Entering the Era Where AI Chooses Products — 10 Tactics to Win with Catalog Optimization

Can AI “see” your catalog?

Ask ChatGPT something like, “Show me a coffee maker that fits a small kitchen,” and you’ll now instantly get a recommended product list with images, prices, and purchase buttons. No scrolling or comparison search required. This is “conversational commerce,” where purchasing is completed through a conversation.

Purchase journeys through AI search engines such as ChatGPT, Perplexity, and Google AI Overviews are already a reality. And what matters most is that AI recommendations are not ads. OpenAI made this clear in April 2025: “Products are selected independently by ChatGPT and are not advertisements.”

In other words, whether AI chooses your products depends on the quality of your catalog data. Messy data, inconsistent listings, and outdated processes—all of these can cause your products to disappear from AI’s field of view.

In this guide, we introduce 10 practical tactics for preparing your business for AI search and conversational commerce.

Why catalog optimization becomes a competitive advantage

Conversational commerce is a new shopping experience in which AI acts as a “smart salesperson.” It instantly gives personalized answers to users’ questions. But the data it relies on is your product data. If that data is vague or poorly structured, AI won’t be able to find it.

There are three major criteria AI uses when selecting products.

• Alignment with user intent: Does your data answer the question appropriately?

• Structured metadata: Are price, specs, and categories organized?

• External signals: Do reviews and word of mouth reinforce credibility?

10 tactics for creating an AI-search-ready catalog

1. Structure your data for LLMs

AI prefers clean, well-organized data. Schema markup and standardized attributes help AI understand products instantly.

Action steps:

• Set up product-specific FAQ sections that answer questions like, “Is this coffee maker easy to clean?”

• Standardize SKUs and titles across platforms

Why it works: Clear data structure helps you avoid being buried by competitors’ “messy catalogs.”

2. Write content that sounds conversational

AI prefers natural language that is close to the words shoppers actually use. Rewrite titles, descriptions, and FAQs to sound like how people speak.

Action steps:

• Write intent-driven titles (example: “Compact coffee maker for single-person households”)

• Add FAQs like “Will a 15-inch laptop fit?”

• Write descriptions that emphasize benefits (example: “Brew barista-quality coffee in 5 minutes”)

Why it works: Matching AI’s query-matching logic and style makes your products more likely to be recommended.

3. Personalize content to the buyer’s situation

AI search matches products to specific needs. Context-aware descriptions improve relevance for niche queries.

Action steps:

• Create content for each variant (example: “black model for commuters” vs. “green model for outdoor use”)

• Use language that evokes usage scenarios (example: “For busy parents looking for a spill-resistant coffee maker”)

Why it works: Better matching accuracy for more specific queries increases the chances of being recommended.

4. Use user-generated content (UGC) in your catalog

AI refers to real ratings and reviews as trust signals. Authentic feedback strengthens your catalog’s credibility.

Action steps:

• Run review campaigns tied to use cases (example: “How was this coffee maker in real life?”)

• Encourage discussions in communities like Reddit and Quora

Why it works: Reviews and external mentions directly influence AI rankings.

5. Build credible backlinks and citations

AI uses authoritative sources as part of its evaluation. Coverage from trusted media increases visibility in search.

Action steps:

• Run test queries in ChatGPT to identify which publishers are being cited

• Pitch guest articles or in-depth reviews to those media outlets

Why it works: Mentions from high-authority sites increase your potential to rank in AI results.

6. Actively manage brand sentiment

AI also reads emotional tone. Negative reputation can hurt recommendations. Proactive sentiment management is key.

Action steps:

• Ask AI tools, “What’s your honest assessment of [brand name]?” to identify issues early

• Respond to misconceptions or inaccurate information with evidence-based rebuttal content such as blog posts or product FAQs

Why it works: Positive sentiment directly translates into being easier for AI to recommend.

7. Optimize off-site signals and feeds

AI crawls not only your own site but also external feeds. Open access settings and well-structured feeds help ensure exposure to AI systems.

Action steps:

• Check whether robots.txt is blocking OpenAI’s OAI-SearchBot

• Prepare structured product feeds now for future integration with ChatGPT

Why it works: AI being able to access accurate data is a prerequisite for being included.

8. Test how your products appear in AI prompts

Run test prompts in actual AI tools to understand how your products appear and where improvements are needed.

Action steps:

• Check with general prompts (example: “What are the best coffee makers under $100?”)

• Try comparison prompts (example: “[Your brand] vs. [competitor] — which is better for single-person households?”)

• Check local availability (example: “Where can I buy a backpack that arrives within two days?”)

Why it works: Prompt testing reveals gaps in your catalog’s AI visibility.

9. Create roundup and comparison content

AI uses list-style content such as “best of” roundups to generate recommendations. Comparison content that includes your products expands exposure opportunities.

Action steps:

• Create roundup articles for key categories (example: “5 recommended coffee makers for single-person households”)

• Include comparative language (example: “More durable than [competitor] for everyday use”)

Why it works: For queries like “What’s a good alternative to ___?”, AI pulls information from comparison content.

10. Leverage video content and transcripts

AI can analyze video transcripts. Multimodal content gives product context much more richness.

Action steps:

• Create short product intro videos (example: “30-second guide to using the ○○ coffee maker”)

• Always include descriptive transcripts and captions

Why it works: Video provides rich contextual information and improves AI recommendation accuracy.

Looking ahead to the future of conversational commerce

AI search is evolving rapidly. By optimizing now, you can stay ahead of the next wave of change.

• Direct merchant feed integration: In feed integrations OpenAI is developing, clean and structured data will be a prerequisite for participation.

• Agent-driven shopping experiences: In a world where AI agents handle the entire buying process, catalogs that AI can read accurately become the source of competitive strength.

• Continuous improvement through performance data: AI continuously refines recommendations based on sales data. A catalog is not something you create once and finish; it is something you keep improving.

Conclusion: Acting now creates tomorrow’s competitive advantage

AI search and conversational commerce are changing the very structure of e-commerce. Companies that optimize early will become the leaders of the next era.

The 10 tactics introduced here — data structuring, conversational content, personalization, UGC utilization, backlink building, sentiment management, off-site optimization, prompt testing, comparison content, and video use — are all actions you can start today.

Catalog quality determines whether AI chooses your products. Let’s turn that reality into a business advantage.

At Lazuli, we also support the design and improvement of product data infrastructure for AI search readiness. If you have concerns about the current state of your catalog, please feel free to contact us.
https://corporate.lazuli.ninja/contact/

Based on the original article by Bijan Vaez (Merchkit)