What Google’s Universal Commerce Protocol Means for Travel

Google’s newly released Universal Commerce Protocol (UCP) is the first real agentic commerce standard that can handle the complexities of multi-day tours, including availability, dynamic pricing, multi-stage booking, and post-purchase modifications.

Given Google’s prevalence and existing forays into Travel, end-to-end planning and booking inside Google is a near-term reality, not a future concept.

Read on to learn:

• What UCP is
• Why it’s capable of handling the nuances of multi-day tours, as well as the specific idiosyncracies of your business
• What it could mean for distribution
• And, how you can take practical steps towards readiness.

What is UCP?

UCP is an open standard that enables AI agents, platforms, merchants, and payment systems to interoperate across the entire commerce lifecycle — discovery to booking and beyond — without custom, proprietary integrations. It is both a common language and a set of API primitives (primitives being foundational building blocks i.e. capabilities like checkout, orders, identity linking) that any compliant system can use to transact with one another.

Rather than building a different integration for every distribution channel (Google, Bing, even OTA integrations, etc.), UCP lets you integrate once and then be available across all agent-enabled surfaces that adopt the standard.

Importantly, it supports secure, tokenized payments which means the tour operator remains the Merchant of Record (i.e. you keep control of pricing rules, taxes, and revenue). UCP helps you transact, but never replaces you.

What Makes UCP Uniquely Suited to Multi-Day Tours?

UCP isn’t the first or only agentic commerce protocol out there. OpenAI released ACP in Q4 last year, promising a similar “buy-in-chat” style customer experience. However, it’s capabilities were limited and the protocol only catered for simple products that can be purchased in one discreet interaction. So, what makes UCP different?

Long-lived workflows

UCP enables an AI agent to hold inventory temporarily, collect missing inputs over multiple interactions, convert a reservation into a confirmed booking, collect deposits and balances, and continue managing that booking over time.

Inventory-aware

UCP allows Operators to expose availability windows, capacity constraints, and conditional pricing logic that lets an agent reason over that information before presenting it to a potential guest.

“This tour exists, but not for your dates, budget, or group size.”

Progressive disclosure

UCP allows Operators to declare Required versus Optional fields meaning deposits can be accepted without full traveler data, with follow-up data collected post-booking.

Extensions enable custom logic

UCP’s extension model lets Operators augment core Checkout capabilities with additional fields or rules specific to their business. This means a tour operator could define custom extensions for:

• collecting specific travel documents
• group relationships (couples, families, friends)
• rooming preferences
• solo traveler surcharges

As a rule of thumb, if the data affects discovery, suitability, pricing, or post-booking changes it probably belongs in an extension.

Embedded Checkout options

Furthermore, UCP supports two checkout paths:

1. Native Checkout: Great for simple or moderately complex flows
2. Embedded Checkout: For complex, bespoke, or regulated purchases

Tours will almost always need #2 at some point. Embedded Checkout lets tour operators leverage their own booking interface when and where it makes sense, while still being natively bookable inside AI and Google experiences.

Post-booking support

After booking, a guest may choose to alter their itinerary or request an upgrade. They may cancel and request a refund. Weather conditions might alter a departure date or time. UCP standardizes Amendments, Cancellations, Refund logic, Order state changes, and Merchant-initiated updates and can therefore explain any changes that occur, trigger allowed modifications, and more broadly stay in the loop long after the transaction.

Importantly, UCP assumes that you have complex back-office systems with complex inventory and operations and the protocol mirrors these constraints. That’s a meaningful differentiator to ACP, which tries to abstract much of that complexity for ease of integration, ultimately falling short of the experience multi-day tours demands and leading to fragmented booking processes.

A Practical Example

Let’s walk through a fictitious example, grounded in principles the protocol supports. Note that since there are no live examples to refer to, I’ve applied some artistic license.

Discovery

A prospective guest asks an AI agent:

“Find a 6–8 day guided trekking tour in Patagonia in March, under $4,000 per person, suitable for intermediate hikers.”

Before showing options, the agent needs to know:

• Which merchants exist
• What kinds of commerce flows they support
• Whether they can handle complexity (multi-day, add-ons, etc.)

Merchants (Tour Operators, in this case) can advertise supported UCP capabilities (e.g., Checkout, Order), supported integration modes (Native vs Embedded checkout) and specific constraints and expectations of their commerce flow. This prevents the agent from recommending a tour operator it can’t actually transact with later.

The agent queries merchants for offers that match the traveler’s criteria:

• Destination: Patagonia
• Duration: 6–8 days
• Dates: March
• Difficulty: Intermediate
• Budget ceiling

UCP’s design intentionally avoids forcing merchants to lock inventory during discovery. Tour availability is expensive to lock down; discovery should be cheap and reversible.

Shortlist

The agent now has multiple candidate tours:

• Tour A: 7 days, $3,800, limited availability
• Tour B: 8 days, $4,200 (over budget, excluded)
• Tour C: 6 days, $3,500, fewer inclusions

The agent can compare itineraries, explain why some options were excluded and ask clarifying questions (“Is $4,100 acceptable if meals are included?”).

The prospective guest responds:

“The 7-day Patagonia trek looks good. Can you check availability for March 9 for two people?”

This is the threshold where we move from discovery to booking, but no order has been created yet.

Booking

Only after the prospective guest signals intent, does the agent invoke UCP’s Checkout capability, creating a checkout session with:

• Tour line item
• Quantity = 2
• Selected departure date

The Tour Operator responds:

• “March 9 has availability”
• Price confirmed at $3,750 per person
• Add-ons available

If constraints fail — maybe the Tour becomes sold out — the checkout session would return an error and the agent can fall back to discovery and suggest alternatives. Crucially, the journey doesn’t reach a dead end. The agent reasons, adjusts and the conversation continues.

Once the guest proceeds, they might need to select add-ons, confirm rooming preferences, or sign waivers. To do this, the Agent transitions to Embedded Checkout and displays the Tour Operators' booking interface. Events are streamed back in real-time and the Agent remains context-aware but not in direct control. This is good for Tour Operators, who retain the opportunity to reinforce buyer trust through use of their brand.

When it’s time for payment, the Agent is removed from sensitive steps and the Order is submitted deterministically. In other words, the agent cannot see or influence payment details and the Tour Operator explicitly confirms the purchase themselves.

Our guest is shown an Order ID and a confirmed booking state.

Post-booking

From here onward, the previously described Order Capability applies which handles any Amendments, Cancellations, Refunds / credits, and Status updates via webhooks.

In practice this means the agent can answer questions such as “What’s my booking status?”, explain any changes that occur, and initiate any permitted adjustments.

From Answering Questions to Executing Requests

By now we’re all aware that Search has evolved, from keyword-based discovery to large language models that synthesise information across many sources and present a consolidated answer. In travel, much like many other industries, this has collapsed the traditional funnel but it hasn’t materially changed how bookings actually happen.

LLMs could describe what a tour looked like, but they couldn’t reliably determine whether it could be booked under specific conditions. Dates, capacity, dietary requirements, deposits, and confirmation rules all lived behind workflows that were opaque to machines.

UCP changes that.

An agent can check availability, validate constraints, and determine next steps as part of the same interaction in which intent is expressed. For tour operators, that distinction matters because so much value is lost in the gap between interest and execution.

Distribution, Data, and a Shift in Leverage

Ultimately, UCP represents a new distribution channel. Merchants promote their capabilities in a JSON manifest, and agents dynamically discover and negotiate which capabilities they can use in service of the traveler’s requests.

As it stands, none of this replaces OTAs, nor does it replace operator websites, but it does change how demand is surfaced. Visibility in this channel is influenced less by page-level optimisation and more by whether your inventory, rules, and eligibility can be validated in real time. This is where data quality and infrastructure become strategic.

Brennen Bliss of Propellic put it perfectly in his NavLog newsletter:

“Universal Commerce Protocol is highly likely to give preference to data feed quality over webpage content, structured data over content depth, and product eligibility over traditional SEO factors. It’s for this reason that I recommend brands remain ready to allocate resources toward feed, API, and inventory infrastructure as this evolves. It’s a matter of adopting agent-first distribution or losing leverage permanently as we continue our charge into agentic travel research and booking.”

He also highlights the downstream challenge this creates for measurement and attribution, noting that traditional channel reporting will struggle to explain performance in an agent-led world. Propellic has been exploring media mix modelling as one potential way to regain clarity which is an area worth following as this ecosystem develops.

Structured Content Becomes Foundational

Once AI systems are expected to act on intent, the nature of content changes. To move from interest to action, AI systems need structured, machine-readable representations of what you sell and how it can be fulfilled.

For a multi-day tour operator, that means treating key parts of the product not as prose, but as data:

• departures and capacities
• pricing and deposits
• guarantees versus “on request”
• constraints like dietary support, age limits, or activity level
• rules about when and how a booking can progress

This is not new information. What’s new is where it needs to live and how it’s used. Historically, much of this logic has been embedded in booking engines, internal systems, PDFs, even human workflows.

UCP assumes that AI systems will increasingly query this logic directly. Not to replace your booking engine, but to determine whether and how far a booking can proceed before a human is involved. Structured content is no longer just about SEO, feeds, or metadata. It becomes the operating layer of your business.

Operators that invest early in clarifying and structuring their product logic don’t just make themselves easier to “understand.” They make themselves easier to work with and more likely to remain visible in an agent-mediated world.

What “Readiness” Looks Like

UCP is new, and travel-specific conventions are still evolving, but competitive advantage will rest with those who are able and willing to experiment and be prepared. Whether UCP becomes the dominant path or simply one of several agentic distribution channels is almost beside the point. AI agents are moving upstream in the customer journey, and businesses that are structurally prepared will benefit first.

The opportunity right now isn’t to “implement UCP” so much as it is to make your business agent-ready. That’s exactly what our Agentic Readiness Audit is designed to do.

We look at how your business operates today (your content systems, booking flows, operational data, and commerce stack) and map what needs to change (and what doesn’t) to be ready for agent-driven discovery and booking. The same work that prepares you for UCP also makes it dramatically easier to experiment with future AI use cases, reduce operational inefficiencies, and turn AI into something genuinely useful inside your organization, not just another tool bolted on.

We specialize in content operations, headless CMS architecture, and agentic commerce. More importantly, we help Tour Operators prioritize investments that create optionality. Agent-driven travel is coming and it will reward operators who are prepared rather than reactive, so now is the moment to act.

Book time with us and let’s talk through what agentic readiness means for your business, and how to take the first, most practical step forward.