For 30 years, selling online meant one thing. Now it means something else.
For most of the internet’s life, if a business wanted to sell online, the answer was the same. Build a website. Put products on it. Pay Google or Meta to send customers to that website. Almost every online business, at every scale, was built this way.
That is no longer where the customer starts.
Today, more and more customers open ChatGPT, Gemini, or Claude — or ask a voice assistant on their phone — and simply say what they want.
The AI answers. Sometimes it names a brand. Sometimes it doesn’t. Sometimes the customer completes the whole purchase without ever visiting a website.
This is the shift. The website is still there. It is no longer where the sale is decided.
What is AI Commerce?
AI Commerce is commerce designed to live inside AI conversations — where a customer discovers, compares, and buys a product through a chat, voice, or messaging interface, without visiting a traditional website.
The term was introduced by Chitrangana as a category framing for a shift already visible in real markets: the moment product discovery moved from search bars to AI answers, and from browse pages to conversation.
AI Commerce is not one platform, one tool, or one protocol. It is a way of designing a business — its products, its data, its trust signals, its transaction paths — so that it lives natively inside conversation, voice, and AI-driven interfaces. It is what comes after eCommerce.
ChatGPT
Gemini
Claude
and more
Barrier repair cream · $49 — in the cart
one basket, carried across surfaces — it belongs to the customer, not to your site
Checkout, inside the conversation
payment completes where the question was asked — no website in between
The Five Layers of AI Commerce
For a business to actually work inside AI conversations — not just be mentioned in one, but be found, understood, trusted, and transacted with — five things must be in place. Chitrangana calls these the five layers of AI Commerce.
Presence
Can the AI find your business?
If your products, your prices, and your policies are not in a form an AI can read, you are simply not in the conversation. The customer asks; you are not among the names that come back.
Understanding
Can the AI read you correctly?
A product page written for a human scanning a screen is not automatically readable by a system that is trying to summarise it in one sentence. AI Commerce demands product data that is structured, complete, and honest.
Trust
Will the AI recommend you?
AI systems weigh reviews, third-party references, the age of your record, and the consistency of what you say about yourself. A brand with a strong human story and a weak digital record often loses to a smaller brand with cleaner, verifiable data.
Transaction
Can the customer buy without leaving the conversation?
AI platforms have begun letting people complete the purchase inside the chat — through open protocols like ACP (co-developed by Stripe and OpenAI) and UCP (led by Google and Shopify). A business that is not transactable on these surfaces watches the sale go to one that is.
Continuity
After the sale, does the customer belong to your business — or to the platform that served the answer?
This is the hardest question in AI Commerce, and the one most businesses have not yet asked.
Every layer is a design decision. Together, they are the architecture of AI Commerce.
This is not a five-year prediction. It is already happening.
Weekly users of the large AI assistants now number in the hundreds of millions, and the fastest-growing markets for them are no longer the ones that led the last internet cycle. A meaningful share of those people already ask an assistant before they ask a search engine.
That means the shift is not coming. It has already arrived — unevenly, market by market, and often fastest in the places furthest from where the technology was built.
In 2026, multi-brand direct-to-consumer commerce went live directly inside ChatGPT, with hundreds of brands following. A customer can now ask for “skincare for dry skin” and complete the entire purchase — discovery, comparison, checkout, payment — without leaving the conversation.
This is what AI Commerce looks like when it stops being a concept and starts being a working channel.
Who this changes things for
For a D2C brand
Your customer used to arrive on your website through a search ad or an Instagram post. Increasingly, they will arrive already having asked an AI what to buy — or never arrive at all, because the sale completed inside the chat. Your product data, your reviews, your third-party record: these decide whether the AI chose you.
For a retailer or brand with existing digital commerce
The website you have built is not going away. But it is no longer the only front door. The catalogue that lived on your storefront now needs to live everywhere the customer is asking — voice, messaging, AI chat, marketplace assistants — and behave correctly in each place.
For a traditional business without digital commerce yet
This is the moment where “should I build a website first” is no longer the right question. The right question is: how does my business become present, readable, and buyable inside the conversations where my future customers are already asking?
For an enterprise
The protocols — ACP, UCP, AP2, MCP — are already live. Each of them is a structural decision about who owns the customer, who owns the data, and who owns the transaction. These decisions are not IT decisions. They are business architecture decisions.
Is your business answerable?
Five honest questions, one for each layer. Nothing is recorded, no email is asked. The verdict updates as you answer.
Every not sure is an answer the AI has already given on your behalf.
How Chitrangana works on AI Commerce
Chitrangana approaches AI Commerce the same way it approaches any major engagement — as a business architecture problem first, and a technology problem second. Every engagement begins with a Viability Assessment.
Viability Assessment
Is AI Commerce the right move for this business, now?
Not every business needs to be in AI Commerce this year. For some, the discipline change is more important than the channel change. The Viability Assessment answers plainly: proceed, modify direction and proceed, or do not proceed — with reasons.
Innovate and assess
What should this business become, inside the conversation?
Most businesses arrive here for the first time — there is nothing to correct yet, there is something to design. We work out the structure: what you sell, how each product is described, priced, and proven, and how the business should present itself when an AI is the one reading. Then we assess that structure against how ChatGPT, Gemini, Claude, and voice assistants actually answer today.
Architecture
The five layers, designed for this business.
The five layers of AI Commerce are then designed specifically for the business: presence, understanding, trust, transaction, continuity. This is not a template. It is a blueprint.
Build, and stay
Architecture on paper is not the deliverable. A working AI Commerce presence is.
We implement — product data structure, schema, catalogue readiness, protocol integrations, trust and review architecture — and stay with the business through launch and stabilisation.
Frequently asked questions
What is AI Commerce, in one sentence? +
AI Commerce is commerce designed to live inside AI conversations — where a customer discovers, compares, and buys a product through a chat, voice, or messaging interface, without visiting a traditional website. Chitrangana introduced the term as a category framing for this shift.
Is AI Commerce the same as eCommerce or agentic commerce? +
No. eCommerce is the model built around a website or app. Agentic commerce refers to the specific mechanic of AI agents completing transactions on a customer’s behalf. AI Commerce is the broader category — a way of designing a business so it is present, readable, and transactable across every AI and conversational surface, regardless of which protocol or platform wins.
Why does AI Commerce matter right now? +
The assistants now carry hundreds of millions of weekly users, and in 2026 multi-brand direct-to-consumer commerce went live inside ChatGPT, with Google’s Universal Commerce Protocol and the Universal Cart opening a shared basket across AI surfaces. The infrastructure and the customer behaviour are both already here.
What do we actually need to change in our business? +
The changes are architectural, not cosmetic. Product data must be structured and machine-readable. Product pages must be answer-ready — written so an AI can quote them correctly, not just so a person can scroll them. Reviews, third-party listings, and marketplace records must be consistent, because AI systems judge you by what is verifiable, not by what your website says about you.
Is AI Commerce only for large brands? +
No — and in some ways, smaller brands have an easier time. A brand with fewer SKUs and cleaner catalogue data is often more legible to an AI than a large enterprise with tangled product records. Structural clarity matters more than size.
Where should a business start? +
Start with an assessment — search for your business, by name and by category, inside ChatGPT, Gemini, and Google’s AI answers. What comes back, or does not, is usually the first honest answer to whether your business is already in the AI conversation.
The category is new. The decision is not.
Every structural shift in commerce — the arrival of eCommerce itself, the rise of the marketplace, the mobile shift, the D2C wave — created a small window where the businesses that moved early held positions the later entrants could not easily take back.
AI Commerce is that window now — and it is measured in quarters, not years.
Chitrangana can tell you plainly, in one conversation, whether your business should move — and if it should, where to start.
Our AI Commerce consulting covers both sides of the same problem: AI in eCommerce for the business that already sells online, and AI Commerce architecture for the business whose first storefront will be a conversation.
