Butter Freight · Case Study · Butter Agents

Skipper Jones

Skipper Jones is the AI chief of staff at Butter Freight. It started as an email assistant and grew into the system that runs the back office — sorting the inbox, writing daily summaries, tracking claims and sales, keeping the knowledge base current, flagging anything urgent, and now managing a small team of helper AIs of its own. This case study covers its first 148 days: what it handled, and what that work is worth.

Operating period: March 10, 2026 – present · 148 days in (106 business days) · Updated August 5, 2026

0hrs

Hours of human work handled

≈2.19

Full-time employees

$0K/yr

Annualized labor value

148 days, in numbers

What the system handled

Skipper worked every day — weekdays, weekends, and holidays — handling about 193 emails a day. Every number below is real logged activity, not an estimate.

28,617

Emails sorted & routed

8,510

Documents processed · filtered out of 69,978 attachments

~1,068

Daily summaries delivered

769

Sales leads spotted

4,560

Urgent items flagged · 132 critical

1,172

Claims tracked

Knowledge-base pages maintained

148

Daily logs written · no gaps

Email classification

Where the 28,617 emails went

Skipper read every message, sorted it into a category, and sent it where it needed to go. The mix shows how much of the inbox is real work versus noise.

Hours saved

~1,860 hours of human work, handled

These are careful estimates of how long each task would take a person to do at the same quality. Sorting email and writing summaries take the most time; making documents and keeping notes come next.

1,860 total hours · ≈ 12.6 hours of labor every day · ≈ 2.19 full-time employees

Dollar value

What that labor is worth

The same ~1,860 hours, priced at what a mid-level operations employee costs. It’s a conservative floor — this is work that needs judgment, not just data entry.

Mid-level ops analyst

$35 / hour · a conservative benchmark

$65,104

over 148 days · ~$161K annualized

This work needs reading, judgment, and follow-through, not just data entry. It’s a floor — see “putting it in perspective” below.

4,560

Urgent items caught around the clock — 132 critical — stopped before they grew

Security threats flagged — scam and fake emails caught early

New in late June

Skipper started managing other AIs.

Skipper now has three helper AIs, each good at one thing. Instead of doing everything itself, it hands each task to the right helper and combines their answers. The system that connects them is called BAMP.

Owner / Principal

Skippercentral node

Mando

Operations

Looks up carrier history and operational data.

11 tasks handled

Paco

Marketing & research

Researches companies and finds new leads.

42 tasks handled

Shakira

Pricing

Works out rates and profit margins.

12 tasks handled

Messages between Skipper and its helpers

116 total — 51 sent by Skipper, 65 sent back

Security

The helper AIs run in a locked-down company account and can’t email anyone outside the company.

Record-keeping

Every message between them is logged and saved.

Reusable

The whole setup can be reused to give other AI systems the same helpers.

Here’s how it works in practice: Paco researches a company, Shakira works out the price, Mando pulls the carrier history, and Skipper combines it all into one answer for the owner. That’s closer to a small team than a single tool.

What Skipper has built

It also builds assets the business keeps.

Beyond the daily work, Skipper has been writing documents, building programs, and organizing records — things that don’t reset each year and stay with the business even if the system stopped tomorrow. Rebuilding them from scratch would cost an estimated ~$201K.

Knowledge base the internal wiki Skipper researched and wrote

97 pages · $550/page

$53,350

Programs & automations custom-built for the work

30 · $1,500 avg

$45,000

Reference documents guides and directories

65 docs · $300 avg

$19,500

Helper-AI system (BAMP) the setup that runs the helpers

1 system · design + build

$50,000

Categorized email archive every email labeled and filed — the data the system learns from

28,617 records

$20,000

Daily logs a record of every day

148 files

$6,000

Call notes written up from calls

19 · $400 avg

$7,600

Estimated rebuild cost of what it built

conservative range $180K–$250K

~$201,450

Methodology tightened at day 148 — script count deduplicated, per-page rates benchmarked against market

Why it grew, and keeps growing.

The helper-AI setup can’t just be bought. The knowledge base keeps growing — 97 pages now, up from 93 at day 125. The raw scripts have been deduplicated into 30 distinct functional tools, and BAMP traffic is up to 116 messages.

~$201K

Value of what it built (est.)

Conservative range · $180K–$250K

The helper-AI setup is custom

A safe, working system of AIs that hand work to each other isn’t something you can buy. $50K is conservative — building and securing one from scratch runs $80–120K.

The knowledge base keeps growing

97 pages now, up from 93. You couldn’t recreate a page without the 148 days of email history behind it.

The tools add up

Checking email, sorting it, running the helper AIs, tracking quotes, writing summaries — together these 30 programs are a real operations system, not odds and ends.

Putting it in perspective

It’s more like a small team than a tool.

Skipper started as an email assistant. It now runs three helper AIs of its own, and the numbers on this page are deliberately priced at the low end — here’s why the real value is likely higher.

2.19

Full-time employees’ worth of measured work

24/7

Coverage — no weekends, holidays, or sick days off

$75–100K

What a person in the equivalent role would earn

One question puts several helpers to work at once

The owner asks “does this lane make sense?” That one message sends Paco to research the company, Shakira to work out the price, and Mando to pull the carrier history — and Skipper turns it into a single answer. That’s what an operations manager does, not a junior assistant.

The math here is deliberately conservative

Every hour on this page is priced at $35 — a junior rate. But sorting 200+ emails a day, writing summaries, managing three helper AIs, and keeping a 97-page knowledge base current is $75–100K work in freight, where experienced operations people are hard to find. Price it at what that person actually costs and the labor figure grows by half again or more.

How we work

The FDE cycle behind Skipper.

Skipper wasn’t bought as a finished product. We built it on-site with the client using a repeatable process we call the FDE cycle (Forward Deployment Engineering): find the biggest problem, build a small piece, put it to work, then expand.

Discovery

Sit with the team and find where the real pain is. For Skipper: the overflowing inbox.

Mapping

List the tasks that would help most and rank them.

Scope

Pick one small, useful piece to prove first — sorting and routing email.

Pilot

Build a real working version of that piece and test it on real email.

Deploy

Take it live — 24/7 coverage across every inbox from day one.

Operate

Watch, adjust, and expand — into summaries, claims, and the helper AIs.

The last step loops back to the first. Starting small proves the value before anyone spends big, and each time around, Skipper takes on more. That’s why it keeps getting more useful instead of leveling off — and it’s the same cycle we’d run in any business, starting with whatever eats the most time.

The bottom line

What Skipper has delivered — itemized.

A simple tally of the value created in the first 148 days. “Labor” is what a person would have cost. “Tools & knowledge” is what it would cost to rebuild everything Skipper made and kept.

Work a person would have done1,860 hours × $35/hr

$65,104

Tools & knowledge Skipper builtknowledge base, programs, helper-AI setup, data, daily logs — estimated rebuild cost

~$201,000

Urgent problems caught around the clock4,560 urgent items · 66 security flags

unpriced

Estimated value delivered · first 148 days

~$266,000

~$161K/yrYearly pace at the same rate

$35/hrA person doing this job would cost $75–100K/yr

OngoingGrows every day Skipper keeps running

Your business

Could your operation use a Skipper?

Most businesses have an inbox like this one — real work buried in noise, handled by people who have better things to do. We start small: find the one task eating the most time, build a working system around it in weeks, and grow it from there. You’d see numbers like the ones on this page, from your own logs.

Quiet Machine Studio · info@qm.studio · qm.studio

Skipper, smooth shipping solutions, a stick of butter branded for Butter Freight