AI didn't replace our marketing. It changed where we compete.

Inside AU Beauty Bazaar's AI playbook: a $3.4M run-rate, Google Ads ROAS up from 2.4x to 9.3x, and where artificial intelligence does, and doesn't, belong in ecommerce marketing.

Every ecommerce founder I speak to is asking the same question: if AI can write the ads, the product copy and the strategy, where is the edge? After two years running AU Beauty Bazaar, an online retailer of Korean and Japanese skincare, my answer is simple. AI hasn't replaced our marketing. It has changed where we compete.

We launched in October 2024 and are now at an AUD 3.4 million annual run-rate. Over that time, our Google Ads return on ad spend has risen from 2.4x to 9.3x. Here is what AI actually changed for us, and what it didn't.

Being found now means being found by AI

The biggest shift in ecommerce search isn't on Google's results page. It's in the chat window. Comparing July to September 2026 with the same quarter in 2025, visits from AI assistants such as ChatGPT, Gemini and Perplexity rose 75%, from 246 to 431. Orders from those visits more than quadrupled, from 5 to 23. ChatGPT alone accounts for more than nine in ten of those visits.

Google is changing too. Comparing July to September 2026 with the same period in 2025, our Google search impressions rose 24% to just over one million, and our average ranking position improved from 17 to 11. Yet clicks stayed flat at around 7,700. More people are seeing us, but more of them are getting their answer before they click, often from an AI summary.

The clicks that do come are worth more. Orders from organic search rose from 90 to 147, and organic search revenue grew 85%.

The lesson for ecommerce brands: ranking is no longer the finish line. Your product pages need to answer the specific questions shoppers ask an AI, such as which Korean serum suits sensitive skin, or what "botox in a bottle" actually means. Several of our top search terms are exactly that specific, including "korean botox serum" and "korean botox in a bottle".

AI for speed to market

Our most valuable use of AI isn't writing. It's research. We use AI to scan what's trending in Korean and Japanese beauty, spot products gaining momentum, and move fast to get them into Australia before our competitors.

The effect shows up clearly in our numbers. Our two biggest jumps in return on ad spend both followed new product launches where we were early to market. In February 2025, ROAS rose from 3.6x to 8.3x. In June 2026, it rose from 5.6x to 11.4x, in our record month for orders. Being first with the right product did more for us than any increase in ad budget.

Ignoring AI's default advice

Ask any AI tool how to grow an online beauty store and you'll get the average of the internet: run sales, post every day, chase volume. For a reseller with tight margins, competing on price is a race to the bottom.

We chose to compete on curation and customer service instead. We stock a carefully selected range, build routines and sets around real skin concerns, and invest in helping customers choose well. Our average order value has risen from $66 to $95, without competing on price. AI gave us speed; it didn't give us our strategy.

Beating AI fatigue

Shoppers can now spot AI-written copy a mile away. When every beauty brand's ads and product descriptions sound the same, specialist knowledge is what stands out: which formulas hold up in an Australian summer, which suit sensitive skin, and what a product can and can't claim. In skincare, sunscreens are regulated as therapeutic goods in Australia, which limits what you can say in an ad. A generic AI draft doesn't know that.

We use AI for the generalist work, such as reporting, customer service drafts and supplier coordination, and keep the judgement calls with people who know the category. With detailed playbooks and AI tools, our virtual assistant's productivity rose about five-fold. A small team can now do the work of a much larger one.

The results

Source: AU Beauty Bazaar's Google Ads, Google Analytics and Google Search Console accounts. All figures in AUD, self-reported by the business.

What ecommerce brands can take from this

1. Use AI for research and speed, not strategy. Let it find what's rising in your category, then decide with your own judgement.
2. Write for the questions shoppers ask AI. Product pages that answer specific, real questions get found by search engines and AI assistants alike.
3. Track AI as its own channel. Check where ChatGPT, Gemini and Perplexity visits land on your site, and what they buy.
4. Don't follow the default playbook. If every AI tool gives your competitors the same advice, doing the opposite with conviction can be your edge.
5. Make specialist knowledge visible. In a sea of generic AI copy, expertise is the most persuasive thing you can publish.

Business Spotlight features are contributed by operators in our network and reflect their own experience and views. Intrigue Digital does not verify the internal metrics or figures reported and they should be read as the author's own account, not independently audited results.