How it works · start to running

Eight steps, four of which you can stop at.

Custom work makes people nervous for a good reason: it is hard to tell, from the outside, whether a bespoke service has a method or is being made up as it goes. So here is ours, in the order it actually happens, with the points where you decide marked.

The configuration is bespoke. The method is not. Every build runs through the same eight steps, in the same order, whether it is one workflow or three.

You are not committed to anything until step three, and the price is fixed before you are.

Before you commitSteps 1–3
First callFree, 30 min
The audit$500, credited
The reportYours either way
Build quoteFixed, before you say yes
Committed at step 1Nothing
Committed at step 2Nothing
If nothing in your week clears the bar, we say so and that is the end of it. An audit that always finds work is a sales call.
The method

What happens, what you hand over, and where you get to say no.

Four answers for every step, because those are the four things you actually need before agreeing to any of it.

  1. 01

    Discovery

    30 minutes · free

    A call with no deck. We ask what eats your week and tell you whether there is anything in it worth automating at all.

    If there is not, that is a perfectly good outcome and we say so on this call rather than selling you an audit to find out.

    You provideHalf an hour and an honest description of your week.
    You getA straight answer on whether automation helps you.

    Your decisionWhether to buy the audit. Nothing is committed here.

  2. 02

    Workflow mapping

    90 minutes · part of the audit

    Ninety minutes on your actual week. Not a questionnaire — we follow the work through whatever tools it already lives in, and watch where it stalls, gets re-typed, or waits for one person.

    You provideAccess to watch the work happen, and whoever actually does it.
    You getYour process written down, which most businesses have never had.

    Nothing to decide yetThis is the audit you already bought. Nothing is being decided yet.

  3. 03

    Proposal & success criteria

    The written report

    Every candidate we found, ranked by payback: hours back per month, a fixed build quote, and a feasibility check. Ranked by what pays for itself, not by what we would enjoy building.

    The ones that are not viable are in the report too, with the reason. You paid for the truth, not for a yes.

    You provideNothing. The work is done.
    You getThe ranked report, the fixed quotes, and the measure each build will be judged on.

    Your decision — the real oneBuild one, build three, or take the report to someone else. The quote is fixed before you commit, and the report is yours either way.

  4. 04

    Build

    Usually 2–4 weeks · sometimes months

    One to three workflows, designed and built inside accounts you own — your workflow platform, your cloud project, your AI key.

    Nothing is built on a platform of ours that you would have to keep renting. Ownership is spelled out in full.

    Most go live two to four weeks after you approve the quote. Some take months — scope, or a system that will not give up its data cleanly. We say which is which before you commit.

    You provideLeast-privilege access to the systems the workflow touches. Never a password.
    You getThe workflows, the prompts, the logic — all of it yours to keep or hand to someone else.

    Already agreedScope and price were fixed at step 3. You will know every number before we start.

  5. 05

    Testing in shadow mode

    5–7 days · zero writes

    The system runs against your live data for five to seven days without touching anything. Watching, not acting. Every decision it would have made is recorded so you can read them before any of them count.

    Our last shadow run caught two defects before they reached real mail.

    You provideA week of normal work. No change to how anyone operates.
    You getA record of what it would have done, on your real traffic, before it can do any of it.

    The safety gateZero writes for the whole run. Nothing it decides during this week reaches a customer, a file, or a payment.

  6. 06

    Training your team

    The people whose work this changes are shown how it behaves, what it escalates, and what to do when it flags something for them.

    It is documented as well as demonstrated, so the knowledge does not live in one person who was in the room.

    You provideThe people who will actually use it.
    You getRunbooks and architecture notes, kept current quarterly rather than written on the way out.

    Nothing to decide yetNothing changes state during training.

  7. 07

    Deployment

    The shadow run ends and the system starts acting: filing, labelling, drafting, escalating. Every action it takes is reversible, because it archives and labels rather than deleting.

    What it still cannot do has not changed, and never will on a build of ours.

    You provideNothing new. It is already running in your accounts.
    You getThe work handled overnight, and the exceptions stacked at the top of your inbox.

    Built without the capabilityNo send node and no payment node. Anything that reaches a customer or moves money is prepared and left for a person. A filter can be bypassed; a missing tool cannot.

  8. 08

    Monitoring & improvement

    Ongoing · cancel with 30 days

    We alert on the run that never happens, not only the run that errors — because on our own systems 4 of 27 scheduled runs once left no record and nothing told anyone. That record is published.

    Then one page a month, same format every time: what was patched, what shipped, what broke, what it cost, and one recommendation.

    You provideA decision each month on what the build days are spent on.
    You getMonitoring, published response targets, and capacity to improve it. Priced separately.

    Your decision, every monthCancel with 30 days for any reason. If build capacity goes unused three months running, we will tell you to downsize.

Stopping

You can stop after any step, and keep what you paid for.

Every process page on the internet is drawn as a one-way arrow. This one has exits, because the exits are the part that should make you comfortable starting.

After the call

You have spent nothing and you have an honest answer about whether automation is worth your time. If it is not, we would rather you heard that from us in thirty minutes than found out in six months.

After the audit

The report is yours. Build one of the candidates, build three, or take the whole thing to another firm and have them quote against it. The $500 is credited if you build with us and simply spent if you do not.

After the build

It lives in your accounts and keeps running with nobody on retainer. There is no licence key to expire and no service of ours to disconnect.

During the retainer

Thirty days’ notice, any reason, no exit interview. Or pause for up to sixty days a year at a quarter of the fee, with monitoring still on, so a slow season does not force you to cancel and rebuild.

Step 01

Start with the call that costs nothing.

Thirty minutes on how your week actually runs. If there is nothing here worth automating, that is what we will tell you.