← Projects

QBO-P01 · Quobo Labs fleet

Events by Le Voyage

Labs partnershipLive

A real events business running on lab-built code: a Louisiana dance-cruise company whose whole back office — bookings, payments, group finance, an AI assistant — is the lab's investment in the business it operates.

Live and running a real business — the story is below.

Chapter one

The problem

The business came first. A founder with thirty years in cruise travel organizes themed sailings for dance communities — a beloved band on board, hundreds of dancers, a block of cabins on a major cruise line, a dance package on top of the fare. Two or three groups run at once, with clients spread across the country. Demand was never the problem.

The operating reality was. Bookings were paper forms — with full credit-card numbers written on them. Files lived in a duplicated cloud-drive tree where one completed sailing's folder was simply missing. The group ledger — deposits collected against cruise-line deadlines, comped band cabins, whether a sailing would even break even — existed nowhere but the founder's head. Every booking was a phone call with her. Every payment was manual. Every entry into a cruise-line portal was fifteen to twenty minutes of retyping.

The constraint on growth wasn't the market. It was that the person whose relationships create the product was also the data-entry clerk, the collections department, and the database. So the lab made an investment — not with a check, but with code. It built the business's operating platform end to end and took over technical operations, freeing the founder for the two things only she can do: the bands, and the cruise lines.

The mechanics

How it works

  1. Every sailing sells itself from one page

    Each cruise gets a public landing page — ship, itinerary, band, cabin cards with live sold-out state and honest all-in prices. Share links are operator-grade: a URL can hide cabin categories, inject a bespoke rate, or carry a promo, so a text message to one longtime client quotes her differently. And every promoter — a band, or each member individually — gets an attributable link with a printable QR square and a clicks-to-bookings-to-revenue funnel behind it.

    A sailing's landing page: cabin cards, live sold-out state, all-in pricing

  2. A booking wizard instead of a phone call

    Cabin → Guests → Add-ons → Insurance → Review & Pay. The price is always itemized and all-in — fare, port charges and taxes, gratuities, dance package, per person, no hidden math. A deposit holds the cabin; the balance can ride an interest-free monthly plan, and bank transfer is free while cards carry a processing fee — steering by design, for a fee-averse clientele. Add-ons are chosen at booking, which eliminated the paper options form entirely.

  3. A portal where every passenger is a citizen

    Travelers sign in with a six-digit emailed code and see their trip: a journey stepper, the itemized price, payment history, documents, and a message line to staff. Unusually, secondary passengers are first-class — a guest on someone else's booking can sign in, view the trip, pay her own share with her name stamped on the payment, and ask her own questions. The copy is deliberately hand-holding; the clientele skews older.

  4. The Console: the whole back office in one surface

    Eight modes — Home, Records, Messages, Billing, Reports, Tasks, Studio, Partners — each badge-counted with the work it owns. Home opens on an AI-written daily digest above urgency-ranked action cards that deep-link straight into the screen that resolves them. Records is the spine: the group → booking → client tree, with holds, splits, and cancellations governed by a real status matrix instead of a sticky note.

    The Console's Home mode: the morning digest above ranked action cards

  5. A ledger that used to be a feeling

    The group-finance engine computes what the founder once held in her head: deposits against cruise-line group deadlines, earned cabin credits, projected commission, package collections, comped band cabins — down to a per-sailing net position and a viability flag that raises an alert when a group trends below break-even. A scenario modeler answers "what if we sell fifteen more interior cabins" and compares cruise lines' terms before anyone signs.

  6. Twenty scheduled jobs do the chasing

    Payment reminders ahead of each installment. Insurance-deadline checks while the pre-existing-condition window is still open. Document chasing, hold expiry every fifteen minutes, post-cruise follow-ups, and a morning alert for money collected but not yet posted to the cruise line. The system remembers and reminds; the humans handle only the exceptions.

  7. An AI back office, on a leash

    An assistant panel rides on every admin page: describe the chore in a sentence — record a mailed check, send a confirmation — and its tools do the clerk's work. Destructive actions require a two-step confirm, and a guard rewrites any reply that claims an email was sent when none actually was. Beyond it, browser automation logs into cruise-line agent portals and does the retyping — under one hard rule: the robot never touches credit-card data.

Field notes

Would this fit your life?

A dancer with a flyer from Friday night

The moment
The band's accordion player hands her a flyer with his own QR square printed on it.
What they do
The square lands her on the sailing's page — ship, itinerary, band, live cabin availability — and she books herself: cabin, guests, a bus seat, insurance, an interest-free monthly plan, paying the deposit by bank transfer to skip the card fee.
The payoff
A paper flyer became a fully priced, self-served booking — and the business can see which musician sold it.

A second passenger paying her own way

The moment
An email says her half of the final payment is due — but her sister made the booking.
What they do
She signs in with an emailed code, sees the trip and the balance for herself, pays her share by card, and messages staff a question about dining — no borrowed login, no relaying through her sister.
The payoff
Shared cabins stop producing "my sister has the password" phone calls; every traveler is a citizen of their own trip.

The operating partner, 8:30 a.m.

The moment
Coffee, and the Console's morning digest.
What they do
Reads the AI briefing — two final payments overdue, an insurance window closing Thursday, yesterday's money not yet posted to the cruise line — then works the ranked cards, each one click from the screen that settles it.
The payoff
The whole business triaged in minutes, and nothing about it waiting on anyone's memory.

The founder, deciding on a new sailing

The moment
A band says yes and the cruise line quotes rates. Instinct says go — but this used to be a leap of faith.
What they do
The scenario modeler runs the prospective group across two cruise lines' terms, finds the break-even passenger count, and shows what a slow first month of sales would do to the net position.
The payoff
The go/no-go moves from instinct to arithmetic — and the founder still picks the band.

No varnish

From the lab notebook

Where it stands

Live, and not in the demo sense: real sailings, real bookings, real money moving through the full sell → book → pay → operate → reconcile loop, with migrations applied to a production database. The Console's eight modes run the business daily. Behind it: roughly two hundred test files, a CI gate on every change, over a thousand commits across five months, and a session changelog that reads like the partnership's diary — entries open with what the founder asked for and close with what shipped.

What operating a business taught the lab

Requirements stop being hypothetical when the requirement is this month's cash. The platform grew features no vendor would ship — a two-directional reconciliation that closes the loop from payment processor to cruise-line remittance, follow-up sequences tuned to this one clientele — because the builder answers for the outcome, not the invoice. Honesty got institutionalized too: an automated audit found that most of the app's signals never reached the old dashboard, and that finding drove the Console rebuild. The founder and the developer even message each other inside the product, so the software's own schema records how the partnership works.

What's still rough

A phone voice agent is built end to end and tested — and dark: no number provisioned, no real calls yet. The cruise-line retyping robot is fully wired but not yet doing the daily work. Insurance quoting deliberately routes to a human. Automation rules run in production, but the screen for editing them was removed in a consolidation and hasn't returned. Fee-free payment requests are settled by hand. And the deepest debt is written down by the project itself: most database migrations were applied manually, so the database can't yet be rebuilt from tooling alone.

Capital as code

The platform is not for sale, and there is no software revenue anywhere in it — by design. It is the moat of one business the lab holds a stake in. Nobody builds a cruise-line reconciliation report or a per-band-member attribution funnel for a monthly subscription fee; those only get built when the builder is invested in the business's own results. That is the partnership thesis in practice: where other investors write checks, the lab writes software — and the return is the business it helps run.

Next in the fleet

QBO-P02

Morgan Land Co.

Pre-launch

A Lafayette landscape company the lab invested in with code: a homeowner gets an AI-measured price from their own lot in about 90 seconds, while a designer studio, a constraint-solving scheduler, and a customer portal run the rest of the business on one data model.

Continue the tour →