- contact@verticalserve.com
How InsightPulse products get built — the AI layers of Home Care 360 for home-based care and Storage 360 for self-storage, the entity graph behind company data, corporate hierarchy and the failures that shaped the rules.
The AI layer, explainable street rates, the rent-increase copilot, the lakehouse architecture and a 90-day pilot plan — in one PDF. See the product.
Lost referrals, the staffing copilot, retention, star ratings on CMS Care Compare, the revenue cycle and a 90-day pilot plan. See the product.
3,044 referrals declined for no staff, $21.8M of expected revenue: how Home Care 360 prices the leak, turns it into 587 hires by branch and discipline, and ranks a caregiver for every open visit — with the copilot's score worked through.
Aides turn over at 53% a year. Eight signals already in the schedule and payroll, a calibrated 90-day leaving score, and a stay-interview list with the drivers and the dollars.
Every branch against the CMS-certified agencies in its county, from the public Care Compare files: behind on quality in 31 of 52 counties, and a patient-survey gap that is specific and fixable.
$13.8M denied in a year, $8.4M of it preventable by the app's estimate. Scoring pending claims, projecting every authorization to exhaustion, and catching LUPAs under PDGM before the period ends.
A 25-day median stay, a 73-day average and live discharges near 40% after six months. Scoring every hospice patient for a long stay and a live discharge, for the IDG.
For CIOs: nine linked entities and one definition per metric, an AI layer written back as tables, 217 read-only data contracts, a cache and an audit log — and what is synthetic, generated or live.
A dashboard says occupancy is 91.3%; it does not say which store to fix this morning. The six-part anatomy of an insight, the nightly loop that builds 251K of them, and the briefs and alerts that put them in front of the right person.
Occupancy-driven moves, an aged-vacancy push, a competitor guard and a ±10% cap — how 3,506 street-rate moves worth about $1.85M a year each carry a reason a store manager can repeat.
A copilot that prices each lease against its move-out risk, fairness caps written as rules, and a simulator that shows $151K of kept uplift becoming $129K net after expected move-outs.
Like-for-like by store and unit type: our index against each operator, the gap to the cheapest store, fair share and rate-move alerts — and how the index guards the AI's street-rate moves.
For CIOs: data that stays in your Databricks workspace, 134 reviewed read-only data contracts, a nightly AI layer, a result cache that keeps cost flat, and what is not in the current release.
Insurance distribution consolidates faster than most datasets refresh. Four broker records were already stale within a week of building — here is the signal that caught each, why membership is not ownership, why announced is not closed, and the coverage numbers we publish instead of our best ones.
Look a company up by domain, ticker or LEI — with no prior lookup step. Corporate ownership trees, brand assets and insight cards through a single REST call, with per-field provenance.
More posts are on the way — entity resolution at scale, what SEC Exhibit 21 does and does not tell you, and how we model announced-but-unclosed transactions. Want a topic covered? Tell us.