Marketplaces, AI agents and the race for range

It feels like every second Australian retailer at the moment is deploying or expanding a marketplace strategy.

Australians spent $18.9 billion on pure online marketplaces in 2025, according to the 2026 Australia Post eCommerce report - representing 23% of all online spend now occurring down under.

Amazon remains the dominant force, while Temu and Shein have helped fuel marketplace adoption. But increasingly, Australian retailers are becoming a much bigger part of the marketplace story too.

Wesfarmers is a good example. Across Bunnings and Kmart, its marketplaces now boast more than 400,000 SKUs and over 700 sellers. That expanded range is translating into sales, with one in four Bunnings online baskets now including a marketplace item and one in six at Kmart. BIG W market has also rapidly expanded its range and now offers more than 3 million SKUs, benefiting from the seller base from MyDeal.

And they are far from alone.

Retailers across the market are racing to scale marketplace propositions to expand range in a capital-light way, capture more customer missions and grow online audiences to monetise through retail media and it is easy to see why.

In the US, Amazon alone now accounts for an estimated 40.5% of eCommerce sales in 2025 whilst Walmart has seen much success through its marketplace offering to fend off Amazon. But as retailers continue to broaden their marketplace offers, one thought increasingly plagues me. If everyone sells everything, we are operating in an increasingly commoditised landscape. But there is a paradox here. The same expansion that risks making retailers less distinctive could also make them more useful to AI shopping agents.

Retailers have traditionally differentiated themselves, at least in part, through what they sell. Specialists build authority around a category. Department stores curate brands. Mass retailers compete on breadth, convenience and price, but their range and assortment strategies are still highly deliberate. They leverage exclusive ranges, private label and if they are EDLP retailers they often adopt the good, better, best product and pricing architecture - to guide assortment and give shoppers the right range to choose from.

But as retailers add hundreds of thousands of third-party products, they inevitably begin to carry more of the same brands, the same products and even sometimes the same marketplace sellers and those architectures can be challenged and retailers can become known for everything but remembered for nothing. For some brands, this may also prompt a rethink of where exclusivity still delivers value. With shoppers buying across more channels, brands need to weigh the benefits of selective distribution against the reach of broader availability.

The upside for the retailer is obvious. But the trade-off is that range itself can become less distinctive, the retailers reputation can be tested even when curation is occurring and, supplier relationships over decades can be challenged.

Ultimately, if the same product is available through three different retailers, the question shifts from:

Who sells what I want?

to:

Who can give me the best proposition for the same product or selection of products I am looking to buy?

That comes down to not only price but availability, delivery speed and cost, returns, loyalty benefits, trust and service and that can become more costly to service overtime as retailers compete on many fronts against a wider pool of competitors. This ultimately makes it easier for the shopper to compare overall propositions across retailers and reinforces that differentiation is delivered not at the product level but the overall proposition level.

The premise of the marketplace has always been simple: bring huge amounts of supply together in one place and make it easy for shoppers to find and buy what they need at a great price. That made the marketplace a key destination for shoppers who were looking for an easy way to discover and buy several products at a great price, for one low delivery fee.

But what happens when personal AI agents are increasingly capable of searching across the web to compare products and complete transactions on the shopper's behalf? If an agent can find what the shopper needs with ease and for the best value, marketplaces may become less important as a destination for the shopper.

Amazon recognised this shift early and has adopted a closed ecosystem approach - unwilling to share its inventory with the major AI platforms to limit their leg up, as well as introducing new features like buy for me to allow shoppers to discover and source products beyond the Amazon catalogue.  

In September, Amazon also reinforced its position, blocking Muse from shopping on Amazon.com on customers’ behalf, citing unauthorised access and privacy and security concerns. Meta Muse’s launch signifies the next big agentic commerce step, recording 2.8 million downloads within two weeks of launch, according to Sensor Tower. With Shopify store integrations already in place, and billions of users within the Meta ecosystem, Meta Muse presents a real threat to Amazon’s audience scale and it is not willing to hand over the keys to the warehouse, at least not yet.

The strategic issue is bigger than access to inventory, even though that is the ticket to the ball game. If an external agent becomes the place shoppers start and make their choices, Amazon risks losing part of the direct customer relationship (and associated data) that makes its marketplace so valuable.

This is where things get interesting. Historically, huge ranges have created complexity. Adding hundreds of thousands of products gives shoppers more choice, but also makes search, navigation and merchandising harder.

An AI agent does not necessarily face that same constraint.  It can potentially search enormous assortments against a very specific brief and reduce them down to a handful of relevant options. That could make assortment breadth and depth more valuable in a world where the agent is looking to recommend the best value and solution for the customer as it weighs up an array of options and value levers – price, range, delivery speed, loyalty and more. 

A retailer with access to 100,000 products, in part curated through its marketplace to extend core range, has 100,000 opportunities to satisfy a customer need – particularly when the shopper is looking for several items as part of that individual shop.

But SKU count alone is not the advantage. Products need to be relevant, accurately described, available, competitively priced and able to arrive when the shopper needs them. Duplicated listings or poorly performing sellers add volume without necessarily adding value.

So perhaps the future battleground is not simply who has the best AI shopping experience, but who can provide access to the broadest pool of relevant products, at the right price and value point that are actually available to buy.

But there is a contradiction.

The broader retailers make their assortments, the more useful they may become to AI agents - while simultaneously making themselves look more like everyone else.

AI may make a vast range easier to navigate, but it does not remove the retailer’s responsibility for quality. Product information, seller standards and fulfilment still need to stand up to scrutiny.

At this stage, I don’t think we know.

Personal shopping agents could weaken the marketplace as a destination while making its assortment more valuable as a source of products. Retailers therefore need to think about how they become valuable to the agent while remaining a retailer the shopper recognises and trusts.

That means making the marketplace an integrated part of the overall proposition ensuring that:

For retailers battling the likes of Amazon, agentic commerce may present a window of opportunity. But breadth alone will not be enough. Brand purpose and promise are more important than ever to guide decision making and ensure the retailer stays true to its proposition and continues to offer something differentiated in market over time.

💬 Ads are starting to talk back. Google and OpenAI have both launched new ways for brands to put AI agents directly into advertising, turning an ad from something shoppers simply click into something they can have a conversation with. Google is testing its Business Agent inside YouTube ads, while OpenAI’s Sponsored Agents allow shoppers to open a separate conversation with a brand after seeing an ad. Meta, TikTok and Snap have also been experimenting with conversational ad experiences, but an interesting divide is emerging around who gets to do the talking. Google, OpenAI and Meta are giving brands a voice through their own agents, while Amazon and Walmart are keeping the conversation within their own shopping assistants. As Roger Dunn explored in his latest AI Commerce Brief, the battle isn’t simply about making ads conversational; it’s also about who owns the customer conversation when they do.

🛒 Shopify is opening the door to AI agents at checkout. While Amazon has been restricting how some third-party agents interact with its ecosystem, Shopify is moving in the opposite direction. Its latest WebMCP capabilities allow browser-based AI agents to move beyond searching a merchant’s products and building a cart to progressing through checkout and placing an order once the shopper has authorised the purchase. It potentially creates an agent-led journey spanning discovery, cart and checkout across Shopify’s merchant ecosystem. As agentic commerce develops, we’re starting to see very different strategies emerge: some platforms are looking to control how external agents access their inventory and customer experience, while others are actively building the infrastructure that allows agents to transact.

🇦🇺 Is TikTok Shop finally getting closer to Australia? TikTok Shop continues to build out the infrastructure required to operate like a fully-fledged eCommerce platform, this week announcing a US partnership with UPS-owned Happy Returns that will allow shoppers to make eligible returns without a box or label across almost 10,000 locations. Closer to home, TikTok recently held its Take Off with TikTok event in Sydney, and there’s renewed chatter that TikTok Shop could finally be on the horizon for Australia. Admittedly, we’ve been hearing variations of this rumour for a few years now, so we’re not putting a launch date in the diary just yet. But with TikTok continuing to invest heavily in commerce globally - spanning discovery, advertising, checkout and now the less glamorous but critical post-purchase experience - TikTok Shop down under could be closer than we think.

I’m an advisor, trainer and thought leader with 25+ years’ experience across digital, marketing, loyalty and commerce, working with consumer, FMCG and retail brands. I’m also the host of Unpacking the Digital Shelf – APAC Edition and have been recognised as a RETHINK Retail Top Retail Expert (2026), and Top 20 CMO (2018).

Before you go…

🧭 Need strategic clarity or expert advice? Let’s chat.

🛠️ Looking to build capability in your team? Explore training & workshops.

🎤 Planning an event or offsite? Book a keynote or thought leadership session.

Get in touch at [email protected] or find out more here, and don’t forget to follow along on LinkedIn and Substack.