●selected work
four engagements,
four verticals.
commercial real estate · operations
automating maintenance across a commercial portfolio
client: Las Vegas commercial property manager (NDA) · 42 properties · status: live · stack: ShoWork
the challenge.
a Las Vegas commercial property manager runs maintenance for 42 properties through outside vendors. requests went out by email, text and phone, updates came back the same way, and proof that a job was done depended on someone remembering to send a photo. the property manager spent the day chasing status instead of running the portfolio.
what we built.
we put the whole portfolio on ShoWork, our own product. the property manager opens work orders for any of the 42 properties and assigns the vendor. crews send voice notes and photos from the ShoWork Field app, in English or Spanish, and check in and out with GPS. ShoWork turns that into verified reports and invoices, so the proof arrives with the work instead of after a follow-up call.
2
vendor companies, one system
en · es
crews report in either language
read the commercial real estate work →
real estate · sales
ai lead scoring for a luxury real estate brokerage
client: las vegas brokerage (NDA) · engagement: 9 weeks · stack: DealWork + custom scoring model
the challenge.
a Las Vegas luxury brokerage was generating thousands of inbound leads per quarter from listing portals, social, and referrals. agents were chasing volume instead of quality, burning the first hour of every day on dead-end follow-ups while the actual buyers got cold.
what we built.
we deployed DealWork and trained a scoring model on the brokerage's two years of historical close data. the model now scores every new lead in real time on price band match, behavior signal, and listing match, surfacing the high-intent buyers to the right agent within minutes. auto-drafted follow-ups handle the rest.
my top agent closed two $2M+ deals in the first quarter that the old system would have buried under cold leads. the model paid for itself before the trial ended.
patient intake automation for regional clinics
client: multi-site clinic group (NDA) · engagement: 11 weeks · stack: custom intake platform + insurance verification api
the challenge.
a regional healthcare network with 12 clinics was losing 2+ hours per front desk per day on paperwork. patients were filling out the same forms by hand at every visit. insurance was being verified manually, and 8% of claims were being denied for trivial intake errors.
what we built.
we built a patient-facing intake portal with ai form pre-fill from prior visits, real-time insurance verification through a payer api, and an LLM that flags incomplete or contradictory entries before they hit the EHR. front-desk staff get a clean, validated record on every check-in instead of a scanned PDF.
our denial rate dropped to under 1% in three months. front-desk turnover stopped almost overnight. the job is finally a job, not a punishment.
predictive scheduling for a commercial general contractor
client: commercial GC (NDA) · engagement: 16 weeks · stack: custom scheduler + weather/supply ml model
the challenge.
a commercial general contractor running six concurrent projects was constantly over budget and behind schedule. their PMs were updating gantt charts by hand on fridays, but the schedules were stale by monday morning. weather, supply chain delays, and crew availability were treated as surprises every time.
what we built.
we built an ai scheduler that ingests live weather forecasts, supplier etas, and crew availability, then continuously re-plans the timeline. PMs see a live 'schedule of record' with drift alerts when reality deviates from the plan. daily reports are auto-generated for owners and lenders.
for the first time we know on tuesday what's going to bite us on friday. that alone changed how we run the business.