Workstreams | AI Hospitality Alliance

Workstream

Workstream #1 — Agentic direct booking and commerce

Representing hoteliers and travelers in conversations with major AI platforms Improving discoverability and enabling direct booking through consumer AI platforms Keeping the guest relationship with the hotel brand, during the entire customer journey Leverage existing connections board members have with AI frontier labs

#	Title	Status	Priority	Owner	Due	Waiting on	Notes
1	Conversation with Anthropic	In progress	High	GL	2026-07-16	—	Connections: Head of strategic partnership (BK) Ryan Libster, head of all industries (per GL) Onboarded Ryan Ball - sales leader GL will speak to RL to see if they'd like to kick it off Start organizing an MCP exchange with the LLMs, Antropic pass through the original prompt. Convo Jul 16 with Pete https://www.linkedin.com/in/petermonica/ Next steps: committees, meeting with hotel co's
2	Conversation with Google	Not started	Medium	BK	2026-07-23	—	BK will kick off
3	Conversation with X	In progress	Medium	BK	2026-07-23	—	Convo started with GTM: Scott Ryan, Bryan Feuling (BK) Looping in head of industries, scheduling end of Jul after their meeting with M
4	Conversation with OpenAI	Not started	High	Ira	2026-07-26	—	Sierra AI - CEO is a chairman of the board of OpenAI Connections: BK, MM, GL, RV Scott Rosecranse - GL will reach out Tyrone is no longer with OpenAI
5	Conversation with Perplexity	Not started	Medium	Ira	—	—	—

Workstream

Workstream #2 — Standards & Technical Guidelines

Dataroom

Resources: https://www.ietf.org/ and https://www.linuxfoundation.org/ https://machalliance.org/ analogy Connections: GL -> Linux foundation Developing open interoperability guidance, technical standards, and implementation best practices that help hospitality systems work together more effectively. Building blocks for the hotel use case Libraries, SDKs, (ie if for security or hospitality or reservation lookups - libraries, developer blocks, components to use take (pick them, plug them in and ship). Foundational developer tooling.

#	Title	Status	Priority	Owner	Due	Waiting on	Notes
1	Definition of Direct Booking	In progress	Medium	Ira	2026-07-26	—	https://docs.google.com/document/d/1eYCWi9bbVKakPp8L29oVCBN2m4vDwrx32EDXybTqk-o/edit?tab=t.0#heading=h.xqfjoskgkb9d Co-brand with AHLA
2	Protocol working groups (agents, MCP, ACP, A2A, etc.) for adapting open standards to hotel use cases	In progress	Medium	Ira	2026-07-30	—	Within Linux or IETF? https://developers.openai.com/commerce https://ucp.dev/latest/specification/overview/ SDK: https://strandsagents.com/?trk=66100455-ac4d-44e9-9d9a-6e8bfed444f1&sc_channel=display+ads&utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=openai-s-most-powerful-ai-is-here-but-not-for-everyone&_bhlid=91f880c999b7770421e1462c71f7a8c22988dac4 Benoit Hediard, founder of Globetrotters.ai (see email) Darko - Sirma (make sure to reach out) https://developers.googleblog.com/announcing-the-agentic-resource-discovery-specification/ https://portal.agenticcoders.dev/ https://www.affiliatecoc.org/code https://strandsagents.com/?trk=0193dc1a-4d7b-4d5a-a23b-660fbc313137&sc_channel=display+ads&utm_source=www.therundown.ai&utm_medium=newsletter&utm_campaign=apple-takes-openai-to-court&_bhlid=679f5fe075ec939bad5f5925c0e09e0e06082122
3	Define and sort order categories of problems to be tackled by AI	In progress	Medium	Ira	2026-08-14	—	https://aihospitalityalliance.com/ai-use-cases Potentially work with AHLA, HEDNA, HSMAI, HFTP as an industry effort to collect use cases for all categories, aggregate, analyze and publish
4	AI readiness framework: how to access for your company	In progress	Medium	Ira	2026-08-30	—	Khaked Koubaa to lead potentially https://www.linkedin.com/in/koubaak/ Darko - Sirma (make sure to reach out)
5	Middleware orchestration layer blueprint	In progress	Medium	Ira	2026-09-30	—	Allows to swap tools if something new comes out (which is often). Convo with RS. Engage Sirma (see Darko). https://portal.agenticcoders.dev/ Property attributes: Amadeus use case (300 attributes in CRS - the system has an ability to hold ~300 property attributes). What other CRS companies are doing? standard is 12, typically not including neighborhood and destination information. Needs to be structured. Integrated with all front-end channels through the orchestration layer.
6	Scalable AI platform architecture blueprint	In progress	Medium	Ira	—	—	Convo with RS Engage Sirma (see Darko email) https://searchengineland.com/ai-decision-layer-agentic-commerce-481862 https://searchengineland.com/author/benu-aggarwal - All Others https://martech.org/the-cms-is-becoming-the-ai-operating-system-for-brands/ https://portal.agenticcoders.dev/
7	An open-source interoperability map across major PMS & CRS APIs	In progress	Medium	Ira	—	—	HotelLogic project: https://www.hotellogic.co/index Collab: Mews, Oracle, Apaleo, Cloudbeds, RMS Consider: Semantic layer. A universal “skill” in between all different systems. If you’re an agent - you don’t need to hit openAPI but can go directly to webhooks or MCP. (how does it know? MCP registry for hotels?) Potentially workgroup within Linux foundation to create an open source API semantic layer https://developers.openai.com/commerce https://ucp.dev/latest/specification/overview/
8	A shared data-cleaning blueprint for AI-ready systems and data dictionary	In progress	Medium	Ira	—	—	Convo with GS Collab: ireckonu, Hapi, RS, GS Engage Sirma (see Darko email)
9	AI visibility measurement criteria and KPIs	In progress	Medium	Ira	—	—	Collab: Curacity, Milestone, Cendyn Benoit Hediard, founder of Globetrotters.ai. Can we measure look-to-book properly (machines vs humans)? What can we currently track in GA4? What KPIs matter? Property attributes: Amadeus use case (300 attributes in CRS - the system has an ability to hold ~300 property attributes). What other CRS companies are doing? standard is 12, typically not including neighborhood and destination information. Needs to be structured. Integrated with all front-end channels through the orhestration layer.
10	AI Impact by cost & revenue bucket	Not started	Medium	Ira	—	—	Prioritized use cases by AI, based on outcomes of the AI use cases knowledge base gathered. Convo with RS. Prop level role Call center Corp labor
11	Tokens per guest - consider introducing as a new KPI	Not started	Medium	Ira	—	—	See Terence Ronson article https://aihospitalityalliance.com/insights/enterprise-ai-token-economy https://aihospitalityalliance.com/insights/hitec-tokenmaxxing-question-for-vendors
12	Linux Foundation convo	Not started	Medium	Ira	—	—	AWS is a founding partner Connection: GL
13	Google's Agentic Resource Discovery specification	Not started	High	Ira	—	—	Google has published an open specification — ARD — for how AI agents discover, verify and safely connect to tools across organisational boundaries. For hospitality, it is the standard that finally lets a booking agent find your property's live inventory without going through an OTA aggregator. https://developers.googleblog.com/announcing-the-agentic-resource-discovery-specification/

Workstream

Workstream #3 — Governance & Responsible AI

Dataroom

Guidance on privacy, trust, transparency, staff impact and sustainability Model AI clauses that protect hoteliers in vendor and intermediary contracts Responsible deployment: GDPR, opt-in, multilingual nuance, human handoff Resource: https://aihospitalityalliance.com/insights/governance-by-design-not-committee

#	Title	Status	Priority	Owner	Due	Waiting on	Notes
1	Auth & Governance layer inside middleware	Not started	Medium	Ira	—	—	Governance: core Microsoft capabilities. Azure authentication layer. Connection: SO'F, RS Engage Sirma (see Darko email)
2	Legal aspect, policies, etc	Not started	Medium	Ira	—	—	Legal aspect, policies, etc. Make it easier for industry to navigate. Interpretations of new regulations, etc. Connection: LC, Oracle
3	Controls to apply to AI-driven functionality	Not started	Medium	Ira	—	—	Blueprint for a controlled approach to AI implementations (i.e. for an RM agent - what are the guardrails, etc.) Connection: LC, Oracle
4	Risks, safeguards, especially when scaling to a number of properties	Not started	Medium	Ira	—	—	It’s very easy to create a sexy demo that lives on your computer. But scaling it to 100 hotels or more - it’s not just an extension of the demo or a pilot. It’s a massive transition to IRL. And the answer changes by the day - how do you achieve this? Connection: MS Engage Sirma (see Darko email)
5	UN: Global Dialogue on AI Governance (represent hospitality)	Not started	Medium	Ira	—	—	https://www.un.org/global-dialogue-ai-governance/en Khaled Koubaa to lead potentially https://www.linkedin.com/in/koubaak/