AI is changing how people buy things
The customer journey has never been straightforward or one-way. But as more people use AI to discover products, what used to often start with a simple search query now begins as a conversation with an LLM.
60% of consumers worldwide interact with AI at least once a week, according to Klaviyo's 2026 AI Consumer Trends Report. Whether they’re making personal decisions, looking for inspiration, or planning activities, at least 1 in 5 consumers now start their research with LLMs.
But the AI customer journey goes beyond just discovering products. AI is also changing what consumers already know before they land on your website, and what they expect from the customer experience when they get there.
The customer journey now has new friction points and fresh ‘aha’ moments, which means new opportunities for your brand to win new customers and keep existing ones who found your website through AI.
Here’s what you need to know about the AI customer journey, so your brand can start seeing the benefits of AI-powered shopping.
In this guide:
- Collapsed discovery and research
- Extended consideration phase
- Real-time answers for conversion
- Personalisation trust signals
- Poor AI and brand trust
- Customer service revenue
Shift 1: AI has collapsed the discovery and research phases
Before AI shopping, most B2C product discovery happened through traditional organic search or social media. When customers entered the research phase, they would usually hop across several websites or social media accounts to compare products.
Now, AI is streamlining this process. With AI chats, consumers are discovering and comparing new products within the same chat interface. In some cases, they can even buy a product without leaving the AI conversation.
43% of consumers have used AI when shopping for non-essential products in the past 6 months, according to our AI consumer trends research. 39% have also bought a product recommended by AI during that period.
When they search with AI, consumers also tend to use longer, more detailed prompts to find what they want. Based on our AI Consumer Trends Report, 30% of consumers use 8 or more keywords for an AI search query, and 78% say they include emotional or personal context at least some of the time.
With this context, LLMs can recommend product listings, generate comparisons, and summarise customer reviews instead of just showing a list of brand websites. This means the consumer doesn’t need to open another tab to move on to the research phase of their customer journey.
LLMs can also follow up with clarifying questions, like, 'Would you like me to share free shipping or in-person pick-up options?' or 'Is there a particular colour or style you prefer?' This is what makes it possible for the customer to discover and research products in the same conversation.
How to adapt
- Structure your content to match how people search with AI.Keyword-optimised product copy alone won’t help your products get discovered through AI. Organise your product descriptions and metadata around use cases, comparisons, scenarios, and outcomes — not just categories and specs.
- Get more detailed customer reviews.LLMs are starting to quote third-party customer review websites as key sources of information when giving product recommendations, according to Search Atlas research. Come up with a more thoughtful strategy to collect and incentivise plenty of new, detailed reviews (EN) ↗, so that AI crawlers always have up-to-date ones to draw from.
Shift 2: The consideration phase is longer
AI can curate product recommendations and build comparison tables in seconds. The irony is that with so much information instantly available, consumers are actually spending more time considering their options.
2025 was the year of the ‘browsing boom (EN) ↗’ for ecommerce brands. Product views across ecommerce sites were up 37% YoY before BFCM, while order growth rose 14%—a widening gap that shows shoppers are taking their time to compare, research, and think carefully before they buy.
This jump can’t be explained entirely by AI shopping, especially since more shoppers are viewing product listings on AI platforms instead of on-site. But other research shows a link between AI and consideration: in a 2026 survey by Gartner, for example, 31% of consumers said that AI overviews made them consider more product options.
AI is a shortcut to finding the best products based on your specific context and needs. But people here still take time to go through the details before making the next move, and they may have more in-depth chats with LLMs while they do so.
This means shoppers may be more informed and have higher expectations when they land on your website to dig even deeper. If the on-site experience doesn’t match these expectations, you risk losing a high-intent visitor.
How to adapt
- Set up AI attribution. Shopify and Klaviyo data (EN) ↗ use UTM parameters to help you see who started a checkout, placed an order, or browsed but didn’t buy from an AI search. You can use this data to deliver personalised abandonment and post-purchase flows with more detailed product information, since you can reasonably assume AI shoppers want it.
- Invest in detailed, contextual zero-party data (EN) ↗ collection. Learn as much as you can about your AI traffic audience, so you can build a strong relationship with them and hopefully shorten their consideration phase. Use quizzes or forms to collect more in-depth preference or lifestyle data, so your browse and cart abandonment flows can better match their context and needs.
Shift 3: Conversion needs real-time answers
Just like how Amazon changed what people expect for fast delivery, AI is now changing what people expect for instant, in-depth access to information.
People don’t drop these expectations when they land on your website. In fact, 75% of consumers have abandoned a purchase because they couldn’t get instant answers, according to the research behind Klaviyo’s 2025 AI Shopping Index.
Brand-side AI customer agents are becoming a must-have for the AI customer journey. But these interactions need to deliver the same value consumers are used to with LLMs—and that means they need to be personalised.
An AI customer agent should be able to:
- Recommend products based on real-time customer data, including past browsing behaviour, purchases, and stated preferences.
- Suggest products that go well with the items already in someone’s shopping basket.
- Answer detailed questions about each product in your inventory, including what they’re made of, where they’re sourced, or when they’re expected to be back in stock.
How to adapt
- Train your AI customer agent on customer and product data.Your AI agent won’t be able to offer high-quality, personalised answers to questions if it knows nothing about your customers or your products. Personalised service means your AI agent has a dynamic, growing knowledge of your customers and products at all times.
- Make it easy to escalate complex enquiries to a human.There will always be certain questions that even the most capable AI agent can’t resolve on its own. Use a shared helpdesk where your AI and human agents have access to the same data and context, so the handover doesn’t cause friction for the customer.
Shift 4: Personalisation can make or break trust
74% of consumers expect more personalised experiences from brands, according to Klaviyo’s 2025 Future of Consumer Marketing Report.
This was already the case before AI shopping became more common, but AI personalisation is now also starting to make or break trust between brands and consumers: more than half (55%) of consumers have had an 'aha' moment where AI impressed them with its accuracy or personalisation, according to our AI consumer trends research.
When personalisation goes wrong, the impact can snowball. The same study found that when people get poorly personalised content, their main response is to be less likely to open future messages from that brand.
How to adapt
- Unify your customer data.Your personalisation strategy is only as strong as your customer data. If your customer profiles are incomplete or fragmented across multiple platforms, now is the time to consider consolidation. A shared, real-time view of each customer makes it easier to deliver consistent, accurate personalisation across every marketing and customer service touchpoint.
- Invest in predictive analytics.Use predictive analytics to go beyond rule-based personalisation. Instead, tailor your messages to each person’s preferred timing, channel, and products. For example, you can send well-timed discounts with personalised product recommendations to high-risk customers by combining their browsing behaviour, channel preferences, preferred send times, and potential churn risk.
Shift 5: Brands can lose trust when they implement AI poorly
Only 13% of consumers fully trust AI, according to our AI Consumer Trends Report. This puts brands in a grey area: as shoppers get used to AI and form their own views on how to use it, AI can be both a help and a hassle.
The same research found that almost half (47%) of consumers think AI has improved the quality of the product recommendations they receive, and 43% say it’s making brands’ customer service better.
But a recommendation that totally misses the mark, a robotic support interaction, or generic AI content can quickly turn people off: nearly 1 in 5 consumers lose trust in a brand’s data practices after a poorly personalised experience, according to our research, and 41% say customer service chats that don’t feel human are the brand interactions that come across as ‘too automated’.
How to adapt
- Keep humans in the loop.Every consumer is at a different stage with AI, so don’t force the technology on those who are sceptical. When you do offer AI experiences, make it clear how shoppers can switch over or escalate the conversation to a human.
- Don’t sacrifice your brand style.Train your AI on your brand’s distinctvoice and tone (EN) ↗guidelines so your AI interactions don’t feel cold or generic.
- Monitor AI-specific trust signals.Keep track of where AI is affecting the customer experience negatively. Watch metrics like unsubscribe rates after AI-generated marketing messages, survey responses after interactions with AI agents, and sentiment throughout conversations with AI customer agents.
Shift 6: Customer service is a revenue generator
AI is shifting post-purchase customer service from a returns and order management function to a proactive revenue and loyalty driver. With AI agents available 24/7, every interaction after the first sale becomes a chance to deepen the customer relationship and drive repeat revenue.
AI agents that have access to real-time customer data can deliver a white-glove experience. For example:
- A hotel brand’s AI agent could proactively offer local restaurant recommendations or assist with transport plans after a guest books their stay.
- A children’s retailer could set up an AI agent to analyse customers’ baby registry information and suggest age-appropriate items based on their child’s predicted age or size.
How to adapt
- Sync service and marketing data. Every customer interaction generates data that can personalise retention touchpoints. Your marketing and customer service data should be accessible from the same place, so that AI and human agents can tailor their conversations to previously shared preferences and product questions.
- Create a 1:1 post-purchase experience. Set up a self-serve customer hub where shoppers can do more than track their orders. Offer personalised product feeds, AI agents that answer questions and recommend products, and a place where customers can redeem loyalty points.
- Anticipate customers’ needs before they even ask. Turn purchase history into more proactive, relevant customer outreach. Use AI to predict the next best products a customer might want based on their previous orders, for example, and embed those recommendations in post-purchase emails or in customer hubs.
How Klaviyo supports the AI customer journey
The customer journey still includes awareness, consideration, conversion, fulfilment, and loyalty phases. What’s changed is the infrastructure underneath it.
AI has made discovery faster, consideration more thorough, expectations higher, and the cost of friction steeper.
Here’s how Klaviyo can help you adapt your marketing and service for the AI customer journey:
- Nurture: Composer plans and launches marketing autonomously, based on a simple prompt or idea and your brand guidelines. You maintain full strategic control—nothing launches without your sign-off.
- Conversion: K:AI Customer Agent draws on your customer data, product catalogue, policies, and brand voice to provide 24/7 personalised shopping assistance, answer questions, and escalate issues to humans with full context intact.
- Personalisation: Features like predictive analytics, channel affinity, and personalised A/B testing deliver 1:1 personalisation without manual segmentation.
- Post-purchase: Klaviyo Customer Hub creates one personalised place where customers can manage their relationship with your brand, from self-service options to product recommendations and rewards redemption.