An AI consulting firm · Truckee / Reno–Tahoe · Bay Area · Orange County
Luxin builds custom AI systems inside the business you already run — email, invoices, scheduling, reporting — on your own accounts, so you own what gets built.
Luxin, an AI consulting firm, has systems running in real businesses today. One client has already referred two more owners.
If someone sent you this link
Then you’re the person this was written for. The short version:
Luxin makes a business AI-native: the work gets done by systems, and the owner’s job becomes reviewing and approving instead of doing.
The rest of this page is proof. Three short reads, in order:
The current way
Three hundred emails a day, and fifty of them need a partner’s personal reply.
Half an hour a day of partner time at $300 an hour: roughly $37,000 a year.
Hundreds of handwritten invoices a year, typed into the books and then logged again by hand.
A catch-up runs eight to ten hours when it gets busy.
Two to three hours of typing for every new escrow, and a hundred follow-ups every two days.
One missed inspection traced back to a hand-typed date.
None of that needs a person. All of it needs a person to check it.
Case studies
3 live · 2 delivered earlier · 1 in progress · 4 proposed
A proposal agent that turns an inbound request and a set of plans into a filed job and a drafted proposal, ready for review.
About 6 hours a week of proposal-writing time given back, Marc’s own estimate as published on his earlier portfolio site, not a measurement
An email operating system: every message read, matched to its matter, and answered in draft before the attorney opens it.
100+ emails deep in the inbox before the system was built
The system Marc demos in every meeting: agents draft the briefs, notes and follow-ups for his own company, and he approves every send.
50+ daily briefs built by the system since 20 July 2026
A sales and investor outreach system the company owns end to end: it finds prospects, ranks them, and writes the first email.
3 tiers that every prospect is sorted into before a person looks at it
One client is named with permission. The rest are held back until each says otherwise.
All the case studiesHow we work
What do you understand about AI, and how does the business actually run day to day. We listen first.
Marc shows the one he runs his own company on, then maps it onto your operations.
Your numbers, your hourly rate, every assumption printed next to the figure so you can change it.
One task, one to two weeks, a fixed price, on your own accounts. Then you decide.
A written runbook (the page that says how to run and fix it), a team walkthrough, and the accounts stay yours.
The mechanism
A system watches the work coming in and does the part that is the same every time. Then it stops and waits for you. You read it, change what you want, and approve. Only then does anything move.
It gets more accurate as you correct it. The corrections are the training.
How it stays reliable, and stays yours
Code does the math. AI writes sentences.
Numbers in reports and invoices are calculated by regular code, every time. AI drafts the words.
Nothing sends or deletes on its own.
Reports, replies, and invoices wait for a person. Originals are never removed.
Unsure means ask, not guess.
An email it cannot place goes to a “needs a home” folder for a one-click answer.
Your data stays in your accounts.
Microsoft 365, Google, QuickBooks: your logins, your data, on commercial AI terms that never train anyone’s models.
Everything is logged.
Every file moved and every draft written, with a timestamp. Nothing is a mystery.
Errors alert us, then fall back.
If a step fails, a person is told and the work goes back to the old way until it is fixed. Never a silent failure.
It learns from your corrections.
Every edit you make to a draft is kept as an example, so the next one reads closer to how you would have written it.
Documented, and you own it.
A plain-English runbook at handover. Change it, extend it, or hand it to someone else. No lock-in.
Pricing
We price against the hours the work gives you back — usually 40 to 50% of what the first year saves you. Every price on this site was sent in a real proposal.
The pilot is a single task, $500 to $1,500, with no commitment past it.
Our advice: start with a pilot. One task, one to two weeks, then decide.
Pilot
One task, automated end to end, on your accounts. The baby step.
$500–$1,500
fixed, one payment
Project
A full system, or several, scoped and built and handed over inside one to three months.
$1,500–$5,000
per project, fixed · full builds $9,500–$11,500
Monthly build budget
For businesses that have decided and want it all built by a date, upkeep included.
$3,000/mo
about 30 hours a month, flat
The people
Founder
Marc scopes and sells every engagement, and builds the small systems himself. He is in the room for every conversation. When you hire Luxin, he is who you get.
Co-founder
Pranav studied computer science and AI at USC and works at a cybersecurity company. He reviews how every build handles your data and leads the larger builds.
Luxin brings in engineers per project rather than carrying a payroll. Each one is paid a fixed price for their piece. You pay for the work, not for overhead.
In the owner’s words.
“Really impressive work! From the first call I had with Marc, I could tell I was working with someone who knows what he’s doing.”
Questions
Start with one thing
Thirty minutes, no deck. You describe how the work actually happens, and I tell you plainly whether there is something worth building.
Know someone?
Send this to someone who…