This insight summarises Budget the Meter, Not the Machine: Planning for 2027 When AI Runs on Usage, an interview with Terence Ronson, Founder and Managing Director of Pertlink Limited and chair of the HFTP AI Collective's AI Education & Training Subcommittee, published on Hospitality Net in October 2026.
AI comes with a meter. Traditional hotel technology behaves like a fixed asset, with licences and annual renewals. AI spend rises with use, so every guest interaction adds cost. For 2027, the hard part is no longer adopting AI but measuring, governing and paying for it.
Token Cost Per Guest (TCPG). Ronson's core metric is the AI cost of serving one guest across the journey, from a pre-arrival question to a post-stay review. "If you can't tell your owner what AI costs per guest, you don't have an AI budget. You have a hope."
Budgeting moves:
- Ring-fence a workforce line for training, role-specific guidance and change management. "Buying AI without training your people is like installing a grand piano in a hotel with no pianist."
- Put AI on a rolling forecast, monthly or quarterly, with a defined trigger for when usage and cost exceed plan. "Flexibility without a trigger is just drift."
- Budget for external forces, including guests shortlisting hotels inside AI assistants and, in Asia-Pacific, data-localisation rules that limit which vendors a group can use.
Measure value, not activity. Ronson separates efficiency, cost saving and value creation, each with its own measure. "A busy chatbot is not a profitable chatbot." He recommends starting back-of-house, where data is cleaner, guest risk is lower and returns are easier to prove, then using the savings to fund guest-facing tools: "Boring AI pays for exciting AI."
Governance belongs in the budget. Citing Moffatt v Air Canada and a Munich court ruling on inaccurate AI-generated hotel information, he recommends:
- A Decision Envelope that defines what AI may decide alone, what it must escalate and what it may never do.
- An AI Override Rate that tracks how often staff correct or reverse AI decisions. "If nobody ever overrides the AI, either it's perfect or nobody's watching."
Why this matters for hotels. By 2030, most hotels will have access to similar AI models and tools. The difference will come from judgment, data foundations, operational discipline and turning efficiency gains into more staff time with guests.