Five ways AI is already changing the customer journey
In May 2026, NielsenIQ reported that 42% of consumers have used at least one AI tool to shop within the past month.
People can ask an assistant to research a category, compare products, explain trade-offs, find a suitable option, or answer a question after purchase.
Although it’s ultimately still people who make the final decision, it’s clear that AI-assisted shopping is already mainstream.
Moreover, it’s a powerful indicator of the delegation of certain parts of the shopping journey to AI agents. This is agentic commerce and, while it’s still considered an emerging trend, the infrastructure for it is being built at speed. OpenAI, Stripe, Google, Visa and Mastercard have all shipped agentic transaction protocols in the past year.
Yet the majority of brands are still unprepared for the shift to a customer journey with not just one but two audiences: humans and AI agents.
At DEPT®, we’ve been tracking the growing role of AI across the discovery, consideration, conversion, and use stages of the customer journey. The following represent the five most critical changes we’ve observed and the practical actions we recommend brands take as part of a comprehensive business-to-agent (B2A) strategy.
1. Organic visibility becomes a knowledge challenge.
Thanks to generative search, discovery now starts with a synthesized answer. And generative visibility reaches further than a collection of optimized pages. AI systems can draw on websites, product feeds, documentation, reviews, earned coverage, and other accessible sources when constructing an answer.
Generative search can break one request into several related searches before assembling its response. Google calls this query fan-out. It allows AI Mode to explore subtopics and data sources that a conventional query might miss.
Your brand may never receive the first visit. Its initial appearance could be a short AI-generated description assembled from product pages, reviews, documentation, third-party coverage, and structured data.
That description needs to be accurate before it can be persuasive.
The larger job, therefore, is organizational. Your public information needs to describe the same thing wherever an AI encounters it.
Build a governed source of truth for company and product information. Give documentation, FAQs, reviews, and structured data the same strategic attention as campaign content. Then audit the answers AI systems produce, tracking factual accuracy, brand association, source citations, and referral traffic alongside search rankings.
Inconsistency creates a practical disadvantage. If product claims vary across a website, a retailer listing, a sales document, and the support center, an AI system has to decide which version to trust. Without a governed source of truth, you could easily ask an AI assistant for recommendations within your brand’s category and receive a list of reasons for why it would recommend a more machine-readable competitor.
This gives CMOs a wider remit. Brand accuracy now depends on coordination across marketing, commerce, communications, service, product, and technology.
2. AI advertising enters the consideration process.
Advertising is moving closer to a point where AI will assemble and explain the customer’s options.
By offering premium placement directly below its response to a question, ChatGPT Ads gives brands a clear opportunity to target users with search intents currently unknown to traditional search engines. Other LLM advertising works similarly, like AI Max, which matches campaigns with more complex searches using creative assets, landing pages, and query signals.
This favors campaigns built around specific, recognizable problems for customers. Assets need enough substance to be interpreted, summarized, and compared. A sharp headline helps, but the surrounding product data and landing-page evidence must support it.
CMOs should test these formats early with clear controls, as we’ve been doing through our partnership with Lumen. Connect paid media to first-party insight and current product feeds. Add recommendation influence and incremental consideration to the measurement plan, rather than asking click-through rate to tell the whole story.
3. Agentic commerce rewards operational clarity.
Agentic commerce is commerce in which a customer delegates a commercial task to an AI agent, and the agent executes it. Already, AI’s role in some journeys is extending from recommending a purchase to selecting, transacting, or managing that purchase on the customer’s behalf, within limits the customer sets.
OpenAI’s Instant Checkout allows eligible users to purchase products via ChatGPT, powered by the Agentic Commerce Protocol. The merchant still controls fulfillment and customer service.
Customers will likely maintain closer control over expensive, sensitive, or emotionally charged purchases. Routine decisions create a clearer opening for agents, especially when the customer has already established their preferences.
Brands should prepare for both modes. The human experience needs confidence, clarity, and a reason to choose. The agent needs current product data, explicit terms, accurate inventory, and a transaction path it can navigate with permission. It also needs a checkout or conversion flow that can distinguish an authorized agent from a malicious bot. Visa, for example, has developed a Trusted Agent Protocol to help merchants make that distinction.
Start by finding the tasks customers already treat as administrative work. Reordering, renewing, assembling a standard basket, comparing plans, and managing subscriptions are credible candidates.
Then fix the infrastructure beneath them. Improve metadata. Standardize inventory and pricing signals. Open appropriate feeds or APIs. Define when the agent can act, when it needs confirmation, and when a person should take over.
4. Generative interfaces organize experiences around intent.
Websites ask customers to understand the organization’s categories. Generative interfaces let them start with their own goal.
A customer can describe what they need, provide constraints, and refine the request through conversation. This opens useful design territory for brands with complex catalogs, technical products, or high-consideration services.
PwC recently launched a generative experience that uses a conversational interface to help customers leverage their company’s knowledge base to unlock their own tech-driven growth.
A chatbot alone won’t improve the journey. Start with your highest-friction decision moment—not your homepage—connect the interface to governed content, and keep human support available where judgment or reassurance matters.
5. Audience insight becomes a continuous capability.
Annual segmentation studies can tell you who your customers are. They struggle to capture what someone is trying to achieve in a specific moment.
AI closes that gap by turning audience insight into a continuous capability. Instead of refreshing your understanding of demand once a year, your teams can read it from the signals customers generate every day: purchase histories, search patterns, service conversations, campaign responses, and act on it while it’s still current.
We did exactly this for Rituals, one of Europe’s largest bath & body brands. By modeling the demand signals in Rituals’ customer data, we were able to predict which product each customer was likely to want next and surface it across channels, engaging the right people at the right time with the right product.
The emerging extension of this capability is synthetic audiences: AI-simulated customer groups that let you generate and test hypotheses about demand, how a new proposition, price, or message might land, before committing budget to find out.
What AI systems say about your brand is now an observable, measurable variable, and the brands that win will be the ones tracking it continuously, not auditing it once.
Building a B2A strategy for your brand
While B2C and B2B marketing will continue to be essential to your strategy, B2A represents an equally essential new addition as AI systems play an increasingly influential role in the customer journey.
Brands still need to earn human attention and trust, while giving AI enough reliable information to understand what they offer.
Those that make their knowledge easier to retrieve, products easier to evaluate, and experiences easier to act on will earn a place in more decisions, whether the customer arrives alone or sends an agent first.
