Agentic Commerce Needs Truth, Not Just Connection
During this year’s NRF 2026 event, Google and Shopify announced the introduction of the Universal Commerce Protocol (UCP).
This new standard is designed to do more than just connect systems; it was created to integrate the full complexity of a commerce journey into a seamless conversational experience.
This represents a strong step forward. UCP provides the possibility of a shared language for agentic commerce, and it enables a world in which AI assistants can browse, compare, and transact on behalf of customers – and do that across retailers, platforms, and channels.
It is therefore part of the answer to addressing one of digital commerce’s most persistent challenges, namely that of fragmentation.
While protocols standardize how agents connect to commerce systems, they cannot however guarantee that what those agents rely on is always true. And we have seen from our own experience that, in agentic commerce, truth is everything.
Agentic Commerce Depends on Trustworthy Information
In other words, AI agents do not ‘browse’ in the way that humans do, but rather they make decisions based on signals. They may rely on such information as availability, delivery dates, fulfilment options, or returns policies. These are inputs, as opposed to marketing messages. If those inputs are wrong, outdated, or overly optimistic, the entire experience is negatively impacted.
An agent that confidently places an order for an item that cannot ship, cannot arrive on time, or cannot be fulfilled as promised doesn’t only fail in the transaction, it damages trust in the system (and indeed the brand) that enabled it.
Agentic commerce fails when agents connect to unreliable operational truth.
The Hidden Risk: Promises Without Operational Reality
On the positive side, UCP standardizes how agents:
- Discover products
- Add them to carts;
- And initiate checkout flows
However, this alone does not guarantee:
- Accurate, real-time inventory availability
- A reliable delivery promise calculation
- Intelligent fulfilment decisions across locations
- Transparent handling of splits, substitutions, or delays
- Consistent post-purchase execution
The items on the latter list are not protocol problems – they are order management problems. And the danger of agentic commerce is that it can dramatically amplify these existing challenges.
OMS as a Key Foundation in a UCP World
If AI-driven commerce is to work at scale, there must be a source of truth behind the protocol (indeed any protocol), and that role belongs to the Order Management System.
OMS provides the operational guarantees that agentic commerce depends on, for example:
- Real-time inventory visibility (So agents don’t sell what can’t be delivered).
- Delivery promise management (So dates committed to at decision time can actually be met).
- Dynamic order routing and intelligent order splitting (So fulfilment decisions are optimized but without confusing customers).
- Multi-fulfilment orchestration (So agents can offer choice without introducing risk).
- Exception handling and lifecycle visibility (So when things change, the system responds before trust is lost).
In an agentic world, we believe that the OMS isn’t just there to manage orders, but – most importantly – to protect promises.
And, as AI agents increasingly act on behalf of customers, brands will be increasingly judged on reliability. We can think of it this way: While UCP is a step forward in enabling the conversation, it is the OMS that ensures the outcome.
The winners in the new era of agentic commerce will not be those brands that promise the most, but the ones that only commit to what they can deliver, and to deliver that experience consistently. This goes beyond protocol alone, as all these additional factors are dependent on operational discipline, real-time intelligence, and systems designed to manage complexity without breaking trust.
Therefore, the brands that will enjoy the most success in the agentic commerce world will be those that understand the difference between connection, which enables commerce to take place; and delivery which earns trust. In other words, being able to make promises and then keep them as part of the customer experience.
OneStock’s Role: Bridging the Gap
This is exactly where OneStock is stepping in. We are currently working on the integration of the UCP protocol in close collaboration with our two key partners, Shopify and Google.
Our goal is to empower retailers to enhance the agentic commerce journey by injecting operational reality directly into the process. By integrating UCP with the OneStock OMS, we provide the true customer promise – including real-time availability and precise delivery options – from the very first interaction on the product page all the way through to checkout.
Ultimately, relying on OneStock is a guarantee of a high-quality, sustainable conversational experience. By feeding clean, accurate data through the UCP protocol, we ensure that the AI agent remains credible over time, avoiding the hallucinations or errors that come from poor data.
- To achieve this, OneStock directly answers the UCP “Capability Check” request. Instead of generic estimates, we respond with valid, executable logistics options – specific shipping methods, accurate costs, and guaranteed dates – tailored to the customer’s precise context.
- Once the decision is made, the integration ensures a seamless transition to transaction. Upon receiving the “Session Complete” signal via the UCP protocol, the order is instantly injected into OneStock, triggering the fulfillment logic without friction or delay.
- Furthermore, this integration empowers agents to effectively manage the post-purchase journey, ensuring that the agent’s visibility doesn’t end when the order is placed. To support this, OneStock exposes a Unified Tracking endpoint that the agent can query at any time to retrieve live tracking links or precise status details, keeping the customer informed proactively.
