Short answer: Yes, but fix your product pages first. Adobe reported on October 29 that in September, visitors arriving at US retail sites from AI tools such as ChatGPT generated 83% more revenue per visit and converted 90% better than other traffic. Small online stores in Taiwan don't need monitoring tools yet; start by writing product data, spec tables and return policies clearly enough for AI to read.
On October 29, Adobe Digital Insights published new figures on its official blog: in September 2026, AI-referred traffic to US retail sites grew 71% year over year, and those visitors converted 90% better than non-AI traffic and generated 83% more revenue per visit, both the widest gaps Adobe has recorded. A year earlier, the same AI visitor was worth less than half a typical visit. The data covers more than 1 trillion visits to US retail sites and 100 million SKUs, plus a survey of over 5,000 consumers.
Why now? AI shopping has moved from curiosity to "done my homework, ready to buy." Adobe's August 2025 report still showed AI visitors converting 23% worse than average. This year, shoppers compare specs and reviews inside ChatGPT, Gemini or Perplexity and arrive close to purchase. Adobe also forecasts US online holiday spending of about $275.1B, up 6.7%, with AI-referred traffic up about 130% and as much as 159% on Thanksgiving. Volume is still small, but each visit is worth more.
Platforms are competing for these high-value visitors. Google's AI Mode routes many shopping links to its own surfaces; a study summarized by Search Engine Roundtable on October 30 found about 57% of product links point to Google Shopping. OpenAI is building shopping and checkout inside ChatGPT, and Perplexity has its own shopping features. The category is moving from "AI helps you search" to "AI picks for you, and may buy for you," and whether a brand gets recommended depends on whether machines can read its product information.
What does this mean for small online sellers in Taiwan? US data is not Taiwan, but the direction is the same: fewer, better-qualified customers will arrive from AI. Below we cover what Adobe's numbers really mean, which pages to fix first, and how to track AI traffic without paying for subscriptions.
The details: fewer AI visitors, but each one is worth more
The story is a reversal in quality, not a traffic explosion. Adobe's September figures:
- Conversion: AI traffic converts 90% better than non-AI traffic; in August 2025 it converted 9% worse.
- Revenue per visit: 83% higher, and Adobe expects the gap to double during the holiday season.
- Engagement: 73% more time on site, 25% more pages per visit and a 53% higher add-to-cart rate.
- Readability: Adobe's new Citation Readability Score puts the average US retail homepage at 63%, electronics at 79%, grocery at 55% and general merchandise homepages at only 40%.
The last point is the most practical. Among retail sites, some pages are readable by AI and some are not. Electronics scores high because spec tables, comparison tables and support pages are already structured. Brands with clear data win; sites relying on images and promotional copy lose.
What each reader should do now: find the pages AI can't read
Everyone starts the same way: check whether AI can pull price, specs and purchase terms from your pages.
- Brand owners and SMB owners: pick 10 key products and ask ChatGPT and Gemini to "recommend a [product] at [price]." Record whether you appear and whether the details are right, and fix the worst ones first.
- Marketers and SEO practitioners: create an "AI sources" channel group in GA4 combining chatgpt.com, perplexity.ai, gemini.google.com and similar referrers, then compare conversion rate and order value monthly.
- Developers and agencies: check that product pages have Product structured data, that price and stock are text rather than images, and that specs are in tables. Fixing those three helps both AI and Google Shopping.
Tool comparison: free GA4 covers about 70% of AI traffic tracking
Small stores should build a baseline with free tools and consider paid monitoring only once monthly revenue passes about NT$1M.
| Tool | What it does | Cost range | Best for |
|---|---|---|---|
| GA4 (custom channel group) | AI-source visits, conversions and revenue | Free | Every store; start here |
| Adobe Analytics | Full source analysis for large sites | Enterprise quote | Brands with very large revenue |
| AI visibility tools such as Profound or Scrunch | Track how often your brand appears in AI answers | Monthly subscription, varies by plan | Multi-brand teams needing competitor benchmarks |
| Microsoft Clarity | See clicks and scrolling of AI visitors on site | Free | Teams redesigning product pages |
What the headline won't tell you: 83% is an average of large US sites
Keep at least three limits in mind:
- Small base, inflated ratios: AI is still a minority of retail traffic, and it converts well partly because only people who have already decided click through. Ten times more AI traffic would not mean ten times more revenue.
- Only clicks are counted: Adobe measures AI traffic by link clicks. People who read the AI answer and then buy in a store or on a marketplace aren't counted, so the real impact could be larger or overstated.
- Taiwan's market is different: much of Taiwan's online buying happens on Shopee and momo, and AI recommendations often link straight to marketplace listings, so a brand's own site may not be the beneficiary.
The no-subscription alternative: one product dataset, used three ways
The cheapest approach is to clean up product data once and feed it to your site, Google and AI at the same time.
- Step 1: list key products in a spreadsheet with name, price, specs, intended use, warranty and return terms.
- Step 2: present that information on product pages as text and tables, and add Product structured data as documented by Google.
- Step 3: upload the same data to Google Merchant Center, and publish a clear FAQ and returns page.
- Step 4: compare AI-source conversion with other sources in GA4 each month, and decide on paid tools after three months.
FAQ
Do Taiwanese websites get AI-referred traffic too?
Yes, though the share is still low for most sites. Filter for chatgpt.com, perplexity.ai and similar sources in GA4 to see your actual numbers.
How do I get ChatGPT to recommend my products?
Nothing guarantees it, but you can raise the odds. State price, specs and use cases in text, and build third-party reviews and mentions.
If AI traffic converts better, should I cut ad spend?
Not yet. AI volume is still small, so treat it as an extra channel and watch three months of data before reallocating budget.
Do I need a developer for Product structured data?
Most e-commerce platforms include it. A custom site takes about 1–3 working days to set up, then verify it with Google's Rich Results Test.
My take: chase "AI-readable," not "AI traffic"
The mainstream takeaway will be "AI traffic converts twice as well, so grab AI visibility now." My view: for small Taiwanese online sellers, AI traffic will stay a single-digit share for the next 12 months, and chasing AI visibility tools won't pay off. What matters is Adobe's unglamorous readability score: turning product pages from pretty images and promo copy into specs and terms machines can read. That improves Google Shopping, on-site search and customer service at the same time, and AI recommendations come as a bonus.
For ScriptWalker (a Taiwan-based Laravel/Flutter custom development studio), the service opportunity is clear: a "product data health check and structured redesign" package that audits 50–200 product pages, adds spec tables and Product structured data, connects Merchant Center and delivers a GA4 AI-source report. That is easier to accept and price than vague "GEO optimization."
Sources
- Primary: Adobe: AI Traffic Drives 83% Higher Revenue Per Visit (2026-10-29)
- Primary: Adobe Digital Insights 2026 Holiday Forecast (PDF)
- Primary: Adobe: Generative AI-powered shopping rises with traffic to retail sites (2025-08)
- Primary: Google Search Central: Product structured data
- Third party: Search Engine Roundtable: Daily Search Forum Recap (2026-10-30)
- Third party: TechCrunch: AI traffic to US retailers rose 393% in Q1 (2026-04-16)
Want an AI-readability and structured data check for your product pages? Contact us:
- Email: [email protected]
- Phone: 0916-224-047
- LINE: @ufv9089p