The AI Product Discovery Playbook: From 66% Searching to 23% Trusting

The Data That Should Change Your Marketing Plan

Two reports published within 24 hours of each other paint a clear picture of how product discovery is changing.

On September 29, 2026, Emplifi and Semrush released a joint report finding that 66% of consumers have used AI to research products or discover gift ideas. That same day, Integral Ad Science (IAS) published its Holiday Shopping Report showing that AI-led product research has doubled year-over-year, with 38% of shoppers in Asia planning to use AI for festive product research.

Combined with Kantar's Media Reactions 2026 study published September 30, which found that only 23% of consumers trust AI recommendations, the picture is stark: consumers are searching via AI at scale, but they do not believe what they find.

This is the gap your brand needs to close.

How AI Product Discovery Works Today

When a consumer asks ChatGPT, Gemini, or Perplexity for product recommendations, the AI does not browse your website. It synthesizes information from multiple sources: structured product databases, review aggregators, social media mentions, and its own training data.

This means brand discoverability in AI systems depends on three things:

1. Structured product data that AI systems can parse (title, description, price, specifications, availability).
2. Social proof signals that influence what AI models surface (reviews, ratings, brand mentions across platforms).
3. Crawlable content that answer engines can access and synthesize (product pages with semantic markup, FAQs, comparison content).

If your product data is locked behind JavaScript rendering, missing structured markup, or scattered across inconsistent listings, AI systems will either skip your products or describe them incorrectly.

Step 1: Fix Your Product Data Foundation

Before any AI visibility strategy, your product data needs to be machine-readable.

Schema Markup

Add Product schema (name, price, availability, rating, review count) to every product page. This is the minimum that AI systems look for when extracting product information.

Structured Feeds

Maintain a clean product feed (Google Shopping, Shopify catalog, Facebook catalog) with consistent naming, accurate pricing, and high-quality images. AI systems increasingly pull from these feeds rather than scraping pages.

Product APIs

If you sell through marketplaces (Amazon, Shopify), ensure your product API exposes complete data. The PYMNTS/Visa GDSI study found that only 15% of SMBs have structured product data ready for AI agents. Being in that 15% puts you ahead of 85% of your competitors.

Step 2: Build AI Visibility Across Platforms

Different AI systems pull from different sources. Here is where your brand needs to be present and correctly structured.

The 73% of Google searches that now end without a click (per UpliftAI data) means traditional SEO is no longer enough. Your brand needs to exist in the answer, not just in the results.

Step 3: Earn the Trust That Turns Search Into Purchase

The 23% trust figure from Kantar is the most actionable number in this dataset. Here is how to close it.

Reviews and Ratings

AI systems weight review volume and recency heavily. A product with 1,000 recent reviews is more likely to be recommended than a product with 5,000 stale reviews. Actively solicit reviews and respond to them publicly.

Social Proof at Scale

The Emplifi/Semrush report highlights the intersection of AI and social media in product discovery. User-generated content on Instagram, TikTok, and YouTube creates the social proof signals that AI systems use to gauge brand relevance and quality.

Transparent Specifications

Consumers distrust AI recommendations partly because they cannot verify the source. Brands that publish complete, accurate, and easily accessible product specifications give consumers a way to fact-check AI claims. This builds trust in the brand, not just in the AI system.

Step 4: Measure What Matters

Traditional marketing metrics (clicks, CTR, CPC) are becoming less relevant as AI intermediates discovery. New metrics matter:

• AI share of voice: How often does your brand appear in AI-generated recommendations for relevant queries?
• AI-driven brand searches: Are consumers searching for your brand name after encountering it via AI?
• Attribution from AI sources: Can you track when a purchase originated from an AI recommendation?

Kantar found that only 51% of marketers feel they are extracting actionable intelligence from their data. The brands that fix this measurement gap first will optimize AI visibility before competitors even know it matters.

Step 5: Test Fast, Learn Faster

The Kantar study concludes with advice from Gonca Bubani: "That shouldn't stop them from experimenting with AI, but it does mean they need to carefully test, measure and learn as they invest."

Practical experiments to run this quarter:

1. Query five AI platforms for your top product category. Document what they recommend and whether your brand appears.
2. Add Product schema to your top 10 product pages. Measure crawl and index changes in Google.
3. Run a micro-creator campaign (per the Kantar shift to micro/nano creators) and measure whether AI systems pick up the social signals.
4. Track brand-name search volume before and after AI visibility improvements.

Read Next

• OpenAI Dots: Always-On Agents Are Here — OpenAI's agent platform where consumers may discover products.
• OpenClaw Enterprise: The Open-Source Agent Control Plane — Governance for agent deployments that touch product data.
• Kantar Media Reactions 2026: Marketers Bet on AI, Consumers Aren't Sure — The consumer trust data behind this playbook.
• The AI Shopping Disconnect: Consumers Ready, Merchants Aren't — Our earlier analysis of the consumer-merchant gap.

Ready to measure and improve your brand's AI product discovery? Talk to Vivid about AI visibility audits and structured data strategy.