Hotel tech vendors have spent two years adding AI to their products. A new survey suggests most hotels are getting their AI somewhere else.

Research conducted for Destination AI and reported by PhocusWire on 14 October 2026 found that 72% of hotel brands get their AI capabilities from general-purpose tools such as ChatGPT, Gemini and Copilot. Forty-one percent build or customise their own solutions in-house. Only 33% adopt AI features from their existing hotel software vendors, and 16% buy from AI-first startups. The survey covered 107 hotel company leaders in August and September.

If you sell hotel software with AI in it, that is the competitive set you are actually in. Your AI feature is not only competing with the rival vendor's AI feature. It is competing with a ChatGPT tab the revenue manager already has open. This post compares the two the way a hotel buyer does, then sets out how to position, price and prove a vendor AI feature against it.

What the survey says about vendor spend

The spending question matters more than the adoption headline. Asked how AI is likely to affect their spending with hotel software providers over the next 12 months, 27% of hotel leaders expect to buy somewhat or significantly more, 33% expect to spend about the same and 17% plan to spend somewhat or significantly less. So AI is not yet shrinking vendor budgets for most hotels, but it is not automatically growing them either. The feature has to earn the money.

The same pattern shows up in chain data. h2c's AI Opportunity Study 2026, reported by TravelDailyNews, found that 91% of participating hotel chains already use AI but only 13% report measurable return on investment so far. External AI tools, including ChatGPT, account for 60% of reported day-to-day usage, against 27% for AI embedded within vendor systems. There is a counter-signal worth noting: 31% of chains now primarily rely on vendor-provided AI solutions, up from 21% in 2025. We covered the wider findings in six myths about selling AI to hotel chains.

General-purpose AI vs a vendor AI feature: how hotels weigh them

A hotel buyer rarely frames it as a formal comparison, but these are the questions they are answering in their head when they decide whether your AI feature is worth paying for.

The price row is the one vendors underestimate. OpenAI's ChatGPT Business overview lists Standard seats at $25 per user per month on a monthly plan and $20 per user per month billed annually, with a two-seat minimum, and states that OpenAI won't train on the workspace's data. That is the anchor a general manager carries into your pricing conversation, whether or not they say it out loud.

The data rows are where the vendor wins. A general-purpose tool only knows what someone copies into it. A feature inside the PMS or RMS can see the booking pace, the guest history and the rate plan, and can act on them without anyone exporting a spreadsheet. If your AI feature does not use that advantage, the buyer is right to compare it with a $20 seat.

What the gap tells you

Read together, the numbers say three things about the market you are selling into.

• Hotels have already decided AI is useful. 63% of respondents in the Destination AI survey use AI regularly in their work. You do not need to sell the idea of AI.

• They have not decided it needs a vendor. Bottom-up use dominates: 32% said bottom-up innovation drives their AI success and 34% cited experimentation culture, against 24% with a dedicated AI lead and 14% with executive mandates. Staff found ChatGPT on their own.

• Almost nobody can prove ROI yet. 35% do not measure AI return at all and only 14% use defined performance indicators measured before and after deployment. A vendor that brings the measurement brings something the hotel lacks.

That last point is the opening. Only 13% of organisations in the Destination AI survey said they have a clear strategy, formal governance and scaled execution across functions. Most hotels are running AI as a set of individual experiments. A vendor feature that arrives with a baseline, a target and a report fits into a gap the general-purpose tools leave open.

How to position, price and prove a vendor AI feature

1. Position on the work it does inside your system

Do not lead with the model or the word AI. Lead with the job that only your product can do because it sits on the hotel's data: the rate it recommends from live pace, the guest message it sends from the actual reservation, the report it writes without an export. If a task could be done just as well by pasting data into ChatGPT, it is not your differentiator. The distinction between AI built into the workflow and AI bolted on is the same one we drew in AI-native vs AI-enhanced PMS positioning.

2. Price it against the seat the buyer already pays for

Assume the buyer will compare your add-on price with a per-seat general-purpose tool. If your AI feature costs more per property per month than a few ChatGPT seats, the extra has to be justified by an outcome, not by access. Options that hold up better than a flat AI surcharge include bundling the feature into an existing tier, pricing it on the outcome it moves, or offering it free for a defined pilot with a measured target. Our guide to hotel tech pricing models covers the trade-offs of each.

3. Prove it with a before-and-after number

Since 35% of hotels in the survey measure nothing and only 14% measure before and after, bring the measurement yourself. Agree one metric before go-live, such as minutes per guest reply, upsell revenue per arrival or forecast accuracy. Record the baseline, then report the change at 30 and 90 days in the hotel's own numbers. A general-purpose tool will never send the general manager that report. You can.

4. Make it safe for the people already using ChatGPT

Staff who use general-purpose AI are not your enemy. They are your most AI-ready users. Show them what your feature does that their chat window cannot, such as acting in the system or using guest history, and let them keep the general tool for everything else. Positioning the feature as a replacement for ChatGPT invites a comparison on breadth that you will lose.

The takeaway

Your AI feature is not competing with the other vendor's AI. It is competing with a $20 seat the hotel already has.

The Destination AI survey shows hotels adopting AI fast but mostly through general-purpose tools, with only a third getting it from their software vendors and most unable to measure the return. Vendors that win this round will position AI on the work only their system can do, price it with the per-seat anchor in mind and arrive with the before-and-after number hotels are not producing themselves.

If you are not sure whether your AI story is helping or hurting your pipeline, a free revenue leak audit will show where deals are stalling and why.

Frequently Asked Questions