML Engineer, AI-native Startup
“Shipping velocity up ~2.5x. Bug rate down. Used to be: one of those, never both. The framework is the difference.”
Ravi G.
CTO, eCommerce SaaS
“I'm not a dev, but the framework's prompt structure forced our AI tools to surface the trade-offs in plain English. I can finally make product calls without an engineer translating for me.”
Maya H.
Technical PM, B2B SaaS
“We'd stalled for two months on a feature we couldn't get the AI to build cleanly. Installed the framework on a Friday. Shipped it by Wednesday.”
Cole D.
Founder, Vertical SaaS
11 · The pricing
Why I’m selling this for $99.
Because the alternative is that I write a 30-page proposal, schedule three discovery calls, charge you $15,000 to come implement it personally, and we both feel like we’ve done something serious.
I’ve done that engagement many, many times. The work is the same every time. The result is the same every time. The only thing that varies is which client is paying for the same 10 minutes of foundation work.
You don’t need me to do it for you. You need to know it exists, understand why it matters, and have the prompts and scripts ready to run it yourself.
That’s what this framework is.
If you’d rather hire me directly to come install this in your organization personally, I’m available. The rate is $15,000. The result is the same. The choice is yours.
12 · Frequently asked
Before you buy
Will this work with my AI tool?
Yes. The framework is filesystem-based and tool-agnostic. The examples use Claude Code, but every prompt and pattern adapts cleanly to Cursor, Aider, Continue, Copilot, or any custom agent that can read files in a repository.
How long does the actual setup take?
About 10 minutes per project. You drop the framework into your AI agent and the agent self-organizes the prompts, scripts, and templates into your repo. You don’t do the install — your agent does.
Do I need to be technical to use this?
The CEO track is written for non-technical readers. You can read and understand the entire methodology without writing code. The hands-on install is done by your AI agent — your team just hands it the framework.
Why won’t you tell me what the methodology actually is on this page?
Because the value of having this foundation is partly that your competitors don’t have it. I’m not going to broadcast the framework for free on a public page. Inside, I name it, document it completely, and hand you the working prompts. The price of admission is $99.
What if my team pushes back on it?
Common. The Coder track addresses every objection developers typically raise. The most effective response is to ask them to run the Phase 1 audit on a single project. The audit is read-only and takes five minutes. The report itself usually changes their minds.
Can I share this with my team?
The framework is licensed for use within your organization. The prompts, scripts, and templates are yours to use, modify, and adapt for any commercial purpose. Redistribution of the framework itself is not permitted.
What’s the refund policy?
If you read the orientation PDF and the first file of either track and don’t think this is worth $99, email me and I’ll refund you. I’d rather you not buy something you won’t use than have you sitting on a framework that isn’t right for you.
THE METHODOLOGY
The AI Foundation Framework
PUSHBUTTON AI · 2026
Buy the framework. Drop it in. See the difference in 10 minutes.
The teams who install this in 2026 will be eating the lunch of the teams who don’t, for as long as AI matters in software development.
Get the Framework $99Instant download.