Retailers, AI search and the future of discoverability

For years, discovery on the digital shelf largely depended on broad search engines like Google, alongside retailer search and category listings built around keyword matching. Shoppers typed in what they wanted, scanned ranked lists, and clicked through options.  

That model is now shifting.  

AI-powered discovery tools are rapidly changing how Australians search for and evaluate products. PayPal Australia data suggests up to 48% of Australians now use AI platforms to search for products online, with adoption rising to 66% among under-45s. One in five Australians aged 25–44 reportedly use generative AI daily for product discovery. 

This matters because AI is not simply layered on top of existing search behaviour. It reshapes the order of discovery itself. Instead of typing keywords and scanning lists, shoppers increasingly expect guided, conversational and personalised recommendations. 

At the same time, retailers are embedding AI directly into their own ecosystems, powering on-site search, recommendations, navigation and shopping assistants. As these systems increasingly determine what shoppers see first - and in some cases whether they see a product at all - discoverability is shifting from visibility to AI selection. 

The question is no longer whether a product can be found. It is whether retailer AI systems decide it deserves to be surfaced. 

This evolution raises critical questions for brands: 

• How is AI reshaping discoverability within retailer ecosystems? 

• Which signals matter most as AI systems curate and guide decision-making? 

• How should brands rethink content, availability and digital presence to remain visible? 

• What does this mean for digital shelf management going forward? 

  
  

A closer look at Australian retailers embedding AI within their ecosystems 

Australian retailers are applying AI to discovery in different ways, shaped by their categories, customers and operating models. Yet despite these differences, a clear pattern is emerging across grocery, home improvement and beauty: retailers are increasingly defining how products are surfaced, ranked and recommended within their own environments. 

The examples below illustrate how AI is reshaping discovery inside retailer ecosystems and why brands must now optimise for retailer-specific AI logic, not just search visibility. 

Woolworths: AI discovery, personalised assistants and the power shift in retail search 

Woolworths is pushing the frontier of AI-driven discovery further than most Australian retailers, embedding advanced AI directly into the shopping journey. However, maturity is still building. 

Through its partnership with Google Cloud’s Gemini Enterprise for Customer Experience, Woolworths is evolving its digital shopping assistant, Olive, from an informational tool into a proactive concierge. Olive will be able to plan meals, interpret handwritten recipes, surface relevant specials and, with customer consent, add items directly to cart. 

This marks a clear shift in how discovery operates. Rather than returning search results, the AI captures intent and acts on it. Shoppers can express needs in natural language and receive curated product selections shaped by behaviour, loyalty data and context. 

Product titles and descriptions are the foundation of TikTok Shop discoverability. TikTok’s internal search blends keyword relevance, content engagement, and trending metadata, which means how you write listings directly impacts where (and whether) you appear. 

Test driving Olive: The Verdict 

For weekly specials, Olive initially returned generic promotions. When prompted to provide options relevant to me, it responded with personalised specials aligned to previous buying behaviour, demonstrating the integration of conversational input with first-party data. 

When asked for products to make a lasagna, Olive presented several lasagna sheet options and generated a comparison table to evaluate them. However, when asked to recommend products across multiple meals for the week, the experience narrowed, often focusing on a single meal rather than building a broader basket. 

What we learnt 

Olive represents early signs of a more curated shopper experience where fewer options are surfaced and richer product information is required to generate meaningful recommendations. 

For grocery brands, the implication is clear: structured product data and detailed content are becoming essential inputs for AI systems that select, compare and surface products. 

---

Bunnings: task-based discovery and the rise of intent-led retail media 

Bunnings’ AI approach reflects a fundamentally different shopping mindset: projects, problems and tasks. 

Rather than centring discovery on product search alone, Bunnings embeds AI across task-based navigation and recommendations, guiding customers toward solutions rather than simply surfacing SKUs. Behavioural signals and project context shape recommendations based on inferred skill level, application needs and prior interactions. 

Visibility is increasingly determined by relevance to the task at hand and compatibility within a broader solution set - not just category dominance. 

Test driving Bunnings AI: The verdict 

We tested Bunnings AI with project-led queries. 

When prompted with “I’m looking for paint for my bedroom,” the assistant asked clarifying questions about colour preference, finish type and price range before narrowing options. 

When prompted with “I’m looking to stain my wooden garden furniture,” it gathered detail on stain colour and surface requirements before recommending products. 

The experience is consultative rather than catalogue-driven. Discovery unfolds through progressive clarification - problem → preferences → solution - reducing choice early to surface a smaller, more relevant set of options. 

Importantly, contextual fit is prioritised over bestseller rankings. 

What we learnt 

Bunnings AI signals a shift from product-led search to solution-led discovery. 

For brands, this changes what drives visibility: 

• Task and use relevance matter more than popularity and this is core | key to driving visibility. 

• Structured attributes such as finish type, surface suitability and application context act as critical inclusion signals in categories like paint. This increases the importance of expanding the depth and precision of attribute data captured at product level. 

• Inclusion in the shortlist matters more than position in a long results page. 

---

Adore Beauty: AI-supported discovery through service and expertise 

Adore Beauty’s AI approach reflects the importance of trust and expert guidance in beauty. 

Rather than positioning AI as a fully autonomous discovery engine, Adore Beauty has embedded AI primarily within customer support. Its AI-powered assistant, Abi, reportedly handles 58% of customer queries, improving response times and freeing human agents for more complex advice. 

This represents a different maturity model: AI as triage and operational layer rather than proactive curator. 

Test driving ABi: The verdict 

We tested ABi with a common query: “I’m looking for a heat protectant spray.” 

Rather than asking clarifying questions about hair type, styling routine or finish preference, ABi provided a general statement about product range and directed us toward live chat or human-assisted support for personalised recommendations. 

Shortly after, the interaction was handed to a human specialist who provided tailored product suggestions and links. 

In practice, the AI layer functioned less as a discovery engine and more as a routing mechanism by capturing intent and escalating to expert advice rather than narrowing options independently. 

What we learnt 

Adore Beauty highlights an AI model focused on operational efficiency and expert enablement. 

For brands, this means discoverability may still depend heavily on strong content fundamentals user by humans to provide recommendations to the end user. Clear benefits, structured attributes and concise positioning remain critical because recommendations are often mediated by human advisors drawing on available product information. 

AI integration is not uniform. In some ecosystems, AI drives proactive curation. In others, it supports service flow while humans remain central to decision-making. 

---

How discoverability signals are evolving inside retailer sites 

Across these examples, a clear pattern emerges. 

Discovery inside retailer environments is no longer governed by static keywords, universal rankings or fixed category hierarchies. It is shaped by AI systems that continuously interpret product data, behavioural signals, contextual inputs and commercial performance data in real time. 

Four shifts define this evolution: 

1. Discovery is driven by inferred intent, not declared keywords 

Traditional search depends on shoppers explicitly stating what they want. AI-led discovery infers intent from a broader set of signals, including browsing behaviour, purchase history, and conversational inputs. 

For brands, this elevates the importance of structured attributes and clearly defined use cases. AI systems rely on machine-readable signals that will differ by category in some instances, such as format, function, suitability, compatibility, finish, skin type, surface type or dietary markers to determine relevance in context. 

Optimising for intent requires making products legible to machines, not just compelling to humans and a clear understanding of how conversations occur in an AI driven discovery world.  

2. Visibility is conditional, not universal 

In AI-driven environments, there is no single ranking that applies to everyone. 

Personalisation engines tailor results based on behaviour, loyalty status, location and prior interactions. The same product can be highly visible to one shopper and absent for another. 

Discoverability becomes an ongoing performance outcome rather than a fixed position. Brands must earn repeated inclusion across different shopper contexts rather than optimising for a single “rank one” moment. 

3. Discovery happens across the journey, not just at search 

Search is no longer the primary or only entry point. 

AI-driven discovery now occurs through navigation flows, curated collections, related product modules, cart prompts, replenishment reminders and contextual suggestions. Each interaction feeds back into the system, shaping future exposure, creating a compounding effect. 

Products that engage and convert are more likely to be re-surfaced across modules. Those that underperform are gradually deprioritised. 

Discoverability is therefore not a moment. It is a performance loop. 

4. Data quality directly affects AI confidence 

AI systems rely on clean, consistent and differentiated product data to make confident decisions. 

Research into Australian retailer catalogues shows that many product pages still fail basic search benchmarks, often due to duplicated descriptions, missing attributes or inconsistent metadata. 

In AI-led environments, data quality is no longer hygiene. It is a performance lever. Products that are easy for machines to interpret are more likely to be trusted, surfaced and reused across discovery surfaces. 

What brands must optimise for when retailers control search, ranking and recommendations 

With retailer AI systems now determining what gets surfaced, brands must rethink the optimisation process. 

Visibility is no longer a function of keyword ranking alone. It is the result of structured data, performance signals, availability and contextual relevance inside each retailer ecosystem. 

That shift demands new priorities. 

Rich, platform-ready product content 

Brands must invest in high-quality, unique and structured product content, including attributes, specifications, benefits, use cases, imagery and semantic metadata. This content underpins how AI systems interpret relevance and decide what to surface. 

However, quality alone is no longer enough. Retailer AI systems interpret signals differently, shaped by their own data, customer behaviour and commercial logic. To remain visible, brands must adapt product content for each retailer ecosystem, tailoring attributes, titles, descriptions and visuals to the way discovery actually works on that platform. 

Without both depth and contextual optimisation, products risk being misunderstood by AI systems and quietly deprioritised in search and recommendations. 

Personalisation-ready signals 

Retailer AI platforms increasingly draw on behavioural and loyalty data to personalise results. 

Brands that understand how signals such as purchase frequency, replenishment cycles, category affinity and contextual behaviour influence recommendations can better align their content and catalogue architecture to these patterns. 

Real-time inventory and availability 

AI systems prioritise products that are consistently in stock and available locally. 

High out-of-stock rates or unreliable fulfilment performance can reduce a product’s visibility over time, as AI models learn to suppress items that create friction. 

Inventory reliability is therefore not just operational hygiene. It is a discoverability lever. 

The future of discoverability 

AI is becoming the gatekeeper of product visibility inside retailer ecosystems. 

Discovery is moving from open search environments into curated shortlists. In these environments, visibility is earned through structured data, performance consistency and contextual relevance. For brands, this requires a fundamental shift. 

Optimisation must move beyond legacy SEO thinking toward AI-ready content, real-time availability, be customised to suit individual retailer environments and more.   

Those that adapt early will not simply be found. They will be prioritised inside the AI-driven journeys that increasingly define modern retail. 

---

At Arktic Fox, we work with brands to make their digital shelf AI-ready by strengthening structured product content and retailer-specific optimisation. Need help with your digital shelf strategy? Let’s chat.