Commerce beyond the website
When discovery happens in an AI conversation, current product information and connected business systems become part of the buying experience.
The buying journey can start elsewhere
“Find me a pair of walking shoes for a week in Spain. They need to be comfortable, fit my budget and arrive before I leave.”
That request combines product research with practical conditions. A useful answer needs more than a persuasive description. It needs suitability, size options, price, delivery information and perhaps a returns policy.
An AI service may help build the shortlist before the customer reaches a retailer's website. The site still has a role, but it is no longer the only place where the business needs to be understood.
A real example, with an important boundary
In June 2026, IHG announced an app in ChatGPT that lets people explore and compare its hotels, with current availability, pricing and amenities. Once someone selects a hotel, they are directed to IHG's own booking channels to complete the reservation. [1]
That distinction matters. Discovery and comparison can happen in the conversation while the transaction remains in the organisation's existing environment.
The lesson is not that every purchase has moved into AI. It is that organisations can connect useful parts of a customer journey without rebuilding the entire experience.
Good answers depend on operational detail
For the walking shoes, the product system may hold the specification, the commerce system the price, and fulfilment systems the delivery options. Content teams explain the features and conditions in language people understand.
If an answer combines the wrong size with an old delivery estimate, a polished conversation will not rescue it. Information needs the context that makes it useful: the right market, product variant, time and conditions.
This is why the work extends beyond generating descriptions. Organisations need to maintain and connect the information behind those descriptions, create content where explanations are missing, and review what is made available.
Three steps towards participation
First, understand the answers customers receive. An audit can reveal omissions and inaccuracies in selected AI experiences. Use those findings to prioritise questions worth addressing.
Next, improve and deliver the information. Connect the relevant sources, resolve gaps and make reviewed content available through suitable channels. This can benefit both your website and experiences beyond it.
Then offer a direct route to your organisation. A Brand Agent can answer from permitted information and supported system connections. As customer agents become part of the journey, the organisation needs a clear way to respond to their requests too.
These are steps you can take at your own pace. Direct answers do not automatically mean permission to place an order. Any transaction needs its own supported integration, authorisation and checks.
Keep the organisation in control
A customer might reasonably ask about product availability without sharing personal information. An order enquiry is different: it may require identity checks and access to a particular customer record.
The experience needs to respect that distinction. Decide what can be public, what requires permission and when a question should move to a person or an existing service flow.
Kodiac connects existing content and data with the work of preparing information and enabling a Brand Agent. It gives organisations a route into direct engagement while retaining the systems and responsibilities behind the answer. It does not require every interaction to become a transaction.
Choose one useful journey
Start with a product category and a customer question that brings research and operational information together. Check which facts change frequently, where they are maintained and what happens when an answer cannot be confirmed.
Test the experience from the customer's side: did it help them decide, and could they take the next step? A good hand-off to an existing checkout or adviser may be more useful than trying to automate the entire journey at once.
The opportunity is to serve customers where the decision begins, using the knowledge and systems the business already has.