Void analysis tenant shortlists for vacant retail spaces

Build an agent workflow that runs void analysis for commercial real estate brokers and landlords. The workflow should poll Airtable for a new record in my Vacant Properties table (each row is a vacant retail space with address, square footage, and asking rent). When a new row appears, kick off a void analysis for that property.

Trigger: Airtable poll on new_record for the Vacant Properties table. On fire, pass the record's address, square footage, and asking rent to the agent.

For each new vacancy, the agent should use Placer.ai to profile the location and generate a ranked list of chain tenants that would thrive there:

1. Resolve the address with Placer's Search for Placer Entities. If no matching venue is returned at those coordinates, use Create Custom POI to define a radius-based custom POI at the property.

2. Run the Trade Area by Drive Time Report and the Trade Area Demographics Profile Report against that entity to characterize the surrounding population (income, age, household composition, daytime vs residential density).

3. Identify the nearest few anchor venues via Search for Placer Entities, then call Visitor Journey - Favorite Chains and Visitor Journey - Favorite Places on those anchors to see which chains already draw this shopper base.

4. Use Ranking Index - Top Ranked to sanity-check that the shortlisted chains overperform in comparable demographics, not just in absolute foot traffic.

5. Run Search for Placer Entities again within a small radius (default 2 miles) around the vacant property to flag cannibalization risk: existing nearby locations of any chain on the shortlist. Skip any chain whose closest existing location is under 2 miles unless the drive-time trade area is large enough to support two locations.

Filter rules: skip any chain that Placer flags with data-quality issues (flagged entities in search results), and drop chains where reports cannot be generated. Do not recommend chains that fail either check.

Output: Create a new Google Doc titled with the property address (for example, "Void Analysis - 1420 Main St, Austin TX"). Populate it via Batch Update Document with a ranked list of 8 to 10 candidate chains. For each chain include a two to three sentence rationale citing (a) trade area demographic fit, (b) foot-traffic evidence from nearby anchors, (c) cannibalization risk, and (d) a suggested pitch angle the broker can lead with.

After the doc is created, use Airtable's Update Record to write the Google Doc URL back to the source row so the broker sees it inline.

Important Placer.ai gotcha: report endpoints use an async cache pattern. A 202 IN_PROGRESS response means the report is generating; retry the same request body after roughly 5 seconds until you get a 200. 202s do not count against the weekly quota. Bake this retry loop into every report call (Trade Area by Drive Time, Trade Area Demographics, Visitor Journey, Ranking Index).

Do not run this for updates to existing rows, only for new_record events. If Placer returns no visitation data (204) for a nearby anchor, silently drop that anchor and continue with the others rather than failing the run.