Six Hotel Owners and Investors Are Skeptical AI Savings Ever Reach Their P&L

A Hospitality Net panel curated by Peter Szabo put a blunt question to six named owners and asset managers: where is AI actually saving hotel owners money, and does it reach the owner’s P&L? The answers were pointedly unconvinced. David Bowd, Owner/CEO of Salt Hotels, calls AI a business-intelligence tool rather than a strategy and is openly skeptical of “20% savings” marketing claims translating into hotel-level financials. Alex Sogno, CEO of Global Asset Solutions, says AI is genuinely useful in capital deployment and risk management but warns rising token costs mean it has to keep proving measurable value, not just capability. Andrew Carey of Newport Hospitality Group calls it “early days,” crediting AI mainly with consolidating data across platforms and freeing managers from desk work rather than moving the bottom line yet.

The sharpest distinctions came from two asset-side voices. Giuliano Gasparini, a HAMA Middle East board member, separates productivity improvement, cost savings, and actual NOI increases as three different things, and points to USALI 12 reporting standards as the test for whether brand-led AI benefits ever reach owners at all. Saar Sharon of West Ridge Asset Management doubts brands outside the US will voluntarily share AI-driven cost reductions with owners, and expects the real power shift to come from owners gaining the analytical capabilities once reserved for brands and consultants.

This lines up with a framework Ben Rafter, CEO of Hotel Equities, laid out in June: his own operator-side research found AI delivers provable returns fastest on back-office work — scheduling, procurement, revenue management, maintenance triage — while guest-facing AI stays comparatively unproven financially. Rafter’s test for any AI claim: what specific KPI moved, over what period, against which baseline, and is the gain durable or a one-time novelty. Three months and six additional owner and investor voices later, that skepticism hasn’t softened — it’s hardened into something closer to a consensus.