This is the keep-it-nearby version of the playbook. Each stage is one decision: the question you're answering, the criteria from the article, the options to weigh, and what disqualifies each one. Open the document when you're facing a choice of direction, segment, or billing model — and don't want to guess.
The guiding rule in one sentence: choose deliberately and stick with it, because each of these paths demands completely different work.
Three terms recur throughout. A retainer is a fixed monthly fee for ongoing work, not a one-off project. A niche is a narrowly defined audience or problem type the business focuses on. A unique mechanism is your own, recognizable way of delivering the service.
Decision 1: lifestyle business or a company to sell
The question: are you building a comfortable income stream, or a company you could eventually sell?
This is the distinction most people in the field skip, and it decides everything else. Both paths are fine, but they demand different work, so pick one and don't jump between ideas every week.
| Option | What it means in practice | What it takes |
|---|---|---|
| "Lifestyle" business | A handful of clients, automations worth a few thousand each, comfortable income | Little structure; execution speed |
| Company with real value | A company you could sell to someone else | A repeatable process, a pipeline, people, a valuation |
What disqualifies it: jumping every week between "lifestyle", a product, and a large company to sell. Indecision costs you, because you end up dragging both paths along half-finished.
Decision 2: what to build value on
The question: what part of your offer survives once software itself keeps getting cheaper to build?
Starting thesis: most of what sells today as "AI work" won't survive to 2027, because building software keeps getting cheaper. If a 67-year-old lawyer can already stitch together a decent app with AI's help, then the "come to me, I'll build it, and I'll bill you by the hour" model is falling apart.
The conclusion isn't that AI is overhyped; the results can be real and large. In one e-commerce deployment, an AI system took over returns handling and the return rate dropped from 21 to 16 percent; at that company's scale, even one or two percentage points meant millions on the bottom line. That's data from one specific project, a single case, not a rule.
| Criterion | Holds up | Loses meaning |
|---|---|---|
| What it's about | A clear line to the client's financial outcome | Add-ons and "features" bolted onto existing tools |
| Where the edge comes from | Knowledge of the industry and what actually hurts the client | The bare fact that "I know how to code it" |
| What the client is buying | Peace of mind, that they have an AI strategy and are ahead | Another "AI system" with no context |
What disqualifies the option: an offer whose value depends on an existing workflow that's about to collapse on its own. If your edge boils down to "I can code it," that's not an edge; coding keeps getting cheaper.
Decision 3: who to talk to
The question: which market segment has success metrics organized enough that you can actually prove an AI system's result?
The playbook targets the mid-market: companies with roughly $10 million to $250 million in annual revenue. The reason is practical: these companies have already had to write down procedures, hire people, and track metrics, so they have clear numbers to attach a result to.
| Segment | Why it does (or doesn't) fit |
|---|---|
| Smaller businesses and solo founders | More often fund "passion projects" that don't add up to a result |
| Mid-market (roughly $10-250M revenue) | Have written procedures and metrics; results are easier to measure and defend |
| Large corporations | Often less organized than assumed; throw more people at problems, too spread out for a coherent strategy |
What disqualifies the option: no clear success metrics on the client's side. Without them you can't prove the AI system changed anything.
Decision 4: the service ladder and the billing model
The question: which rung of the ladder do you offer a given client, and how do you bill for it?
The heart of the approach is packaging the service into a named, repeatable process, the way large consulting firms do. Instead of "I'll do whatever you want," you offer your own framework. In this playbook that's an "agentic operating system": a structure that recognizes events and routes them into as-predictable-as-possible workflows, bringing in a language model only where it's genuinely needed.
On that foundation you build a ladder of offers you climb step by step, polishing each rung before moving to the next:
| Rung | What it is | Range / model |
|---|---|---|
| AI workshop | Cheaper entry service, an hour or two, to build a shared picture and trust | Introductory service |
| Blueprint (discovery stage) | Paid analysis and plan ending in a concrete document | $15,000-35,000 |
| Custom project | Building the system and handing it to the client's team | Project pricing |
| Technology partnership | A model with a share of the outcome | Outcome-based billing |
Value pricing, not hourly: hourly billing trends toward zero, while outcome-based billing becomes possible, but only if the systems genuinely work.
What disqualifies an outcome-based partnership: the lack of a single, agreed KPI with a direct line to the financial result. A guess isn't enough, and not every company will be ready for this.
Decision 5: whether to aim for a sale
The question: are you building the company to be valued and sold, and do you understand how that valuation works?
Value-based pricing applies here too. Above a certain revenue threshold, roughly $5-6 million a year, the multiple a company is valued at rises noticeably; typical examples show a jump from roughly one times to around five times profit.
| Annual revenue | What the valuation looks like |
|---|---|
| Around $2 million | The company tends to sell for a comparable amount |
| Around $6 million | The valuation can reach a multiple of profit |
The second argument is about the economics. Classic software captured a small slice of a client's budget: the license fee. AI, as a partial replacement for labor, lets you reach for a bigger piece: the part of the budget that used to go to salaries. That's a different economics than traditional software. (The figures illustrate the mechanism and the target, not a guaranteed outcome.)
What disqualifies this path: no appetite for risk, no prior experience, or no genuine curiosity about the subject. Without curiosity, hard work just burns you out. Most attempts to build a company fail; that's an honest caveat, not a promise of quick wealth.
The decision path — one page
- Who do you want to be? Choose: lifestyle business or company to sell. Stick with it, don't jump between ideas.
- What are you building value on? On a clear line to the client's result and knowledge of the industry, not on "I know how to code it" alone.
- Who are you talking to? Mid-market (roughly $10-250M revenue), because it has written metrics you can attach a result to.
- What do you offer and how do you bill? Package the service into a named process; climb the ladder: workshop → blueprint ($15,000-35,000) → project → outcome-based partnership. Bill for outcomes only with a single agreed KPI.
- Are you aiming to sell the company? Keep in mind the roughly $5-6 million revenue threshold where the multiple rises. Enter only with an appetite for risk and genuine curiosity about the subject.
Build with your reputation in mind, and with what you want to be known for, not only the number at the end. Treat this as a frame for assessing your own situation, not a ready-made recipe for everyone.