Free AI Workshops
AIHA Academy — 8 free live AI workshops for hospitalityFrom AI basics to building your own agents and appsLive online · One hour each · Recordings included
Register free
Back to Insights

2026-10-18 | Wei Manfredi

Your Next Customer May Never See Your Marketing. But You Can See What Agents Are Seeing

In Part 2 of the series, Wei Manfredi uses a Denver hotel search to show how buried policies, hidden fees and facts locked in images push AI agents toward other hotels, and proposes a platform-neutral 30-day test of what agents can actually see.

This insight summarises Your Next Customer May Never See Your Marketing. But You Can See What Agents Are Seeing, by Wei Manfredi, Senior Vice President of AI and Architecture at IHG Hotels & Resorts and an AIHA advisor, published on Substack on 18 October 2026. It is Part 2 of the series that began with Your Entire Marketing Stack Is About to Be Rewritten by Agents.

The Denver test. A customer asks an AI agent for a refundable hotel near an 8 a.m. meeting in Denver: under $250 all-in, parking included, a gym that opens by 6. The hotel has a gorgeous website and an attractive nightly rate. But the cancellation deadline is buried on a policy page, the parking fee is hard to find, the gym hours appear only in a photo of a sign, and the final price becomes clear only near checkout. A person might call the hotel. An agent might check third-party sites, ask the customer to loosen a requirement or recommend somewhere else.

The hotel did not lose because its brand is weak. It introduced doubt at the moment the customer asked someone else to make a confident choice. "Your next customer may never see your homepage. But your next customer's agent will still need the facts."

The plumbing is still moving. OpenAI and Stripe introduced the Agentic Commerce Protocol, Google introduced the Universal Commerce Protocol in January 2026, and Visa and Mastercard are developing ways to recognize trusted agents and support authorized payments. Early in-chat checkout experiments ran into commercial and operational friction, and it is too early to assume a single winner or a universal checkout experience.

Start with what you control. Manfredi treats that uncertainty as a reason to begin with accurate products, transparent terms, reliable availability, usable policies and a business that delivers its claims, before investing in the next AI commerce platform.

Checks drawn from the Denver example:

  1. Is the cancellation deadline stated in plain text next to the rate?
  2. Are parking and other fees part of a clear all-in price?
  3. Are amenity details such as gym hours written as text, not only shown in images?
  4. Does the price an agent finds early match the price at checkout?

The 30-day test. The rest of the piece sets out a practical, platform-neutral stress test, starting in Week 1 with seeing what agents see and finding every point where an agent has to guess. The full plan continues on Substack.

Read the full article on Substack →

Stay Updated

Enjoyed this insight?

Get more AI hospitality research and trends.

For Contributors

Submit article for consideration

Share an article, perspective, or industry analysis for possible inclusion in the Alliance Insights section.