Fixed fee, in writing

AI Capability Audit

An AI readiness assessment for owner-run companies

We find the chore that's costing you margin, price it from your own data, and put the verdict in writing — build, or don't build.

$7,500, fixed · A fixed window, start to verdict

Term sheet

The terms

Everything a diligence question would ask for is on the page. If a term below changes before you sign, you'll have it in writing first.

The fee$7,500, fixed. We quote it before you commit and write it into the engagement letter. It is not a starting point.
The durationThe window is fixed, agreed before kickoff so the audit can't quietly become a retainer.
The deliverableYou get a build-or-don't-build verdict and a spec you keep either way — worth having especially if the verdict is don't build. The findings are numbered and priced from your own data.
What you keepYou keep all of it: the verdict in writing, the numbered findings, the spec. Take the spec to another builder if you want. It's yours.
DecouplingThe fee is never credited toward a build. The moment it is, the audit has a sales agenda again. That's the point of charging for it.
Book the audit

Booking puts a time on the calendar. The engagement letter is what commits you.

Sample pages

The verdict

The deliverable is a findings memo, not a slide deck: verdict first, numbered findings, your own figures.

Every finding follows the same shape. First, what the work is now. Then what it costs, where we can price it from your data. Then a verdict: build, or don't build. Don't-build items are listed with the same weight as build items. The memo also shows where each finding sits on the framework we score against, and the full map is here /framework/.

Finding 01 · Build

Daily rebate-terms check

Now: one person reconciles rebate line items against contract terms when there's time. Priced from your own claims data. Verdict: build — a tool inside the ERP runs the same check every morning, a person approves anything it flags.

Finding 02 · Don't build

Ad-hoc pricing exceptions

Now: a handful of one-off pricing calls a month, each needing judgment. Verdict: don't build — the volume is low and the exceptions rarely repeat. A tool here would cost more to trust than it saves. Kept, with reasons.

From a sample deliverable: one build finding, one don't-build finding.

Before you book

The builders

You're handing margin data to whoever runs this audit. Here's who that is, what we've built, and what you'd hold afterward.

Who you're hiring

Throughline is a small team of builders, part of Keller Group, the family of operating companies it grew out of. Ten production systems plus an agent platform run every day inside real operating companies, built by a team of two or three. Primary runs the ops desk, Scorecard the weekly numbers, Cadence the leadership meeting. The full roster is on the systems page /systems/.

The proof

At Kelsan, a multi-state distributor in our own group, running the Epicor P21 ERP, someone walked rebates line by line against contract terms, by hand, when there was time for it.

We built the rebate-and-margin tool inside that P21 system. The tool uses AI to read the rebate terms in the contracts and check every line item against them, every day. It surfaced over $100,000 in recovered margin the manual process was missing.

That tool is one of ten. The systems above are the habit. Read the full case study →

What you'd hold

The audit's deliverable is yours either way. If the verdict is build, that's a separate decision, priced on its own, and what we build you own. Every build is a fixed fee, agreed in writing before work starts. It runs in your tenant. There's no per-seat rent and no proprietary platform. None of that is the last word, though, because software needs maintenance. Here's what's yours to run, and here's the optional support arrangement if you'd rather not /security/.

Your people

Nothing changes in your business without a person's sign-off. The people who know the numbers best are the ones with the least time to chase them. The tool does the chasing. Your people keep the judgment. We don't build dashboards to watch your own staff, and when we hand a tool over, we hand it to your team, with training.

Campaign Manager, one of the systems that runs our own companies, tries to prove its own recommendation wrong before it asks a person to approve anything — and the result is measured against a prediction written down in advance.

Your IT and MSP

Keep your MSP. They should get a veto. They'll say yes faster with the one-pager we wrote for their team: where the data lives, what we never touch, how the sign-off chain works. Send it to them before you book, if that's the order your company does things in. /security/

Stated terms

Not a fit

Some of the fee's value is knowing when not to pay it. These are terms, not small print.

  • 01

    There's no core system to build inside. Tools like it live inside software you already run — an ERP, an accounting system, an operations platform. If the business runs entirely on email and memory, an audit can't wire a tool into it.

  • 02

    You want the fee credited toward a later build. It never is. The reason is in the term sheet above, and we hold to it even when it costs us the sale.

  • 03

    You're after a company-wide strategy engagement. The deliverable here is a verdict and a spec for one working tool. A big-picture plan for the whole business is a different hire.

  • 04

    You need a national firm's letterhead for board cover. That's a real need and we can't meet it. We're a small team of builders. The proof is the systems, and the systems have names.

  • 05

    You're shopping free assessments. Ours is priced on purpose, and the why is just below, in the FAQ.

Book the audit

Fixed fee, fixed window, and the verdict is yours no matter which way it lands.

Two minutes

Self-check

Six plain questions about how the money-adjacent work in your building gets done. Two minutes, no email required.

Most of the work these questions ask about was never anyone's actual job. Nobody built a job around chasing it — it lands on whoever knows the numbers.

01  Does someone still check margin, rebates, or contract terms by hand?

02  Does that work happen only when someone has time?

03  If that person left tomorrow, would the checking stop?

04  Do problems surface after the money is already gone?

05  Does the check get skipped in a busy month?

06  Has anyone put a dollar figure on what that chore costs?

FAQ

Questions

How much is an AI readiness assessment?

Ours is $7,500, fixed. That covers the whole engagement: a fixed window, the numbered findings priced from your own data, and the verdict in writing. There's no hourly meter behind it, and no follow-on fee is needed to understand what you bought.

Why isn't it free?

A free assessment is paid for by the build it recommends. A fixed fee is what independence costs. Your ERP channel's free assessment works the same way: it's the front end of an implementation funnel. So before you pay anyone for one of these, check three things — the fee is fixed, the verdict is decoupled from any build, and you keep the deliverable. Every one of those is on this page.

Is the fee credited toward a build?

No. The fee is never credited toward a build. The moment it is, the audit has a sales agenda again. A verdict you can trust costs the same either way, and the spec is yours to take to another builder.

Couldn’t someone on staff use ChatGPT?

For one contract, once: probably, and it's a fair experiment. What the audit prices is different work — the same check running every morning, for anyone on the team, inside the ERP you already run, with a person signing off on anything that changes. If an afternoon and a chat window genuinely close the gap, the verdict will say don't build, in writing.

Who is this not for?

The stated terms are above. The short version: there's no core system to build inside, or you want the fee folded into a build. Both are dealbreakers we'd rather surface now, before the fee, than mid-engagement.

Why not wait for Epicor's built-in AI?

The built-ins are genuinely useful, and generic by design. That limit is the whole question here. The honest version of that answer, including what waiting costs and what Epicor's own tools are actually for, is on the distribution page.

How do consultants assess AI readiness in businesses?

Most start with interviews and a maturity questionnaire, then hand back a report that scores where you sit. Ours differs in one way that matters: instead of a score, we find the specific chore leaking money, price it from your own data, and return a build-or-don't-build verdict you keep either way. The assessment is fixed-fee and decoupled from any build, so the recommendation is never a sales step.

Booking

Scheduling, and nothing more, until an engagement letter is signed.

Who to reach

A few quick fields. The rest can wait for the call.

Not ready to book? We'll send the one-page terms and a sample finding.

Fixed fee · in writing

One step

The fee, the window, the deliverable, and who it's not for are all on this page. What's left is picking a time.

$7,500, fixeda fixed windowa verdict you keep either way