A Hotel Investor Used AI to Read Guest Reviews Before Acquiring a Property — and Found the Elevators Problem the Inspection Missed

Drew Bridges, a principal at hotel investment firm JMI Realty, described on the Hospitality Daily podcast — in an episode built around a deal the Wall Street Journal previously spotlighted — using AI to analyze thousands of guest reviews before acquiring a property, surfacing capital problems a standard inspection wouldn’t catch. At the Hilton Garden Inn Austin Downtown, review analysis flagged the elevators as a genuine capital-expenditure issue, while AC complaints turned out to need only preventative maintenance rather than full replacement — a distinction that changes how a deal gets underwritten.

Bridges named his workflow tools directly: Avion Hospitality, ElevenLabs, Twilio, and Claude Code, alongside other properties discussed in the conversation — Hotel Van Zandt and Hotel Indigo Nashville – The Countrypolitan. He also uses the same review-and-data analysis to vet prospective third-party management partners before signing them, extending AI’s role from acquisition diligence into ongoing operator selection. He drew a clear line: AI is moving into deal diligence, but the human service the reviews are analyzing still can’t be replaced by the tools analyzing it.

That’s the same distinction Customer Alliance drew in its own launch framing when it rebuilt its guest-feedback platform around AI in August, in an announcement covered by Hospitality Net. The relaunch, branded “guest feedback intelligence,” now serves more than 5,000 hospitality businesses — CEO Steffen Schmickler described turning feedback “into insights, priorities, actions, and measurable improvements,” not replacing the judgment that acts on it. Bridges’ use case is the acquisition-stage version of the same shift: guest reviews as a structured data source for capital decisions, not just a reputation-management input.

For asset managers, the practical takeaway is narrower than “use AI in diligence” — it’s that review text, read at volume, can separate a genuine CapEx problem from a maintenance non-issue before either shows up in a purchase price.