Woolworths’ AI customer problem-solving a question of semantics


Woolworths is introducing semantic search capabilities into its online shopping portal to catch a trend that is seeing consumers increasingly use ecommerce sites as problem-solving tools.



Semantic search is a subset of AI technology that uses advanced mathematical techniques and natural language processing to reveal the intent behind online queries to generate results, rather than relying on traditional keyword matches.

Speaking at Elastic’s Elasticon Tour in Sydney, Woolworths experience product lead Meredith Murphy said that use of the portal had progressed with customers now expecting it to provide more than dry merchandise and pricing information.

Instead, she said, customers were seeking answers based on more sophisticated enquiries that which require a deeper understanding of their individual missions.

That could, for instance, be information about cooking or opening hours in particular areas, she said.

“I think when you’re thinking about scaling support and using AI, it’s how you take a customer search term, unravel that and identify what the intent is of that search and deliver on it. We’re proactively solving a what a customer wants before it becomes a problem,” Murphy said.

However, she said that the supermarket giant still remained cautious about letting AI make the search results overly complicated for customers, especially in cases when their queries could be handled in simpler ways.

“Sometimes the search query is really straightforward, and I think often it gets forgotten when we’re talking about natural language, that not everything needs to be inherently complex. It’s about having a really qualitative set of results which are relevant and are ideally personalised to the customer experience.

“We’ve given them the right tools to refine that result so that they’re relevant to their shopping history, etcetera. What we’re doing is bringing in semantic capabilities and also starting to add features through the [user experience] such as auto applying filters.

“If we know you’re seeking specials, we can apply that filter proactively for you. It’s that fusion between what we can do with technology unpacking a complex query and then still giving the customer that element of control,” she said.

As iTnews reported earlier this year, Woolworths has positioned itself to start incorporating Google’s agentic AI into its portal’s chatbot, Olive.

Woolworths chief executive announced in January that Woolworths was would be the first Australian retailer sign-up for Google’s Gemini Enterprise for Customer Experience.

Murphy’s comments resonated with those Bardwell made at the time the retail giant announced its partnership with Google.

“As the first Australian retailer to partner with Google’s agentic platform, we are evolving our digital shopping assistant Olive into an intuitive partner that won’t just answer questions, but actually anticipates your needs,” Bardwell said.

Elaborating, Bardwell said that the agentic capability would give its virtual assistant the ability to help customers plan meals and look for bargains.

It’s not clear to what extent Woolworths will seek to incorporate the capabilities Google’s agentic customer experience technology into its ecommerce portal.

According to Google, however, its shopping agent is able to handle complex reasoning, to parse image, video and voice inputs, and accommodates “consented actions”, including  adding merchandise to digital shopping carts.

Woolworths’ interest in semantic search capabilities arrives amid a broader shift which is seeing some major brands attempt to boost engagement with consumers by experimenting with other types of enhanced search.

For example, digital property advertising specialist REA Group recently started offering conversational search capability on its flagship website realestate.com.au.



Source link