How do I track ChatGPT and Perplexity traffic and sales in Shopify and GA4?

What part of AI traffic can you actually track?

You can track clicks that reach your store from an AI assistant; you cannot track every time an assistant mentions you. That distinction matters. A shopper who reads a ChatGPT answer, then searches your brand on Google two days later, shows up as branded search. Treat tracked AI referrals as a measurable floor, not the full effect.

ChatGPT makes the floor easy to measure. OpenAI's publishers FAQ says ChatGPT automatically includes utm_source=chatgpt.com in referral URLs so publishers can track traffic in tools such as Google Analytics. Because the tag is in the URL itself, it survives cases where the browser drops the referrer header. Links a shopper copies out of ChatGPT and pastes elsewhere will still lose it.

How do you separate paid ChatGPT ads from organic AI traffic?

Separate them by UTM medium before anything else, or your AI search numbers will include paid clicks. OpenAI's ads basics says advertisers can add static tracking parameters such as UTMs to landing page URLs and that they persist on ad clicks. On the stores we run, ChatGPT ad clicks arrive tagged utm_source=chatgpt with utm_medium=cpc and an oppref click ID, while organic answers arrive as utm_source=chatgpt.com. The difference is one ".com".

Then report the two side by side. Paid ChatGPT is a media buying line with a cost per order. Organic AI referrals are the output of AI search work and have no click cost. Mixing them flatters both.

How do you set up an AI assistants channel in GA4?

Create a custom channel group with an AI assistants channel, then order it above Referral. Google's custom channel groups help page walks through this exact example and notes that traffic is included in the first channel whose definition it matches.

1. In GA4, go to Admin, then under Data display click Channel groups.
2. Copy the default channel group or create a new one, so the default stays untouched.
3. Add a channel named ChatGPT Ads: session source exactly matches chatgpt AND session medium exactly matches cpc (adjust to the UTMs you set on your ads).
4. Add a channel named AI assistants with a condition on session source that matches regex. We use an explicit domain list: chatgpt\.com|chat\.openai\.com|perplexity\.ai|gemini\.google\.com|copilot\.microsoft\.com|claude\.ai
5. Click Reorder and drag ChatGPT Ads to the top, then AI assistants above Referral (Google also suggests considering placement above Organic Search).
6. Save the group, then open Reports > Acquisition > Traffic acquisition and switch the primary dimension to your new channel group. Custom channel groups apply to your reports retroactively, per Google.
7. Add purchases, total revenue and key event rate as metrics, and confirm Shopify purchase events are reaching GA4 before trusting revenue.

Google's own example regex is broad (it matches patterns like any source containing gpt or copilot). Broad patterns can catch unrelated sources, so test any regex against your full source list before relying on it, and update it when new assistants appear, as Google's page also advises. GA4's traffic acquisition report is session-scoped, which is the right view for which source drove the visit that converted.

For Google's own AI features, AI Overviews and AI Mode clicks generally sit inside Google organic. Google points site owners to the Generative AI performance report in Search Console to see how content performs in those features.

Where do ChatGPT sessions and orders show up in Shopify?

Shopify shows AI traffic in its marketing and sessions reports, grouped by UTM source or referrer, and on each order. Shopify's marketing reports documentation confirms data can be grouped by UTM campaign, source, medium or referrer, with referrers split into first and last interaction.

1. Open Analytics > Reports and look at sessions by referrer and sales attributed to marketing. Filter for chatgpt, perplexity, gemini and copilot.
2. Compare first-interaction and last-interaction views. AI often introduces a shopper who converts later through email or branded search. Shopify resets the first-interaction referrer if a visitor does not buy within 30 days of a session.
3. Open any individual order and check its conversion details to see the recorded source for that sale.
4. If you sell through Shopify's agentic storefronts, orders from AI channels display in your admin with channel or referrer attribution, per Shopify Help.

ShopifyQL: the query we actually run

For a clean monthly view we query sessions and orders by UTM source with ShopifyQL (schema in Shopify's ShopifyQL docs). A sessions query looks like: FROM sessions SHOW sessions, conversion_rate WHERE utm_source = 'chatgpt.com' GROUP BY landing_page_path SINCE 2026-08-01 UNTIL 2026-08-31 ORDER BY sessions DESC LIMIT 30. For orders, query the sales dataset on order_utm_source = 'chatgpt.com'. Use explicit dates; relative single-day ranges can return empty results. If you only sell in one market, add a country filter so crawler and non-target traffic does not dilute the rate.

This is how we measured the shift at one supplement store we run: ChatGPT visits grew 6x in under a year (141 sessions in September 2025 to 885 in August 2026, US Shopify data), and most landed on product pages. The landing page split is the most useful output, because it tells you which pages AI is sending shoppers to and therefore which pages must convert.

Does ChatGPT traffic convert better than Google organic?

Sometimes, and the published studies disagree, so measure your own. The table puts the main datasets side by side. They use different samples and definitions and should not be averaged.

Practical Ecommerce's comparison explains much of the gap: Kaiser and Schulze used last-click attribution across 49 countries and isolated ChatGPT, while Adobe measured post-click behavior on US data and grouped AI tools broadly. Our own first-party number sits on the high side: at one brand we work with, ChatGPT traffic converted at 4.99%. We do not present that as typical. It is the reason we always look at the landing pages AI sends traffic to before we look at the channel average.

How we verify AI traffic in our own server logs

Analytics shows the clicks; server logs show the visits AI makes before anyone clicks. We read both, because the logs catch what GA4 and Shopify never see, and they are the only way to tell a real AI agent from an impostor.

1. Pull raw access logs from your host or CDN and filter by user agent for the OpenAI, Perplexity and Google agents you care about.
2. Verify every hit against the published IP ranges for that bot. A matching user agent from an unlisted IP is not the bot.
3. Drop spoofed traffic. On our own site, most requests claiming to be AI bots were scanners faking the user agent. Counting them inflates AI interest and hides real crawl problems.
4. Separate the live agent from crawlers. ChatGPT-User is ChatGPT's live browsing agent, so its hits line up with conversations rather than background crawling.
5. Join logs to referral data. Match verified fetches by timestamp to sessions tagged utm_source=chatgpt.com, and keep paid ChatGPT clicks (advertiser UTMs such as utm_medium=cpc plus an oppref click ID) out of the organic line.

What it looked like on succession.media

Between September 14 and 24, 2026, ChatGPT-User requests verified against OpenAI's published IP ranges opened our homepage 21 times, about twice a day. One became a click-through tagged utm_source=chatgpt.com, and that visitor booked a strategy call about a minute after landing. Twenty-one fetches and one click is a small sample, and we present it as a sample, not a rate. What it shows is the chain working end to end: the fetch in the logs, the tagged session in analytics and the booking, all lined up by timestamp.

Crawlability belongs in the same check. If key content only renders with JavaScript, an AI agent may read an empty page. Animated counters are a common case: they show 0 in the static HTML, so the agent reads 0. We render the real values server side. The full technical list is in our guide on llms.txt, schema and robots.txt for Shopify, and our take on what this traffic is worth to a budget is in is GEO worth it for a DTC brand.

What should a monthly AI traffic report include?

A useful monthly report shows AI referrals as a channel with the same metrics as paid and organic, plus the context tracking cannot capture. This is the layout we use for clients.

• Organic AI sessions, orders, conversion rate and revenue by engine (ChatGPT, Perplexity, Gemini, Copilot).
• Paid ChatGPT ads as a separate line with spend and cost per order.
• Top 10 AI landing pages with their conversion rate against the site average.
• First-interaction AI orders, to show assisted value that last-click hides.
• Share of prompts where the brand is named, from a fixed prompt set re-run monthly. Our guide on checking whether ChatGPT recommends your brand covers the method.
• Branded search trend from Search Console, as a proxy for mentions that never click.

If you want this built and reported for you next to your ad numbers, that is part of our AI search optimization work. The free growth audit includes a check of how your current setup classifies AI traffic.

FAQ