Automation & CRM Blueprint: What to Build for Your Business Model

How this blueprint is put together

It is not a list of everything that could be automated. Every business could automate forty things; the useful question is which three to build this quarter, and that is decided by how the business makes money, where its demand arrives, and what it is running on today.

1. It sorts by how you make money, not by your industry

   A dental clinic and a tattoo studio have the same operational problem — a calendar is the scarce thing and no-shows are the leak — while a dental clinic and a dental laboratory have almost nothing in common. Sorting by sector puts the wrong pairs together, which is why the sixteen models here are cut by how the money arrives and how the work is delivered. If you would rather read it by sector, industries that can grow through automation covers twelve of them in prose.

2. The four questions each move something specific

   Who sells decides how much belongs in the first phase and whether a CRM is worth buying at all. The channel decides which automation comes first and what has to sit underneath it. What you run today decides whether there is work to do before any build starts. How you want to pay decides the platform — Zapier, Make, or self-hosted n8n. Every rule that fires writes a line into “what your answers changed”, so nothing about the result is a black box.

3. The phases are an order, not a wish list

   The first phase is the work that pays for itself on its own, and none of it depends on the rest being finished. The second assumes the first is running and the record it produces is trusted. The third is only worth building once volume makes the manual version genuinely expensive. Most stalled automation projects are a phase-three build attempted first — which processes are worth automating goes through the shapes that tell them apart.

4. It will tell you not to buy a CRM

   Three of the four possible CRM answers are not a purchase. One person selling out of a spreadsheet does not need one yet, and the result says what would have to change first. A business that already has one should keep it, because the cheapest CRM is always the one the team already opens. An online store, a marketplace and a creator already hold the customer somewhere else, and a second copy of that person goes stale within a month. Everything it does recommend is something we implement — the same stack behind our GoHighLevel, Pipedrive and Airtable work.

Where this sits next to the other tools

This one answers “what should we build”. Once you have a candidate, the automation fit assessment asks whether that particular process can actually be handed over and which half of it can run unattended, and the automation ROI calculator puts a number and a payback period on it. Run them in that order: a build with an excellent return can still be the wrong thing to start, and this page is what stops you starting with the wrong one.

What the plan assumes about your systems

Every automation here reads from or writes to something you already run, so the questions worth asking about those systems come before the build, not during it — the integration checklist is the list. And each of them will eventually meet a case its rules never covered, which is a design decision rather than an accident: failure-ready automations covers what should happen to the run that dies at 2am. If most of your enquiries arrive as messages, settle the WhatsApp Business app or Cloud API question before anything else, because the rest of the plan sits on top of that decision.

What this looks like when it is built

The first-phase builds are not hypothetical. Instant first response and lead routing are what this lead response automation does; booking and reminders on a business number are this WhatsApp appointment booking system; and invoicing from completed work is this billing and receipts automation.

Automation and CRM by business model

Why does it ask for a business model instead of an industry?

Because the model predicts the work and the industry mostly does not. A dental clinic, a tattoo studio and a driving school all sell slots in a calendar, so all three lose money the same way — an empty hour that cannot be sold twice — and all three fix it with the same three builds. A dental clinic and a dental laboratory sit in the same sector and have almost nothing in common operationally: one is booking-led, the other is a made-to-order production queue. Sorting by sector groups the wrong pairs together, which is how you end up with generic advice. If you would rather read it by sector, the field guide on industries that can grow through automation covers twelve of them.

What if my business is not one of the sixteen?

Then the tool says so rather than guessing, and shows you the three closest by how they make money. That is a real answer, not a failure — some operations genuinely combine two models, and a few are their own thing. The common case is a business that runs two models at once: a restaurant that also sells catering, a store that also sells wholesale, a practice that also runs courses. Run the tool twice, once for each half. Those halves usually need different systems, and treating them as one business is often the actual problem.

Do I need a CRM from day one?

Usually not, and this tool will tell you so. Three of its four possible CRM answers are not a purchase. If one person is selling and it currently lives in a spreadsheet, the honest recommendation is a better spreadsheet plus a rule about what a lead record must contain — and the three specific things that would justify buying one later. If you already have a CRM, keep it: the cheapest CRM is the one your team already opens, and a migration costs weeks and loses history. If you run an online store, a marketplace or an audience, the customer record already exists in that platform, and a sales CRM beside it becomes a second, staler copy of the same person within a month.

Why does it only recommend a handful of tools?

Because it only names software we actually implement. There are good products this page never mentions, and that is deliberate rather than an oversight — recommending a tool we could not build on would make the page more comprehensive and less useful. Where the honest answer is a category we do not sell, the result says that plainly instead of substituting something we do. The full list of what we work with is on the services pages, and every tool the blueprint names is drawn from it.

How is the order of the three phases decided?

By dependency and by payback, not by ambition. A first-phase build pays for itself on its own and needs none of the others finished — instant response, one record for every lead, invoicing from completed work. A second-phase build assumes the first is running and that the record it produces is trusted, which is why routing and reporting sit there rather than at the start. A third-phase build only earns its cost once volume makes the manual version genuinely expensive. Most stalled automation projects are a third-phase build attempted first, usually because it was the most interesting one.

What do the four questions actually change?

Each one moves something you can see, and the result lists every change it made. Who sells decides how much belongs in the first phase and whether a CRM is worth buying yet — one person gets a shorter plan, because a plan nobody finishes is worse than a short one. The channel decides which automation comes first and what has to sit underneath it. What you run today decides whether there is work to do before any build starts, and if you already have a CRM it removes the recommendation entirely. How you want to pay decides the platform underneath everything, and at the lowest budget it removes the third phase rather than pretending it fits.

Is this the same as the automation fit assessment?

No, and they answer different halves of the question. This page decides what to build: given how your business makes money, which three automations are worth the quarter. The fit assessment takes one of those and asks whether it can actually be handed over — whether anyone can state the rule it follows, whether the data sits somewhere software can read, and what it costs when a run goes wrong. The ROI calculator then puts a number and a payback period on it. Run them in that order, because a build with an excellent return can still be the wrong thing to start with.