vanhausBook a free discovery call

Work

Real tools for real businesses.

Each one below is a live recreation of the real product, running on made-up data. Clients are described by industry until they say it's fine to use their name.

HVAC & generator company, NJ

Demo coming soon

An AI receptionist that picks up every call

Problem
Calls came in while the crew was on job sites. Voicemails piled up and some leads never called back.
What we built
An AI phone receptionist built on Twilio and OpenAI Realtime. It answers every call, collects job details, routes urgent calls to a person, and emails a summary after each one.
Result
1,000+ calls answered so far.
Read the case study

Tri-state area HVAC & generator company

Generator proposals in minutes, not hours

Problem
Every quote was rebuilt by hand from a spreadsheet, a Word template, and the owner's memory. Slow quotes meant missed bids and jobs lost to faster competitors.
What we built
A proposal generator. Describe the job by form, chat, photos, or a bid PDF, and get a priced, branded proposal the customer signs online.
How it learns
Built on 121 of their past proposals, with every new one added back in. Repeated price edits become new pricing rules the owner approves.
Result
Hours by hand became minutes. Tested against a real job they had won, it landed on the exact same total.
Read the case study

HVAC & generator company, NJ

A QA/QC checklist that checks itself against Housecall Pro

Problem
Every employee had a locked Excel checklist workbook per month, filled in by a call team overseas. Jobs were typed in by hand, and nobody could see which checklists were still missing before payroll.
What we built
One app for every job. Jobs sync from Housecall Pro, the call team works a daily queue, Housecall Pro data suggests answers a person confirms, and managers track completion per technician and pay period.
Result
Every job's checklist in one place, scored exactly like the old workbook, with no more typing jobs in by hand.
Read the case study

Tri-state area HVAC & generator company

An AI operations manager that watches 18 systems at once

Problem
The owner's picture of the business was spread across more than a dozen systems: jobs, money, sales, email, meetings, and timesheets. He checked each one by hand, and problems surfaced late, usually by accident.
What we built
A private AI command center connected read-only to 18 systems, including the receptionist, proposal tool, and QA/QC app we built for them. It sends a morning briefing, flags what needs attention, and texts critical alerts.
Result
One morning briefing instead of a dozen systems to check.
Read the case study

E-commerce football pad maker, supplier to the NFL

Website sales into their accounting system, without retyping a thing

Problem
Years of their books live in Sage 50, an older desktop accounting system, so switching wasn't an option. But it couldn't connect to their website, so 4,500 to 5,000 online orders a year were re-keyed by hand.
What we built
A WordPress plugin that turns each day's orders into files Sage 50 accepts, including custom shoulder pads where one product stands for 360 possible inventory SKUs. They import cleanly, as if someone had keyed them in.
Result
80 to 90% less time spent on reporting. The files are in the inbox every morning.
Read the case study

NJ business news publisher

Hundreds of profile pages from one spreadsheet

Problem
The publisher runs influencer lists across many categories, some with more than 700 people. A page for each person makes the list far more shareable and brings in more page views and revenue, but it meant building hundreds of articles by hand.
What we built
A WordPress importer. Drop in the master spreadsheet and it creates every article with the same formatting and the matching headshot, in alphabetical order so the list pages through cleanly.
Result
A list with hundreds of people goes from spreadsheet to scheduled articles in one sitting.
Read the case study

NJ business news publisher

Advertiser reports that credit the right ad

Problem
Advertisers buy slots in the publisher's newsletters and are owed a report. Mailchimp only logs clicks against links, the ad slots rotate between advertisers, and the old report script ran on one computer.
What we built
A web app any authorized staffer can open. Pick the advertiser, dates, and newsletters, and it credits every click to whoever held that ad slot on the send date, then hands back a CSV per newsletter.
Result
95% less time on reporting. Checked against a known advertiser report, all 204 rows matched.
Read the case study