What if we could make the whole of global commerce 20, 30, 40% more effective?

Algolia Chief Technology Officer Sean Mullaney poses this possibility and discusses the impacts of AI on ecommerce in this new episode of the AI Today podcast.

That seismic effectiveness shift is entirely possible — if not likely — with the rollout of AI, and one area where this could happen in spades is ecommerce search.

In this lively conversation, hosted by Kathleen Walch and Ronald Schmelzer, Mullaney touches on:

• The magic of vectorization, the new technology that turns words into concepts and  enhances the relevance of ecommerce search results
• How hyper personalization is changing everything
• What conversational commerce means
• The future of AI in general

Search in the Dark Ages

As recently as a few years ago, ecommerce sites and marketplace sites alike were characteristically overflowing with merchandise, but site-search technology was still based on simple keyword matching.

Search engines couldn’t grasp context or process all the shopping parameters that someone might want, such as an item in stock, in a particular color and a certain version, and hey, how about free shipping? Entering a query in an ecommerce search box produced a slew of possibly decent results, and shoppers often had to sort through the deluge, perhaps backtracking and re-searching, to find exactly what they wanted.

“Just give me an option that’s good enough, and I want this price range,” says Walch. “And then you come back with, like, 30,000 options.”

This problem of “analysis paralysis” has led shoppers to stash items on their wish lists (as Ronald admits he does), to evaluate for a while, or worse, simply go find another online retailer.

AI meets search

Mullaney says that two applications of AI, in particular, relate to upgrading search functionality:

• Asking a question: “generating” a response in a natural-sounding way
• Retrieving information about items in an ecommerce catalog

The search industry renaissance

Instead of matching keywords, AI and new large language models can turn keywords and phrases into concepts. Vectors — what’s powering ChatGPT — mathematically represent these concepts.

Thanks to breakthroughs in natural language understanding (NLU), people on ecommerce sites can ask for what they want by just talking naturally. And the software can then respond in a natural, conversational way as well.

Now, people using ChatGPT, for instance, feel like they can relate better when they talk to a computer. And as a result, they often search using long-tail, or descriptive, queries, like “I need a pair of jeans in size 14.”

When someone enters a long-tail query, “they’re telling you ‘I really want this thing, I’m describing it to you exactly,’” says Mullaney. In fact, he adds, in the past six months, users of sites powered by Algolia search have doubled the numbers of keywords they’re entering. Why? they feel like they’re being understood by the search engine, so they try to provide the mechanism with as much detail as possible.

The “long-tail [keywords]…should be the ones that convert into sales the most easily,” continues Mullaney. “And about 80% of the queries that are happening on these ecommerce sites are these longer queries.…with our vector search approach, we’re getting, like, 30% more sales….Customers are…rewarding this with their loyalty and their money.”

Hyper personalization

Another beauty of AI is that it lets ecommerce sites treat each shopper as an individual — an advertiser’s dream, notes Walch.

In this era of hyperpersonalization, in terms of search, says Mullaney, you can use all available clickstream data from a user’s previous sessions to rewrite the next session and personalize recommendations and other aspects of the shopper journey. You can rerank, pull from brands the shopper has bought in the past, adjust for the person’s typical price points. You do need to have AI from end to end, he adds.

“Conversational commerce”

Generative AI, as in many sectors, is slated to profoundly impact search technology. By feeding information into a generative AI model, you can create the experience of a personal shopping advisor that talks to you about products, answering your questions, giving expert advice, and making smart suggestions.

Algolia calls this phenomenon conversational commerce. “Generative AI is going to lead you through the buying process,” explains Mullaney.

What could be better than having a competent personal shopper help you buy some fancy running shoes — maybe in the middle of the night when stores aren’t open — without  having to leave your computer (or mobile phone)?

Lots of exciting perspective on what the power of AI can do to transform search, and “Algolia is in the center of making all that happen,” concludes Ronald.

Check out the complete interview with Mullaney.