Case Studies

Real systems. Running in real businesses.

Real systems, live in production or actively in development — most of the finished ones built for our own companies first. We build for ourselves before we build for you, so we know exactly what breaks and what doesn't.

12+

Systems in production

Voice agents on the phones 24/7

Properties priced nightly

Companies running these systems

HospitalityBerlin Encore Hotel

The Hotel That Stopped Missing Calls

A 24/7 voice agent that answers every call the front desk can't, quotes live tax-inclusive pricing straight from Cloudbeds, and texts a ready-to-book link before the caller hangs up.

24/7

Every call answered — nights, weekends, holidays

< 2 rings

Average time to pick up

Read case study

Live Entertainment & TicketingAmish Country Theater

The Box Office That Never Closes

An after-hours voice agent for a 538-seat theater's box office — answering only when the desk is closed, quoting live show data, and texting a ticket link before the caller hangs up.

After-hours only

Never answers while the box office is staffed

~90 sec

Average call, ticket link included

Read case study

Travel & Group ToursLand Cruise Vacations

Owning the Checkout Instead of Renting It

A custom registration and checkout platform for multi-day motor-coach trips — built to replace a per-transaction SaaS fee with direct Stripe processing, atomic inventory, and an operations layer that runs itself.

Platform fees — Stripe's processing rate only, nothing layered on top

Zero holds

Inventory claimed atomically at order completion, not reserved on entry

Read case study

Vacation Rentals & HospitalityAmish Country Lodging

Dynamic Pricing That Can Explain Itself

An explainable pricing engine for 75 cabins, lodges, and treehouses — every price traceable to the exact chain of rules that produced it, with a dry-run safeguard before anything reaches a guest.

Cabins, lodges, and treehouses priced automatically, every day

Every price

Explains itself — the full chain of seasonality, demand, and rules behind it

Read case study

Hospitality & Local BusinessFive Star Group Properties

Waking Up a Dormant Google Business Profile

A reputation agent that monitors Google Business Profiles in real time, drafts on-brand replies and Q&A answers, and routes anything sensitive to a human — replacing a $300–$1,500/mo reputation-management SaaS.

24/7

Every new review, question, and photo monitored within minutes

80%

Of routine review and Q&A volume runs fully autonomously

Read case study

Executive OperationsFive Star Group — Founder's Office

The AI Chief of Staff That Got Us to Anfield

Zeke, the founder's own AI chief of staff, holds every confirmation, email, and calendar detail across the business — and, on one cancelled-train morning in London, replanned an entire day in under a minute.

1 minute

From "our train was cancelled" to a full working plan

Every morning

Inbox triage, calendar, and a daily brief across every business

Read case study

Internal ToolsFive Star Group

The Chat App With AI Teammates in the Room

A cross-platform team chat app built to end the per-seat SaaS bill — with AI agents wired in as first-class teammates through a bot API, not bolted on through an integration marketplace.

4 platforms

Mac, iPhone, Android, and web from one codebase

89 commits

Zero to a team-wide beta in about a month

Read case study

Live Entertainment & TicketingAmish Country Theater

The Seat Map That Can't Double-Sell a Ticket

A seat-map ticketing platform built for the Amish Country Theater's 538 seats — an interactive SVG seat picker, a ten-minute checkout hold with atomic seat claims, and a staff point-of-sale, owned outright instead of rented from a vendor.

538 seats

Rendered as a real, interactive seatmap — every row, every section

Zero double-sells

Atomic seat claims inside a single database transaction

Read case study

Internal ToolsFive Star Group

Company Card Spend, Categorized Two Ways at Once

An internal spend-management app for a holding company running several LLCs on shared cards — every transaction gets a Category and an Entity, pre-guessed automatically at ingest so the accountant reviews instead of starting from zero.

2 dimensions

Every transaction gets a Category and an Entity, not just one

1 day

From empty folder to a working, seeded, deployable app

Read case study

People OperationsFive Star Group

People Ops for a Company That's Really Five Companies

An HR and people-ops platform built for a holding company where the same employee can work shifts under two entities in the same week — multi-entity as a first-class data model instead of a spreadsheet-and-filing-cabinet workaround.

Multi-entity

First-class in the schema — an assignment per entity, not a workaround

Full ledger

Every PTO balance traceable back through its complete history

Read case study

MarketingFive Star Group

The SEO Agent Watching Every Property's Rankings

An always-on SEO capability that audits every Five Star Group brand site on a weekly cadence, tracks rankings over time, and turns three separate data sources into one digest — replacing the occasional, reactive burst of SEO work that used to happen only when someone found the time.

Weekly, not quarterly

Audit cadence across every brand site

3 data sources, 1 digest

Ahrefs, Search Console, and Analytics unified into one weekly readout

Read case study

AviationJMC Billing

The Invoice That Builds Itself

A personal contract-pilot billing tool — not a generic small-business bookkeeping product — that pulls flight data automatically, reconstructs the month into billable days, and generates per-owner invoices without an evening of spreadsheet archaeology.

3x daily

Automated flight-data sync, deduplicated by provider ID

48 commits

Spec to a working invoice, over about three weeks

Read case study

Construction EstimatingA Regional Rebar & Concrete Accessories SupplierIn Development

The Takeoff That Has to Show Its Work

An AI-assisted rebar takeoff and estimating system for a concrete accessories supplier — still mid-build, being validated job by job against real historical estimates before it ever touches a live bid.

17 jobs

Real historical client estimates run through validation

2 estimators

Independent AI takeoffs, cross-examined before either sees the answer

Read case study

Also built

Spam Slayer

A multi-tenant IMAP spam filter with a hybrid rules-plus-AI pipeline that thinks before it quarantines.

Want something like this for your business?

Tell us about your business and we'll show you exactly where a system like these fits — no pressure, no jargon.

Get Your Free AI Assessment Read the FAQ