The AI Shortlist Index · September 2026 · preview edition
Every month we ask ChatGPT, Perplexity, Gemini and Google AI the questions a property manager asks when choosing software, three times each on separate days, and record who gets named. This is the answer.
This is the preview edition. One engine, ChatGPT, has answered 25 prompts in each of three categories, 75 runs in all; Perplexity, Gemini and Google AI Overviews are shown as missing until their runs exist. Every number here is one run deep, where the protocol calls for three across at least two days, so read the rankings as a first reading rather than a result.
The engine mostly reads vendors' own websites. Across the twenty most-cited domains in each category, 73% of citations went to vendor-owned sites and under 3% to forums. The thing that decides whether a vendor gets named is largely the vendor's own site, not a review platform.
The leaders are Second Nature in resident benefits at 86%, ButterflyMX in access control at 83%, and Property Meld in maintenance and vendor networks at 91%. In access control every one of the 19 vendors in the database was named at least once; in resident benefits, 19 of 20, so the engine spreads its answers rather than repeating one shortlist.
Bundled resident-facing services (renters insurance compliance, HVAC filter delivery, credit reporting, rewards, move-in concierge, pest control) sold by a property manager as a monthly package, delivered and billed through a platform.
25 runs · 3 of top 3 named · source map
Keyless entry, smart locks, video intercoms, gate and amenity-door access, and the credential platform that ties them to move-ins, move-outs, vendors, and self-guided tours across an apartment community.
25 runs · 3 of top 3 named · source map
Platforms that intake, triage, and coordinate resident maintenance requests, dispatch in-house technicians or a managed third-party vendor network, and track work orders, approvals, invoices, and vendor compliance.
25 runs · 3 of top 3 named · source map
| Engine | Status | Model | Runs | Note |
|---|---|---|---|---|
| ChatGPT | degraded | – | 75 | ok in 3 of 10 categories |
| Perplexity | missing | sonar | 0 | no runs recorded |
| Gemini | missing | gemini-3.8-flash | 0 | no runs recorded |
| Google AI Overviews | missing | – | 0 | no runs recorded |
Each category runs 25 prompts written the way a property manager asks, across four engines, three times each. A vendor’s mention rate is the share of runs that named it; rank is by mention rate; consistency shows whether an engine flips between runs. The methodology states every rule, the prompt library is public and versioned, and the publication policy covers corrections and re-checks. Editions are archived at the archive and method changes at the Index changelog.
Paying WebPossible never affects Index placement. The Index measures what AI engines say. Shortlist clients buy the work that earns a position, not the position.
Published by WebPossible. The practice that acts on these numbers: Shortlist.
Seven days. Twenty-five prompts across four engines. The cause behind your row and a 90-day fix. Ryan replies with an agreement and an invoice; no call required.