Your Next Customer Might Not Be Human: How to Win in the Age of AI Shopping

A practical guide to getting discovered, getting chosen and measuring what works as AI assistants and agents reshape ecommerce.

For twenty years, ecommerce ran on a simple idea: get a person to your website, then convince them to buy. You optimized for search rankings, polished product photos, wrote persuasive copy and tested checkout flows, all for a human with a browser.

That idea is breaking down. More and more shopping journeys now include a third party: an AI assistant or agent sitting between your brand and your customer. It reads your product data, compares you with competitors, summarizes your return policy and, increasingly, makes the recommendation or completes the purchase itself.

Microsoft's new guidance on what it calls "the era of agentic commerce" lays out how merchants can adapt. This post turns the key ideas into steps you can take, whether you lead ecommerce, manage paid media or support brands through an agency.

The shift you can't sit out

The numbers make a strong case:

• Automated traffic is growing 8x faster than human traffic, and AI-driven sessions nearly tripled in 2025 (HUMAN Security, 2026 State of AI Traffic and Cyberthreat Benchmark Report).
• Over 72% of consumers expect brands to offer AI-powered shopping experiences within a year (Microsoft Advertising US consumer survey, n=1,028, Aug–Sep 2025).
• AI-referred shoppers convert 42% better than non-AI traffic (Adobe Analytics, Q1 2026 AI Traffic Report).

That's a channel that is growing fast, that customers expect, and that converts unusually well. Microsoft calls it one of the most significant changes in digital commerce since the rise of mobile, and argues that brands that build an agentic commerce foundation early gain a compounding advantage competitors will struggle to copy.

The mobile comparison fits. Brands that treated mobile as an afterthought in 2010 spent years catching up, and agentic commerce looks set to move faster.

Your product data is now your most valuable marketing asset

If you remember one idea, make it this: AI doesn't shop the way people do.

People respond to lifestyle imagery, design and storytelling. LLMs and AI agents judge products mainly on structured data: titles, descriptions, identifiers, prices, availability, categories and policies. If that information is thin, incomplete or hard for a machine to read, your product may never appear in a recommendation, let alone make the sale.

So your product feed is no longer a back-office file for Shopping ads. It's a primary marketing asset, and often the first thing an AI "sees" when deciding whether to show you to a customer.

The three eras of the web, running at once

It helps to think of the web in three "eras". Your customers move between all of them, sometimes in the same week.

These eras run side by side, and one doesn't replace another. The biggest opportunities today are in the LLM and agentic webs, where AI increasingly shapes what gets discovered, what gets chosen and what gets bought. Winning there comes down to three stages: Get discovered, Be chosen and Understand.

Chapter 1: Get discovered

When a shopper asks an AI assistant for a recommendation, which brands appear is decided long before the prompt is typed. It depends on your product feed, your site signals, how your products are represented across the web, and where you use paid advertising to extend your reach.

Appear in organic AI results, accurately

AI assistants can only recommend products they can find, read and trust. A better website alone won't fix this. Agents can crawl your pages, but those pages were designed for people, so agents often have to infer product details, and inference leads to mistakes. Structured, machine-readable product data is the answer. In Microsoft's ecosystem, that data flows through Microsoft Merchant Center into both paid and organic experiences across Bing and Copilot.

Make your feed AI-ready:

• Submit 100% of your catalog, not just the products you advertise. A partial feed leaves the rest of your catalog invisible to AI. You can upload manually, use FTP/SFTP, have Merchant Center fetch from a URL, use the Content API, or import from Google Merchant Center.
• Get the core attributes right. Accurate titles, descriptions, GTIN/MPN, real-time availability, pricing and product category carry the most weight in AI recommendations.
• Audit your top sellers with Product Explorer in Merchant Center.
• Refresh feeds at least daily. Outdated prices and out-of-stock mismatches are the top reasons products lose visibility.
• Turn on Merchant Center alerts to catch processing failures, disapprovals and price discrepancies early.

Enable the Universal Commerce Protocol (UCP):

This is the most forward-looking recommendation in the chapter. UCP is a common language that lets AI assistants and shopping platforms talk to each other without every company building custom integrations. Think of your catalog as a book: UCP translates it into the language LLMs and agents read, so they can understand, trust and recommend your products.

UCP builds on your existing feed and adds richer context, such as return policies and customer support details, across both ads and organic results. To start:

1. Share brand-wide context. In Merchant Center, go to Settings → UCP settings and add your return policy and support contacts.
2. Share product-level context through your feed. Product warning attributes (consumer_notice_type and consumer_notice_message) are required where they apply, for example in toys or regulated goods. Add a unique merchant_item_id for each product so you can track it across systems.

Currently available to US advertisers (excluding aggregators). More attributes are coming.

Extend your presence beyond organic discovery

Organic visibility is one route in; paid placement is the other. Ads in Copilot extend your existing Microsoft Advertising campaigns into AI-native placements across Copilot Search on Bing, Copilot in Bing and Edge, and the Copilot app. These placements include chat answers, product carousels and Offer Highlights.

Offer Highlights surface your differentiators, such as free shipping, price, availability or convenience, inside the conversation and matched to what the shopper wants. If someone asks for running shoes that arrive by Friday, your free-shipping offer can appear in the answer.

To start, keep a complete and accurate feed in Merchant Center, then set up Performance Max, AI Max and/or Shopping campaigns. AI-native ads are included automatically in these campaign types. To qualify for shipping and pickup highlights, add "free shipping" to your feed or enable Local Inventory ads. (US only.)

Chapter 2: Be chosen

Put simply: getting discovered is table stakes; being chosen is where revenue is won. AI is shortening the path to purchase, and the brands that win will help shoppers evaluate with confidence and buy easily when intent is highest.

Capture purchases inside AI experiences

Once a shopper finds the right product in an AI conversation, every redirect or login screen is a chance to lose them. Copilot Checkout lets shoppers buy directly inside Copilot, on Copilot.com and the mobile app, without leaving the conversation.

What matters most for merchants: you remain the merchant of record. Your existing payments, fulfillment and reconciliation systems handle the transaction, and you keep the customer relationship and data. As Stripe's Head of Payments, Kevin Miller, puts it, "AI is changing how commerce works, and as with every technology shift, it needs new infrastructure."

There are three ways to onboard:

• Shopify: eligible merchants are enrolled automatically, with controls in Shopify admin.
• PayPal: Store Sync handles cart creation, updates and checkout.
• Stripe: the Agentic Commerce Suite offers a turnkey setup.

US only (English, USD), currently in pilot.

Give every customer a personal shopper with Brand Agents

Shoppers who arrive on your site with questions have always had two options: dig through your pages or leave. Both can cost you the sale. Brand Agents in Microsoft Clarity offer a third: a turnkey AI shopping assistant that answers questions in your brand's voice on every page.

It's more than a generic chat widget. A Brand Agent is trained on your catalog, brand guidelines and selling points. It helps shoppers discover and compare products and find the right fit. It learns from real conversations and responds to shopper behavior as well as prompts. Once enabled, it also connects your Shopify or WooCommerce catalog to Microsoft's commerce ecosystem, so your products can appear in Copilot and Bing.

The reported impact is significant: 2x higher conversion rates in Brand Agent-assisted sessions compared with unassisted ones (internal Microsoft analysis, March 2026). Tony Baldwin, CTO of Alexander Del Rossa, called it "the best example of AI integration seen to date."

Setup takes four steps:

1. Install Clarity on your Shopify or WooCommerce store.
2. Go to Settings → Brand Agents. The agent builds itself from your store data, syncing your catalog, matching your voice and running a quality check.
3. Customize the welcome message, branding and tone, and A/B test with rollout controls.
4. Publish and monitor conversions, AOV and engagement in the Brand Agents dashboard.

You need a Shopify or WooCommerce store. Brand Agents is currently in beta, so check eligibility.

Chapter 3: Understand

Everything above works better when you can measure it, and this is where most businesses struggle. Traditional analytics tell you what happened. They don't tell you why competitors win in AI, which content influences AI recommendations, or how AI-driven discovery connects to what shoppers do on your site. AI traffic also behaves differently from organic traffic, which makes the gap wider.

Microsoft Clarity is built to answer three questions: Discoverability (is AI citing my content, and which pages?), Visibility (which topics and queries help AI understand me?) and Website engagement (do AI-referred shoppers convert, and where do they drop off?).

Clarity's AI Visibility insights cover these with four reports:

• AI Bot Activity: which AI platforms crawl you, your top pages, and AI versus human traffic.
• Citations: where and how often AI cites you, plus the grounding queries behind each citation.
• Share of Authority: your citations benchmarked against competitors by topic.
• Topic Insights: content gaps where competitors are cited and you aren't, with prioritized fixes.

Here's a simple example. A retailer notices that AI-referred shoppers compare products but leave before adding to cart. Session replays and heatmaps show them scrolling in search of sizing information. The fix is to move the sizing chart above the fold. That's the difference between knowing what happened and understanding why.

Clarity is free, with no credit card and no session cap, and it works alongside Google Analytics rather than replacing it. To start: install the tracking code (directly, or via Google Tag Manager or your CMS), check the Bot Activity and Citations reports, segment your AI-referred traffic, link Clarity to Microsoft Advertising, and set a weekly review of heatmaps and recordings for your top product pages.

Your 90-day activation roadmap

Here's how to sequence it all in two phases.

Days 1–45: Build the foundation

• Bring your complete catalog into Merchant Center, refresh it daily, turn on alerts and fix errors.
• Enable UCP: add store policies, warning attributes and merchant_item_id.
• Launch Performance Max and AI Max campaigns and enable Offer Highlights.
• Install Clarity, customize your Brand Agent, test it on a share of traffic, then roll it out to 100%.
• Choose a Copilot Checkout path, complete the integration and test the end-to-end flow.
• Connect Clarity to Microsoft Advertising and Google Ads, set up no-code Smart Events for purchases and add-to-cart, and review your AI Visibility reports.

Days 46–90: Measure, test and optimize

• Keep auditing top products and refreshing prices and inventory daily.
• Update offer attributes as promotions and fulfillment options change.
• Measure Brand Agent conversion uplift and AOV, and read transcripts for merchandising insights.
• Monitor checkout success rates.
• Set an AI visibility baseline, fix friction points such as rage clicks and drop-offs, and hold a monthly AI visibility and conversion review.

The finish line and where to start

After 90 days you'll have done what many brands are still planning, with a presence across AI search, AI shopping conversations and AI-mediated transactions, plus the data to keep improving.

None of this replaces what already works. The human web still delivers most of your traffic. But the brands that win over the next few years will treat AI assistants and agents as a real audience, one that reads structured data, cares about accurate prices and stock, and favors merchants who make buying easy.

If you only do three things this quarter:

1. Fix your feed. Submit your full catalog with accurate core attributes and refresh it at least daily.
2. Make it easy to buy. Add on-site AI assistance and, where eligible, in-conversation checkout.
3. Measure AI separately. Segment AI-referred traffic, track your citations, and learn why AI shoppers convert or leave.

Your next customer might be a person, an assistant working for that person, or an agent buying on their behalf. The work you do now decides whether your brand is the one they choose.

Source: Microsoft, "Prepare for the Age of AI Shopping: Your playbook for the era of agentic commerce" (2026). Features and availability vary by market; several are US-only or in beta.