How Do I Measure ROI From AI Search Visibility?

You measure AI search ROI by adding up two things and comparing them to what you spent. The first is direct value, which is the enquiries and sales GA4 can trace to AI tools like ChatGPT and Perplexity. The second is influenced value, which shows up as growth in branded searches, direct traffic and customers who say "I found you on ChatGPT" when you ask how they heard about you.

The formula is simple. The hard part is being honest about the numbers. So rather than start with theory, let's start with a business.

Meet Leanne and her Brisbane plumbing business

Leanne runs a plumbing business in Brisbane's northern suburbs with four vans on the road. Her average job is worth about $650 and roughly one in three quotes turns into a job. Like a lot of tradies she gets most of her work from Google, word of mouth and a couple of directories.

Early last year a customer told her they'd asked ChatGPT for "a reliable plumber near Chermside who does hot water systems" and her business was one of three names it gave. That got her attention. She decided to spend some money making sure AI tools understood her business properly. She worked with an agency on her service pages, her Google Business Profile, her reviews and her directory listings. All up it cost her 9,000.

Six months later her accountant asked the obvious question. Was it worth it?

Step one: she set a baseline before judging anything

The smartest thing Leanne did happened before she spent a cent. She wrote down her numbers for the 90 days before the work started. Branded Google searches for her business name, direct visits to her website, AI referral visits in GA4 and how many new customers she got each month.

This matters more than people think. If you don't know where you started you can't tell whether anything changed. And three months is about the minimum to smooth out seasonal ups and downs. A plumber's winter in Brisbane looks very different to summer, so comparing July to January tells you nothing.

Step two: she counted the direct value

Leanne had already set up an AI Assistants channel in GA4 (we explain how in How do I track traffic from ChatGPT in Google Analytics?). Over the six months GA4 showed 410 visits from AI tools and 26 quote requests from those visits.

Here's the maths.

• 26 quote requests × 1 in 3 conversion = about 9 jobs

• 9 jobs × 5,850 in direct revenue**

On its own that doesn't cover the $9,000. If Leanne stopped here she'd say AI search lost her money. That's where most people get it wrong.

Step three: she counted the influenced value

Yes, GA4 caught 26 enquiries. But it misses a lot. People use the ChatGPT app, copy a business name and search it on Google later. Those visits land as Direct or Organic and the AI gets no credit.

So Leanne did three things to catch the rest.

She asked. She added "AI tool (ChatGPT, Gemini etc.)" as an option on her website enquiry form under "How did you hear about us?" Her office manager also started asking it on every phone call and ticking a box in their job software. Over six months 31 customers picked the AI option. Some of those were already in the GA4 number, so she removed the 9 that overlapped. That left 22 extra jobs.

She watched branded search. Google Search Console showed searches for her business name were up 38% on her baseline. Nothing else had changed in her marketing so she treated part of that as AI influence. She was careful not to claim all of it.

She watched direct traffic. Direct visits were up 22%. Again, she treated this as a supporting signal and didn't add it to the dollar total. It just gave her confidence the other numbers weren't a fluke.

Now the maths.

• 22 extra jobs from "how did you hear about us" × 14,300 in influenced revenue**

Step four: she ran the ROI formula

The basic ROI formula is:

ROI = (Value gained − Cost) ÷ Cost × 100

But there's a catch. Revenue isn't profit. Leanne's gross margin on a job is about 45% so she used profit, not revenue.

• Total revenue: 14,300 = $20,150

• Gross profit at 45%: about $9,070

• Cost: $9,000

• ROI = (9,000) ÷ $9,000 × 100 = roughly 1%

So after six months she'd basically broken even. That doesn't sound exciting until you remember two things. First, most of the work she paid for keeps working after she stops paying for it. Second, the trend was going up. Month six brought in more than months one and two put together.

Her accountant was happy. Leanne was happy. Not because the result was huge but because she could actually see it.

Leanne's story is made up but the numbers are realistic for an Aussie small business. Here's the method stripped back so you can use it yourself.

1. Record a 90-day baseline. AI referral visits in GA4, branded search impressions in Search Console, direct traffic and new customers per month. Don't skip this.

2. Track direct value. Set up an AI channel in GA4 and mark your enquiry forms, calls and purchases as key events. Multiply enquiries by your close rate and your average sale.

3. Track influenced value. Add an AI option to your "How did you hear about us?" question on forms and in your call script. Remove any overlap with the GA4 number so you don't count the same customer twice.

4. Use supporting signals. Watch the trend in branded searches and direct traffic. These help you judge whether the effect is real but don't turn them into dollars unless you have a solid reason.

5. Use profit, not revenue. Multiply revenue by your margin before you compare it to cost.

6. Judge over six months, not six weeks. AI tools update what they know slowly. You can't guarantee results online and anyone who promises AI wins in a month is selling you something.

Direct value is solid but small. Influenced value is bigger but softer. The honest answer is you report both and you're clear about which is which. If your boss or your accountant only accepts direct numbers that's fine. Show them that, then show them the influenced numbers as "likely additional value" with the method you used.

What you shouldn't do is count every new customer during the period as an AI win. That's rubbish and it'll come back to bite you when the numbers don't hold up.

For help choosing what to report each month, read What KPIs should I track for AI search?. If you want to know whether you're gaining ground on the businesses you compete with, see How do I benchmark my AI visibility against my competitors?

How long before I can measure ROI from AI search? Give it at least 90 days to set a baseline and then six months of activity before you judge. AI tools refresh what they know about businesses slowly so early numbers are often misleading.

Can I measure AI search ROI with Google Analytics alone? Only partly. GA4 shows direct referrals from tools like ChatGPT but misses visits from apps and people who search your name later. Adding an AI option to your "How did you hear about us?" question fills much of that gap.

What's a good ROI for AI search visibility? There's no fixed benchmark yet. For most small businesses breaking even in the first six months and growing after that is a reasonable result because the work keeps paying off after the spend stops.

Should I include branded search growth in my ROI? Use it as a supporting signal rather than a dollar figure. It shows people are hearing about you somewhere but other marketing can also lift it, so be careful about giving AI all the credit.

Proficlix runs AI visibility audits for Australian small businesses. We'll look at how AI tools describe you, what you're already getting from them and what's worth spending money on. Book a chat with us about your business and we'll give you a straight answer.