AI that works the way you do.
In 7 days, turn one real workflow into a Claude skill that carries your process and your judgment. You build it. You own it.
Then you stop rebuilding the same workflows with deliverables like RFPs, SOWs, status reports, QBR decks, and solution designs from scratch, and let Claude do the heavy lifting your way every week.
BYOStack is a methodology for hands-on technical operators in the enterprise: solution architects, sales engineers and technical account managers, technical sales leaders, IT directors and managers who still do real work, and architecture and engineering leads. It is for people who already use Claude or other LLMs for real work (or are committed to making it part of their weekly workflow), think in systems, and are done getting generic output from a model that has never been shown exactly how they work.
Not in these roles? Will this still work for me?
Four Coaches (Claude Projects: Compile, Analyze, Strategize, Execute) walk one real workflow from spoken narration to a built skill. They extract the two things generic prompts never carry: your process, and the judgment calls underneath it that you no longer notice making.
Fair warning: this is a thinking gym, not a prompt pack. You will work. That is why it works.
Who this is for, and who it is NOT for
Most people will not do this work. That is fine. BYOStack was built for the operator who would rather think for an hour than rebuild for a month.
This is for you if
- You run complex customer, technical, or operational workflows for a living (pre-sales, solution design, architecture, ops, or IT leadership that still does individual contributor work). Different role?
- You already use Claude (or another LLM) weekly, and you have a Claude Pro plan or will get one.
- You are willing to document how you actually work, and to iterate Version 1 (V1) of a skill over two to four cycles until it earns its place.
- You can commit real working time to your first workflow this week (Compile alone is a two-to-four-hour session), plus two to three hours of optional reading you can spread across the week.
This is NOT for you if
- You want a library of done-for-you skills you drop in and run. BYOStack is NOT that, and you should NOT buy it.
- You hate documenting, dislike architecting, or resent being asked to explain your reasoning.
- You want something that works perfectly on the first run, with no iteration and no hours invested.
- You are shopping for a prompt pack or a list of tricks.
What you actually get
- The four Coaches - Compile, Analyze, Strategize, and Execute: four Claude Projects, one per stage, that interview you, push back on thin answers, and hand you each stage’s files as downloadable Markdown (md) files to carry forward until a skill exists.
- The 7-Day Quickstart - the map from zip file to a working skill: setup in about 15 minutes, your first Compile run on Day 1, Version 1 (V1) by Day 7. You do not have to read the long-form guides first.
- A complete worked example - a weekly status report workflow run through all four stages, start to finish, so you see what a good run looks like.
- The reference shelf - the Methodology Guide, the Decomposition Primer, the Onboarding Guide, and Installing Your Version 1 (V1): the full depth, there when you want it, never a gate.
- The outcome - Version 1 (V1) of a real Claude skill you actually use, built on your own workflow and your own judgment. Not a library of generic prompts. After two to four iteration cycles on your saved real inputs, it is doing your work your way.
How it works
The spine is CASE: Compile, Analyze, Strategize, Execute. Compile captures one real workflow with deliverables (an RFP response, an SOW draft, a weekly status report, a vendor evaluation, an incident write-up) the way it actually happens: you narrate it out loud, then a second pass extracts the judgment underneath it: what gets extra scrutiny, what gets escalated, what you check before you trust the output. Analyze computes the leverage from what you captured and shows its arithmetic. Strategize turns it into a build plan. Execute builds the skill in Claude and guides the iteration.
Compile, the first stage, is a two-to-four-hour working session (not necessarily in one sitting), because that is where your judgment gets captured and every shortcut there becomes a defect in the skill: measure twice, cut once, or pay the iteration tax every week thereafter. Do it right once, up front, and the payoff compounds on every run after.
Before anything gets built, the workflow has to pass the Four Lenses: forcing function, time savings, cost savings (money or energy), revenue generation. Three of four or you do not build it. This single gate is why BYOStack produces skills you use instead of a folder of abandoned experiments. And if you already build with Claude’s skill-creator: BYOStack is the architect upstream of that builder. It decides which skills deserve to exist and what they must carry.
That is the whole system at a glance. The full theory (CASE in depth, the Four Lenses, The Outcome Statement, the Architecture) lives in the Methodology Guide inside the bundle, and there is a section for the method nerds (like me) at the bottom of this page.
The honest time cost
- Your first workflow - real working time spread over the 7-Day Quickstart, with Compile the longest leg. The first Compile session alone produces a documented Standard Operating Procedure (SOP) of your own workflow, worth having on its own.
- The optional reading - two to three hours, spread across the week as you want it. It deepens the runs; it does not gate the first one.
- After that - the setup is done, the method is familiar, and the universal skills you extract compound, so later workflows move faster.
The alternative is NOT zero hours. It is the ongoing iteration tax of AI that never quite works the way you do: the corrections, made by you, every time it runs. That is the actual trade.
Price, requirements, and the guarantee
The price is $97 USD, one time. Not a subscription. Nothing expires. The methodology is yours to keep, and so is every skill you build with it.
One requirement: the Coaches run in Claude, and their sessions are long and deliberate by design, so you need a Claude Pro plan or higher (currently around 20 US dollars a month, paid to Anthropic, not to me). The free plan cuts a Compile session short in the middle, and a half-captured workflow is worse than none.
14-day money-back guarantee, if the method fails you. This is a do-the-work guarantee. Follow the 7-Day Quickstart: set up the four Coaches and run one real workflow through all four stages. If a workflow honestly fails the lens gate, that still counts; pick another and keep going. If you do that and still do not have a skill you actually use, email support@byostack.ai within 14 days of purchase with screenshots showing the four Coach Projects and at least one run inside each, and I will refund you. Start the Quickstart in your first week so your run fits inside the window. If you never ran the method, the guarantee does not apply: I am guaranteeing the methodology, not the purchase of good intentions.
Not sure yet? The Decomposition Primer is free; it is the first chapter of the same thinking, and the exercise inside it stands on its own.
If this belongs in your organization
BYOStack is a personal-use product. If, after using it, your read is "we need this across the organization" (25 or more seats, or a larger AI workflow problem), that is a separate conversation: BYOStack Scale. No pitch, just a real conversation.
Questions, answered
I do not use Claude at work. I am on Copilot, ChatGPT, Gemini, Grok, or something else. Will this still work for me?
Three honest cases. If you are willing to run Claude Pro personally (around 20 US dollars a month, paid to Anthropic), BYOStack works exactly as designed: the four Coaches are Claude Projects, and everything on this page applies to you as written. If your company standard is another tool, Claude works as your personal thinking lab: the methodology (decomposition, the Four Lenses, judgment extraction) is not bound to any model, so you design and iterate the skill in Claude, then port the documented process, the judgment rules, and the output standard into whatever your company runs. If you will not use Claude at all, BYOStack is probably not the right product for you yet: start with the free Decomposition Primer (the exercise stands on its own with any model), and submit a feature request for your platform so the demand gets counted.
One plain observation to close it: run the method once on a single painful weekly workflow, and the returned hours cover $97 USD without much trying. That is the whole bet.
Will there be versions for Copilot, ChatGPT, Gemini, Grok, or other platforms?
Today, BYOStack is built for Claude: that is where the Coaches live and where the long, deliberate sessions run. Other platform versions are possible if demand supports them: a possibility worth naming, not a promise. If you want a version for your platform, submit a feature request. Every request is read and counted, and a platform build gets prioritized when the volume justifies the certification work each one requires.
What if what I do is not listed on this page? Will this still work for me?
The role list is a signal, not a gate. The method keys on behavior: you run recurring, complex knowledge workflows (reports, reviews, summaries, decisions with judgment inside them, etc.), you can articulate what good output looks like, and you are willing to document how you actually work. If that is you, the mechanics are identical whatever the title says: internal strategy decks, incident summaries, vendor evaluations, board updates, and more. And the honest inverse: if you hate writing, dislike thinking in systems, and will not document anything, you will not enjoy a minute of this. Take the free Primer instead; it costs nothing and the exercise alone is useful.
Still undecided? Get the free Decomposition Primer.
For the method nerds (like me)
This section is optional. The method works whether or not you ever read the theory; it is here for the operators who want to know what is underneath, before or after they buy.
BYOStack applies CASE and the Four Lenses from the Certainty Operating System to AI skill-building, adds a first-principles pass that asks whether a workflow should exist at all before automating it, and opens every build with an Outcome Statement. This step is adapted from Dan Nicholson’s Solvable Problem™, covered in Rigging the Game and The Solvable Problem™ Book. The adaptation is called The Outcome Statement: it turns a vague ambition into a problem statement with defined parameters that can actually be solved for. BYOStack implements that idea for skill-building, converting a vague workflow ambition into a surgical build target with measurable results. The full argument lives in the Methodology Guide inside the bundle, and the Decomposition Primer, the first chapter of the same thinking, is free.
The architecture: four Coaches run the four stages of CASE, every stage gated by the Four Lenses, with Universal Skills and the Deployment Decision Tree supporting the build. The output is Version 1 (V1) of a Claude skill you iterate.
About Jeff Merck
Jeff Merck is a Certified Certainty Advisor, executive professional, and serial entrepreneur with multiple decades across enterprise sales, operations, technology, and delivery. He solves complex problems using his own methodology, and BYOStack is that methodology made teachable: the same thinking he brings into organizations as a Strategic Advisor. He runs his own daily business operations on skills built with this same method. He is driven by a mission to help others succeed and make a global positive impact through his collective works. When he is not working or pondering problem-solving, he spends his time with his family.
If you have read this far and it fits, you are who this was built for. Start with one workflow that matters.
Take one workflow apart. Find where the leverage actually lives.
The Decomposition Primer is one of the guides inside BYOStack. It is free, in full, delivered to your inbox. It ends with a 15-to-20-minute exercise you run on one of your own workflows; finish it and you are holding a decomposed mapping of your own work, judgment calls named, leverage point found. That mapping is useful whether or not you ever buy anything.
You will get a confirmation email first; the Primer follows the moment you confirm.
No tracking of any kind: no open pixels, no click tracking, no cookies. Unsubscribe works instantly. The details are in the Privacy Policy.