AI Commerce Won’t Be Won by the Best Chatbot. It Will Be Won by Context.
By Tom Reidy
Key takeaways
- Agentic commerce shifts discovery outside brand-owned websites.
- AI assistants drive sales, but brands risk losing relationship visibility.
- Winning in AI commerce requires structuring data for external AI agents.
- Website chatbots provide control only after the customer decides to visit.
As AI begins to research, recommend and act for consumers, brands face a new challenge: staying useful when more of the customer journey happens outside their own website.
Most brands are still treating AI as a website upgrade. Add a chatbot, connect it to a knowledge base and call it innovation.
But the bigger shift is not what appears on the website. It is where the customer journey begins, who guides it and how much of that relationship the brand can actually see.
Agentic commerce describes the move from AI simply answering questions to helping people complete outcomes. A customer might ask an AI assistant to compare mobile plans, find a suitable gift, narrow down a vehicle shortlist or recommend the right product for a particular need. Over time, the agent may do more of the work: checking availability, arranging a booking, completing a purchase and managing what happens afterwards.
That means a brand’s website is no longer guaranteed to be the first—or even the most important—stop.
The customer journey is moving beyond the website
A recent WhatsApp Business article drawing on IDC research commissioned by Meta predicts that 64% of customer interactions will run through agents and answer engines rather than brand-owned websites by 2029. It also cites research suggesting 63% of consumers now begin product and service research with an AI assistant instead of a search engine, while 87% say AI plays a role in their purchase decisions.
Those are striking figures. Because the research was sponsored by Meta and published by WhatsApp Business, they should be read with the appropriate commercial context. But the direction of travel is difficult to dismiss: discovery is becoming more conversational, and AI is beginning to sit between customers and brands.
The report identifies three main places where these interactions are developing: general AI assistants, chatbots on brand-owned websites and conversational messaging platforms.
Each solves a different part of the problem.
AI assistants offer reach, but brands lose visibility
General AI assistants make discovery incredibly convenient. A customer can describe what they need in their own words and receive a shortlist without visiting ten websites or opening twenty browser tabs.
For brands, that creates a new visibility challenge. They need accurate, useful and accessible information that AI systems can understand and confidently recommend. But even when a brand is included in an answer, it may learn very little about the customer’s wider research journey, the alternatives considered or the conversation that led to the recommendation.
The brand may win the transaction without owning the relationship.
Brand chatbots offer control, but only after someone arrives
An AI assistant on a company website gives the business much more control. It can use approved product information, connect to customer systems and collect valuable first-party data.
The limitation is reach. The customer must already know the brand, choose to visit its website and decide to interact with the agent. That makes a website chatbot useful for converting existing intent, but less effective at reaching people who are still browsing, comparing or simply not ready to buy.
It is ownership without guaranteed attention.
Messaging could become the continuity layer
Messaging platforms sit between these two extremes. They are already part of people’s daily behaviour, yet they can also give a business an ongoing, permission-based conversation with an identifiable customer.
The important feature is not the chat window. It is continuity.
A persistent conversation can retain what the customer asked, what they were shown, what they purchased and what was decided last time. Instead of forcing someone to repeat the same information at every step, the next interaction can begin with useful context.
That turns messaging from a campaign channel into a relationship layer.
For Meta, the obvious answer is WhatsApp, Messenger and Instagram. But New Zealand businesses should not mistake the platform recommendation for the entire strategy. Depending on the audience, the right channel could also include SMS, email, a logged-in customer portal or another messaging service.
The strategic question is not simply, “Should we build a WhatsApp bot?” It is: “Where do our customers already hold conversations, and can we make those conversations more useful over time?”
Context is the real competitive advantage
Context is more than a customer’s name and purchase history. It includes their preferences, previous questions, current intent, known constraints, earlier offers and the decisions already made.
Without context, AI can automate a generic response. With context, it can help resolve the actual problem.
That distinction will become increasingly important as AI tools become easier for every business to access. The technology itself will not remain a meaningful point of difference for long. The advantage will come from the quality of the information connected to it, the history it can responsibly use and the experience it creates for the customer.
Generic automation scales repetition. Contextual automation scales relevance.
What should New Zealand brands do now?
First, make product and service information easy for both people and AI systems to understand. Pricing, availability, policies, specifications and frequently asked questions need to be accurate, structured and kept current. If an AI assistant cannot interpret the offer, the brand is less likely to enter the consideration set.
Second, decide where the ongoing customer relationship should live. A website visit may begin the interaction, but it should not necessarily be the only place that interaction can continue.
Third, connect customer-facing AI to the systems that make it genuinely useful. An agent that can answer a basic question is helpful. One that can check an order, change a booking, confirm stock or smoothly involve the right person is far more valuable.
Fourth, build trust into the system from the beginning. Customers should understand when they are dealing with AI, what information is being retained and how they can reach a human. Context should make the experience feel easier, not invasive.
Finally, measure more than chatbot activity. Resolution time, qualified leads, repeat purchases, customer retention and successful human handovers reveal far more than the number of conversations started.
The next front door to your business
Websites will remain important, but they may no longer carry the entire customer journey. AI assistants could become the place where a need is first expressed. Messaging could become where the relationship continues. The website may increasingly serve as one trusted source within a much wider system.
The brands that win will not necessarily be those with the flashiest AI agent. They will be the ones whose information is easy to discover, whose systems can remember what matters and whose next interaction is more useful than the last.
In an AI-mediated market, visibility gets a brand considered. Context helps it get chosen. Trust keeps the relationship.
Source: “Building agentic commerce: The channels your AI strategy can’t ignore,” WhatsApp Business, 12 August 2026. The article draws on IDC research commissioned by Meta.