Case study, composite, process-led

One install, start to finish.

A service company with three tools that did not talk, one owner reading messages at midnight, and a hiring inbox nobody opened. What we found, what we chose, what happened at the 30, 60 and 90 day marks, and what it cost. Details are drawn from the companies we hold and run, blended so no single business is exposed.

No screenshots here on purpose. The screens are on the systems in production page and we share them live on a call.

One install, start to finishScrollSalt Lake City · Mountain time
The companyLocal service business

Roughly forty staff on the road, one office, seasonal demand, most jobs booked by phone or form.

The stack beforeThree tools, zero links

A booking site, a spreadsheet for scheduling, a job board for hiring. A person was the integration.

Engagement shapeInstall, grow, then operate

Two operating systems in the first sixty days, the growth layer in the third month, a care plan from day ninety.

Price$12k to $20k, then fixed

The vertical install range on the pricing page. Two things move it: how many source systems we connect, and whether the data needs cleaning first. Fixed once quoted.

Week zero

The teardown found four leaks, not one.

The owner asked for a chatbot. The teardown, a recorded ninety-minute pass through every tool and inbox, said something else. Leads arrived after hours and sat until morning, and a share of them had masked contact details that nobody recovered. Quotes were priced from memory, sometimes below the margin floor. The hiring inbox had well over a hundred applications marked new. Meetings ended with intentions instead of tasks.

A chatbot would have fixed none of that. We wrote it down, with numbers, and gave the owner the recording to keep. That document is the free teardown we offer every company, and it is the same one we would have handed over if the answer had been "you do not need us".

  • LeadsAfter-hours arrivals, masked contacts, no recovery, no pricing.
  • QuotesPriced by memory, no floor, no record of win and loss.
  • HiringApplications unread for weeks, no scoring, no scheduling.
  • MeetingsDecisions made, nothing assigned, nothing followed up.
The decision

Two systems first. Then the growth layer.

We do not install everything at once. The lead system went first because it was the largest number in the teardown; the hiring system went second because it removed the bottleneck that would have stopped the company using the first one. Once the office trusted what it saw every morning, the third month went to growth: the content engine, search visibility and paid landing pages, so the new capacity had demand to fill it. Meetings, finance and the rest of the operating layer came after.

The install timeline

Thirty, sixty, ninety.

  1. Days 1 to 30Install

    The first system goes live and the owner sees it work.

    Week one is access, data and a clean read of the source systems. Week two the lead engine runs beside the old process, in shadow, so every recovered contact and every priced quote can be checked by a human. Week three it goes live with a text to the owner on every lead. Week four we sit with the office and watch.

    • Lead recovery and pricing live
    • Owner briefing every morning
    • First win-loss records
  2. Days 31 to 60Second system, training

    The hiring funnel clears the backlog and the team stops asking us.

    Applications are scored against the role, invited to a questionnaire, then offered interview slots the manager already agreed to. The old backlog is worked through in the first week. The office manager runs the second week without us on the call. We tune thresholds from what the manager overrides.

    • Scoring, questionnaire, slot booking live
    • Team trained on the briefing and the queue
    • First month of numbers compared to the teardown
  3. Days 61 to 90Growth layer, then review

    The growth layer goes in, then the whole thing is measured against what it promised.

    With capacity to sell, the third month builds demand: a content engine that turns the owner's walkthroughs into weekly videos and posts, search pages for the services and areas that convert, one paid landing page with tracking that ties spend to booked jobs. Day ninety is a review with numbers: leads recovered, jobs priced above floor, applicants moved, cost per booked job, hours the owner got back. The owner chooses a care plan or a hand-off. Either way the system stays theirs.

    • Content engine and search pages live
    • Paid landing page with attribution
    • Ninety-day numbers against the teardown
    • Care plan or hand-off, owner's choice
What changed

Before and after, in the owner's day.

MomentBeforeAfter
A lead arrives at 9:42pmSits until morning. Masked contact stays masked.Contact recovered, job priced above floor, owner texted the number. Customer answered by 9:47pm.
QuotingFrom memory, sometimes under cost.Rules engine with a floor. Every quote a record with win or loss.
HiringInbox of unread applications, interviews by phone tag.Scored, questioned, booked into slots without a call.
MondayOwner reads every inbox to know what happened.One briefing: what needs you today, with links to act.
MeetingsIntentions in someone's notebook.Transcribed, decisions assigned as tasks with owners and dates.
MarketingA post when someone remembers. Ads with no idea what they return.Weekly video and posts cut from the owner's own footage. One landing page, spend tied to booked jobs.
What it cost, and what it returned

Fixed price, then a monthly number the owner chose.

The install landed inside the vertical install range, $12k to $20k. Two variables set the number: three source systems to connect, and a booking export that needed cleaning before it could be trusted. Once quoted it did not move. The care plan afterwards was a monthly number the owner could cancel; they kept it. We measure return on what a system earns or recovers rather than on hours, which is why we could price it before we started.

The honest note: the first month is slower than the owner wants, because shadow mode is boring. It is also the reason nothing had to be undone later.

  • QuoteInside $12k to $20k. Moved by source-system count and data cleanup. Fixed once quoted, no change orders.
  • PaybackMeasured at the ninety-day review against the teardown numbers.
  • OwnershipTheirs. Code, data, documentation, recordings.
  • Done differentlyStart hiring in week two, not week five. The backlog was the real bottleneck.
The rest of what we bring

Operations was the leak here. It is not the only thing we install.

None of this is sold by default. Every company gets the same teardown, and the answer is different each time. A podcaster needs the content engine and the store before any operations layer. A clinic needs intake and compliance first. A retailer needs search and paid pages with attribution. These are the practices we bring to every engagement; the teardown decides the order.

The companies behind the composite

A composite, and here is what that means.

The company in this story does not exist. The events do: the 9:42pm lead is a real one from our cleaning company, the applicants were scored inside the same company, the meetings that assign work run on our operations platform. We assembled them into one arc so the shape of an install is clear without exposing any single business. The screens are on the systems in production page.

The companies we run Systems in production

Two real ones

The composite above is assembled. These two are not.

Our own cleaning company and our own platform, with the numbers as they are counted on the return page and the thing that failed first on each.

Wasatch Cleaners: quote to cash Akovex: the platform itself

Your teardown is
day zero.

Ninety minutes on a recorded call inside your tools: what to automate, what to replace, what it is costing you. Free, specific, and the recording is yours.

Book an AI teardown