[jm]Japheth Miller
← Writing
May 10 2026
·
5 min read
·Notes

Otto and JunkSavior — two projects, opposite ends

A pre-press release: why I built Otto out of sixteen years of utility-billing pain, and how I modernized JunkSavior — a two-truck Florida operation — from a Gmail account to a full operational and AI stack in two weeks.

This is the first post on this site. Before any of the longer-form thinking I want to publish here, I want to use it to introduce two projects that have shipped over the last twelve months. Different scales, different industries, same shape of work — design and engineering, end to end, by one contractor.

Otto — utility billing, reconciled to the dollar

I spent sixteen years inside a multi-family software company before I started writing production code seriously. Most of that time, the part of the operation closest to me was utility billing. It was a mess.

There was no history of what had happened on a given account before this month. No paper trail. Spreadsheets in five different formats, kept by three different employees, none of them speaking to each other. When a new hire showed up on day one, nobody had a good answer for where do I start. No automations. No guardrails. No integrations between the things the operator was already paying for and the things they actually used. Read estimates were done by hand against last month’s number on a printout. Meter failures were caught by the resident phoning in. There was no logical model — for read estimation, for failure reporting, for keeping track of cycle history — for any of it.

The customers had it worse. A challenge to a charge meant fishing through six emails to figure out what the operator had done, why, and whether the math was even right. Half the time it wasn’t.

Otto is what I wish that operation had had.

One canonical pipeline — Bills, Reads, Rent Roll, Calculate, Approve, Ship, Close — that every screen and every report ladders to. Seven allocation methods, audited on real vendor bills before any UI was wired up. An audit log on every mutation, abort-on-failure. A “show your math” derivation behind every charge so when a resident calls, the operator can answer the question with the page already open.

Where Otto stands today
Live in production. First paying customer onboarded. Three Founder Partner slots open. The full case study lives at /work/ottomated-app.

JunkSavior — modernizing a two-truck operation, end to end

The other end of the spectrum, same year.

JunkSavior came to me with a phone number, a truck, and a Gmail account. Two weeks later they had a wordmark that works at favicon size and on the side of a truck, a 20+ page WordPress site the owner edits himself, schema-marked-up location pages, and verified profiles on Google Business, Yelp, and Nextdoor across three Florida counties.

But the marketing site is just the front door. The work that runs the business sits behind it.

Jobber handles dispatch and the staff schedule. When a crew is on the way to a job, the customer gets a text. When the job is done, another text — with a one-tap link to leave a review. Common customer questions are answered before they ever reach the owner’s phone. Facebook Ads run on AI automation that both improves the campaigns it has and writes new ones based on what’s working.

That’s the arc: a small business going from a Gmail account to a modern operational stack in two weeks. Old-school by background, modern by tooling.

WordPress on purpose, by the way. The owner has to update his site without calling me — that was the only constraint that mattered. WordPress beats a clever stack every time when the question is “can the owner outlive my engagement.”

Full case study at /work/junksavior.

One thread

These two projects look nothing alike on the surface. A multi-tenant SaaS for property management companies, and a two-truck junk-removal operation in Florida. But the shape of the work is the same: figure out what’s actually broken, decide what’s in the first version, design and build it end to end, ship it.

Otto is the software-design-and-engineering case study above. JunkSavior is the AI-automation-for-small-business one. Both are real, both are live, both are how I work.

I’m freelance. I take on a small number of client projects each year. If either of those tracks sounds like the thing your project needs, /contact is the next stop.

More posts to come.


Japheth Miller
Japheth Miller
writes when there’s something worth writing.