Perplexity AI Shopping Assistant has arrived. Here's how to prepare your catalogue for superior visibility and findability
LLMs/RAG are particulary good at product research, will Perplexity close the loop and convince people to transact "here and now"?
Perplexity Shopping has arrived.
A new AI personal shopper beyond Amazon Rufus.
Visual product cards, dynamic feature charts, comparison tables, and snap & shop features make shopping, particularly product research, a lot more conversational and fun.
In our own language.
From discovery to transaction in a single space. Thanks to one-click checkout. Pre-bundled with Shopify and other ecommerce platforms.
Same as other marketplaces, great visibility and product findability starts with a great catalog. This is done via the Perplexity Merchant Program.
Make sure your catalog includes:
👉 Detailed product descriptions to perform well in dynamically built charts and comparison tables.
👉 LLM inferred customer-led attributes using semantics AI from product reviews, FAQs and other relevant content.
👉 High quality photos, taken from different angles to do well at Snap & Shop visual search.
👉 Related content links to blog posts, youtube videos etc
👉 Don’t forget adding structured data, such as EAN codes, GTINs, MPNs and other model references.
What’s my take?
I enjoyed the experience, found it particularly useful for product research during early discovery stages of the shopping journey.
Charts and comparison tables are great, yet due to their structured nature, I would still complete research (double checks) in other sources.
As per CX, my initial perception was too much text and very little images/video. Ecommerce is a highly visual experience and I’d suggest to keep exploring the design of vertical shopping UIs per category/industry where products predominate.
LLMs + RAG are great for intent formulation in a question and answer format, as if you were speaking to a sales associate, yet, are they trustworthy enough to transact and jump straight into purchase?
Needless to say, same as with AI agents, levels of trust vary depending on the product category and job to be done.
All in all, great to see how quickly the space of AI shopping assistants is heating up!