Skip to content
RANGE

A working operating tool built by one of your own people, with RANGE on the work — and the standard that stays: which decisions justify a tool, how to prove one against numbers you already trust, and how it survives the person who built it.

Fixed-scope install

From $7,500

One rented tool at $200 a month per location, across eight, is about $19,000 a year.

$1,000 discovery week, credited toward the install.

If the week concludes an install isn’t the fix, you’ll hear that plainly — the $1,000 bought the read that saved you the rest.

A 30–60 day install, scoped after a discovery week: the first tool specified, built, proved, and running on live shifts before we step out.

For groups with a systems-minded operator on payroll and work still running on memory, spreadsheets, and rented software that doesn’t fit.

Every group already has in-house tooling; most of it just isn’t admitted to. An AGM builds the spreadsheet that really runs ordering. A kitchen manager keeps the prep math in a workbook only he can read. It works, quietly, until the week that person resigns — and then the operation discovers it has been running on a tool with no spec, no documentation, and no owner. That is worse than no tool at all, because everything downstream kept moving as if the number were being checked. Meanwhile, the decisions that would actually justify a build keep running on memory, because nobody is sure which ones are worth it.

This install is the second path into bespoke tooling. The first is RANGE building the tool for you; here, one of your own people builds it, with us on the work the whole way. The teachable part was never the code — AI-assisted methods have collapsed that. It’s the operator discipline around the code: which recurring decisions justify a tool, how to specify one so the answer is checkable, how to prove it against numbers you already trust before it touches a shift, where it lives, and who maintains it when the menu moves. Your builder finishes with a working tool running on live shifts, and your group keeps the written standard for building the next one.

The first thing the discipline teaches is the arithmetic that decides whether a tool is worth building, so run it on the install itself. One rented point tool at $200 a month per location, across eight locations, is about $19,000 a year, every year, for software built to fit the average restaurant. A tool your own builder makes costs the operation once, and what gets built stays with the business — running in accounts you control, with no subscription behind it. From $7,500, the capability is the deliverable. The first tool is proof it took.

Start with a conversation.

Take the Diagnostic

The Straight Read shows you which install to run first. Back within 48 hours.

What gets installed

  • —A first working tool, built by your builder on a live recurring decision — chosen in discovery because it’s already costing you money
  • —The tool-worthiness discipline: which recurring decisions justify a build, with the arithmetic that decides it
  • —A written specification standard: what the tool answers and how to check it, so the output is provable, not taken on faith
  • —The proving protocol: the tool runs against numbers you already trust before a shift depends on it
  • —The maintenance standard: where the tool lives, who owns it, and what changes when the menu moves — documented, with your builder trained to run it

How the install works

DiscoveryFind the decision that justifies a tool, name the builder, and price the install from both.
BuildYour builder writes the spec and builds the tool, with us on the work the whole way.
InstallProve the tool against the numbers you trust, then run it on live shifts.
HandoffDocument the standard, watch the tool run without us, and leave the capability in the building.

What this isn’t

A fixed-scope install with a defined start and end, not an open-ended engagement. It does not include:

  • ●RANGE building the tool for you (that’s the full AI-Assisted Tools engagement, scoped separately)
  • ●A coding course or general AI training (the discipline is operator judgment, not a curriculum)
  • ●Replacing your POS, accounting, or inventory systems (the tool works from the data they already produce; nothing gets ripped out)
  • ●The food-cost or labor discipline itself (that’s the Food-Cost Install or the Labor & Scheduling Install — a tool serves a discipline, it doesn’t substitute for one)
  • ●Your second tool (the standard your builder keeps is what that’s for)

You’re in the right place if

  • —Somebody on your team already built the spreadsheet the operation quietly leans on, and nobody else can read it
  • —You keep renting point software that almost fits, then bending the operation around the gap
  • —You have a systems-minded manager or analyst who could build the tool, if someone showed them the discipline

Common Questions

What does the In-House AI Tooling Install cost?

From $7,500. The first tool’s reach is what moves the number: how many systems it has to read from and how many units it touches. Fixed once discovery has scoped it, all-in, never billed hourly or by the visit. The $1,000 discovery week is credited toward that number if you move forward.

We don’t have a developer. Who actually builds it?

You don’t need one, and that’s the point. The builder is a systems-minded manager, analyst, or ops lead who already understands the work — AI-assisted methods carry the code, and the code was never the hard part. What can’t be bought off a shelf is knowing which decision justifies a tool and whether its output can be trusted, and your operator already has half of that. The install supplies the other half. Most groups know exactly who this person is before discovery ends; some knew before they called.

How much of the builder’s time does this take?

More than a course and less than a second job. Figure a real block of hours weekly across the thirty to sixty days — this is working on a live decision, not exercises — with the heavier stretch in the weeks the tool is proved against your numbers. Their manager needs to know that going in: a builder doing this in stolen hours is how a tool ends up private, which is the failure the install exists to prevent.

What happens the week the person who built it resigns?

The tool keeps running, because the install never let it become private. That’s what the standard is for: a written spec that says what the tool answers and how to check it, documentation a second person has already worked from, and a named maintenance owner. A tool built quietly by one manager who then leaves is worse than no tool — the numbers keep moving as if someone were checking them. The standard is the difference between a tool the operation runs and a tool one person happened to make.

Why not just have RANGE build the tool?

That path exists — it’s the full AI-Assisted Tools capability, and for some tools it’s the right call. This install is for groups who want the capability inside the operation, not just the tool: the first build takes the same discipline either way, but the standard your builder keeps is what makes the second and third tools cheap. If discovery shows the first tool is bigger than an install should carry, we’ll say so, and scope it as a build instead.