A follow-up to “Why your AI travel agent can dream up the perfect trip but can’t book it”

I published that article on December 29th. Two weeks later, back at my desk with inbox overflowing, I find that Google has launched the Universal Commerce Protocol, co-developed with Shopify, Walmart, Target, Etsy, and Wayfair, and endorsed by twenty-plus others including Visa, Mastercard, and Stripe (apparently I wasn’t the only one busy during the holidays). Their roadmap explicitly mentions “Travel, Services” as upcoming verticals.

My LinkedIn feed is full of takes. Everyone in agentic commerce has an opinion. So here’s mine: this changes things, and travel has a narrow window to shape how.

Why UCP is different from what came before

When OpenAI and Stripe launched the Agentic Commerce Protocol (ACP) last year, I was skeptical. ACP was maintained by two companies, initially live only in ChatGPT, with no stated ambition beyond retail checkout. It felt contained.

UCP is different in ways that matter.

First, there’s an open contributor invitation. Google’s documentation states: “We invite developers, businesses, and platform architects to join us in building the future of commerce.” They’ve published a “Become a UCP contributor” call to action. This isn’t “implement our spec”; it’s “help us build the spec.”

Second, vertical expansion is explicit. Travel isn’t an afterthought; it’s in the roadmap. The GitHub roadmap lists “New Verticals: Applications beyond Shopping (e.g., Travel, Services)” as a future enhancement. They’re thinking about us, they just haven’t built the schemas yet.

Third, transport agnosticism. UCP supports REST, MCP, and A2A. It’s not locked to one AI platform’s stack, which means the same capability schemas can serve Google surfaces, Anthropic’s Claude, and whatever emerges next.

Fourth, broader coalition from day one. ACP was OpenAI plus Stripe. UCP was co-developed with Shopify, Etsy, Wayfair, Target, and Walmart, and endorsed by payment networks, processors, and major retailers. That’s infrastructure players convening, not just one AI company extending its reach.

Is UCP truly “open” in practice? Probably not entirely (it’s still Google-led, initially live on Google surfaces). But the structure is designed for ecosystem participation in a way ACP wasn’t.

What travel needs that UCP doesn’t have yet

UCP’s own roadmap is transparent about what’s missing: product discovery, loyalty integration, and vertical extensions beyond retail. Their current schemas are retail-focused (line items with SKUs, quantities, shipping addresses), and they know it.

Travel doesn’t work that way.

When an agent books a trip, it’s not “add to cart, checkout.” It’s multi-service coordination where a flight, two hotels, a car rental, and three activities have interdependent availability and pricing. It’s tour operator intermediation, where regulations require a licensed merchant of record for dynamically packaged offerings (the “package this for me” request that no retail checkout can handle). It’s compliance with the Package Travel Directive in Europe, ATOL in the UK, and consumer protection requirements that vary by jurisdiction. It’s fulfillment that means check-in dates, boarding passes, and vouchers rather than tracking a package.

And it’s discovery with travel-specific steering: not “show me products matching keywords” but “find me a romantic getaway in Italy within budget, considering seasonality, visa requirements, and my airline loyalty status.” UCP’s discovery capability is explicitly on the roadmap but not yet specified. Someone needs to specify it for travel.

These requirements won’t emerge from retail commerce assumptions. They require people who’ve spent years in travel distribution to define them.

What I’m doing about it

I’m responding to Google’s invitation.

Over the past 2.5 years at Chain4Travel, we’ve been building travel distribution infrastructure on Camino Network: 200+ partners including Lufthansa, TUI, MTS Globe; twelve releases of production message schemas for accommodation, transport, and activities; a B2B messaging protocol handling the whole search to book funnel; participation in ECB digital euro and EU digital identity efforts so we know that showing up early to specs and standards matters. Pablo Castillo, Sam Jaarsma, Anke Hsu, Helena Vossen, Alexander Koch and the team bring many decades of combined travel tech experience; mine includes tenures at companies like Sandos Hotels and Alpitour, wrestling with the same coordination problems from the supplier and tour operator side.

The pitch to Google isn’t “use our infrastructure instead.” It’s: “You’ve invited vertical expertise. I have some, and in Camino you can find hundreds of years of expert experience. Let’s build dev.ucp.travel so that when AI agents book trips, the schemas reflect how travel actually works.”

Where blockchain fits

Spawning a CMAccount (an on-chain identity for agents) is a novel UX paradigm. It’s an honest attempt at establishing how humans delegate to autonomous systems with enforceable spending limits and cryptographic accountability. Our children (or more probably given the current adoption curve ourselves in 5 years?) will interact with AI agents as naturally as we interact with apps; they/we’ll understand delegation in ways we’re still learning now. But in the short term, requiring consumers to navigate wallet-based flows risks adding friction.

So I think about this in two timelines. For consumer payment, let UCP handle it (Google Pay, Visa, whatever rails people already trust: this is the pragmatism the industry wants). Where blockchain infrastructure provides clear value is B2B: settlement via signed cheques, atomic delivery-versus-payment for supplier transactions, on-chain audit trails for dispute resolution. The blockchain layer works best as invisible infrastructure enabling trust between businesses, not as a consumer-facing hurdle.

The window is now

UCP’s roadmap says travel is coming. The question is whether travel expertise shapes that standard or we react to it after Google ships something based on retail assumptions.

I’ll be contributing. If you’re a travel company who wants to be part of that conversation, reach out and let’s do it together. If you’re on Google’s commerce team and want to see what twelve production releases of travel schemas look like, the GitHub is public and my calendar is open.

Standards get set with or without us. I’d rather be in the room.

Mik Ruberl has spent 25 years building technology, the last decade at the intersection of travel and tech: CTO at Sandos Hotels & Resorts, Head of Architecture at Alpitour Group, now VP Product at Chain4Travel where he leads work on digital payment and identity infrastructure, and agentic AI for travel distribution.

Originally published on LinkedIn.