Spranic wants buildings queried like ledgers

Jordi Spranger, founder of Spranic, told AI Engineer at AI Engineer NYC 2026 that every building is a balance sheet missing a live record.

Spranger said global real estate totals about 400 trillion dollars, more than equities, debt and gold combined, yet financial books close quarterly while the building underneath may not be rechecked for five years.

He demoed walking a conference hall and a kitchen with an off-the-shelf sensor that builds a semantic graph in Neo4j in about a minute, then aligns repeat walks on a time axis to query what changed object by object.

Per Kroll research on 1,500 properties cited by Spranger, 70 percent of buildings are underinsured and one in three commercial buildings carries the wrong construction class, a gap he ties to stale PDFs and notes after a site visit that backs 4 trillion dollars in US commercial mortgages.

Unlike static scans filed as pictures and PDFs, the graph keeps every walk as evidence and makes the map chatable, and connecting that memory through knowledge graphs could help robots interpret spaces across tasks according to SiliconANGLE.

The payoff Spranger pitched is hyper-personalized insurance priced on verifiable upkeep, with 35 percent hurricane premium cuts for certified buildings and 73 percent of insurers reporting fewer claims when certified, though someone still has to walk the building every quarter and keep the sensor calibrated.