which debuted back in 2019<\/a> – \u201cechoes how Google’s capability has evolved over three key conversational AI waves.\u201d\u00a0<\/p>\n\u201cIn the beginning, there was deterministic AI,\u201d Moger said.<\/p>\n
\u201cOlive was born as a deterministic chat and voice bot, and we’ve got a huge amount of value out of this capability. Olive actually handles over 70 percent of interactions to our contact centre. She also has a really high customer satisfaction because she’s solving for those pain points, like refunds, in a delightfully easy way.<\/p>\n
\u201cThen generative AI entered the chat. And for us, this was an opportunity to supercharge our contact centre team with meaningful AI tools. That might be knowledge retrieval, call wrap-up, and proactive nudges. We’ve been able to take 90 seconds of admin out of a call to reinvest back into the customer.\u00a0<\/p>\n
\u201cBut we hit a ceiling. We reached the limit of what deterministic AI could do, and we were struggling to figure out how to bring these generative experiences to our customers in a safe, reliable way.\u201d<\/p>\n
GECX – which, itself, is a collection of agentic technologies – is enabling the retailer to step past these limitations.<\/p>\n
Olive’s new architecture<\/p>\n
Digital experiences head of technology Ducas Francis provided a look behind-the-scenes at the GECX architecture and how the various elements powered upgrades to Olive\u2019s capabilities.<\/p>\n
AI Commerce Search is the \u201cfoundation of the agentic shopping experience\u201d, Francis said.<\/p>\n
\u201cOur existing search platform was already punching above its weight, but we just needed some more capabilities to be able to deliver this type of [agentic] experience,\u201d he said.<\/p>\n
\u201cMainly what we needed was to be able to map natural language queries to semantic intent and deliver relevant results at agent scale.\u201d<\/p>\n
Next, Shopping Agent in GECX is the \u201cprimary orchestrator for Olive’s product discovery capabilities.\u201d<\/p>\n
\u201cIt integrates directly into AI Commerce Search and uses function calling to talk to our core retail APIs, so the agent is deeply grounded in our own data and retail guardrails, which is our first line of defense as well against things like hallucination,\u201d Francis said.<\/p>\n
\u201cThanks to our strong partnership with Google, we were able to anchor the agent’s reasoning directly into our own proprietary knowledge as well, so that includes our unique recipe catalogue, our verified product and nutritional information, and things like real-time stock levels and pricing.\u00a0<\/p>\n
\u201cUsing the Google Agent SDK, we’re also able to integrate [Olive] seamlessly into the Woolworths app as [well].\u00a0<\/p>\n
\u201cThis brought a whole new set of capabilities to life, like reasoning, personalisation, multi-step tasks, and multimodality.\u201d<\/p>\n
The result isn\u2019t just what Woolworths has previously demonstrated on-stage at the US version of Google Cloud Summit.<\/p>\n
There, the demonstration showed that Olive could participate in meal planning by analysing a photo of a spaghetti carbonara and suggesting the items needed to make it.<\/p>\n
Fridge contents analysis<\/p>\n
At the Sydney version of the summit, Moger showed that Olive could analyse a photo of the contents of a customer\u2019s fridge, and suggest some ideas for what they could make with what\u2019s left.<\/p>\n
Francis said the combination of technologies in GECX meant \u201cthat when you upload a photo of your fridge and say, “What can I make?”, Olive understands that and can build a plan out of it.<\/p>\n
\u201cIt looks at the recipes that we have in our catalogue and tunes those to the needs of our customers, suggests lists of products or lists that also can be added in to make different things as well, if needed, and personalises all those results,\u201d he said.<\/p>\n
\u201cIt does all that in seconds, something that would usually take people half an hour or so, or even longer, potentially.\u201d<\/p>\n
Customer query containment<\/p>\n
Francis said that Woolworths is also using CX Agent Studio to improve the customer support capabilities of Olive.<\/p>\n
These are used to respond to customer questions and to automatically run certain processes, such as refunds.<\/p>\n
\u201cOur previous Dialogflow implementation [of Olive] was based on deterministic flows, which meant that our bot containment rate, while super impressive, was starting to hit a ceiling because of its inability to handle some more complex scenarios,\u201d Francis said.<\/p>\n
\u201cReplacing this with a modern agentic framework has allowed us to overcome these limitations, break through that ceiling, and hopefully will help us move beyond our target 80 percent containment rate for our half million conversations a month.\u201d<\/p>\n
In addition to now applying these technology upgrades to Mandy, Moger said that additional improvements are destined for Olive as well.<\/p>\n
\u201cWe gave Olive to our team first, so they’ve been using her for the past couple of months, and they were overwhelming [in their feedback] that we need better support for our in-store experience,\u201d Moger said.<\/p>\n
\u201cWe’re also going to continue to evolve our voice and vision capability to support more ways to shop.\u201d<\/p>\n
Ry Crozier attended Google Cloud Summit Sydney as a guest of Google Cloud.<\/p>\n","protected":false},"excerpt":{"rendered":"Woolworths is shaping to upgrade the \u2018Mandy\u2019 virtual assistant used in its \u2018Everyday\u2019-branded loyalty, mobile and insurance businesses…\n","protected":false},"author":2,"featured_media":767319,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[256,254,255,64,63,105],"class_list":["post-767318","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-au","tag-australia","tag-technology"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/posts\/767318","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/comments?post=767318"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/posts\/767318\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/media\/767319"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/media?parent=767318"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/categories?post=767318"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/tags?post=767318"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
![\"Woolworths](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2026\/06\/1782718225_769_ImageResizer.ashx.jpeg\")
<\/p>\n