Martin Williams

[Martin Williams @martincoderra]

@VeenaGandhi9 Love this. Two leads in a week without trying means the model already gets what you stand for. Worth asking every new lead how they found you, and checking the same prompt in Gemini and Perplexity. The engines don't always agree.

[Martin Williams @martincoderra]

@connections8 @profound Point 2 is the one I'd underline for agencies here. Clients say AI sends them nothing, then you look at Direct and brand search in GA4 and it's there. Point 4 is why we report engine by engine. Great to see this conversation happening in Sydney.

[Martin Williams @martincoderra]

@glenngabe This is the conversation we keep having with agencies. One prompt checked once is a coin toss. Track a wide prompt set, engine by engine, and report the trend. Clients take "up 6% across 200 prompts" far better than "we vanished on Tuesday".

[Martin Williams @martincoderra]

@DougManoel89 We provide per engine

[Martin Williams @martincoderra]

Three studies this week say AI search sends almost no traffic. Fair enough. But your clients aren't asking about clicks. They're asking why ChatGPT recommends someone else. Agencies that win here will sell visibility, measure it honestly and stop promising traffic.

[Martin Williams @martincoderra]

@JShehata When a publisher turns AI visibility into a product, you know brands are asking for it. Tracking is the easy part. The value is telling a client what to fix, in which engine, and showing it moved next month.

[Martin Williams @martincoderra]

@psinke This is the opening for smaller brands and the agencies behind them. Nobody wins "best X" against the giants, but specific use-case questions are wide open. Map those prompts first, then build pages that answer them properly.

[Martin Williams @martincoderra]

@aleyda @semrush This is the bit Aussie brands miss. Same language, different sources, different winners. Benchmarking an Australian client against US AI visibility data tells you very little. Report it market by market, engine by engine.

[Martin Williams @martincoderra]

Clients keep telling agencies ChatGPT sends them nothing. Check Direct in GA4. Plenty of AI visits arrive with no referrer, so they land there. Add "how did you find us?" to the enquiry form and AI shows up fast. The channel is bigger than the dashboard says.

[Martin Williams @martincoderra]

@Ldnbox Agree on point 5. Once tracking is commoditised, the agency edge isn't the dashboard, it's turning per-engine gaps into a plan the client will pay for. Agencies that start offering it now get two years of proof before it's table stakes.

[Martin Williams @martincoderra]

@Arfadia That 17% overlap is the number agencies should lead with. A client can be the go-to answer in ChatGPT and invisible in AI Mode, and a blended score hides both. Report it per engine and the client conversation gets a lot more honest.

[Martin Williams @martincoderra]

@pressgazette Interesting that Australia is in the test set. When publishers start selling AI visibility, agencies need a baseline before that pitch reaches their clients: where does the brand show up today, engine by engine? Otherwise nobody can judge what a test actually moved.

[Martin Williams @martincoderra]

@DougManoel89 Exactly that. The cited URLs are the bit clients act on, because they show what to fix, not just a score. Keep the question set the same each month too, so any movement is real and not noise.

[Martin Williams @martincoderra]

@TheCoolestCool @ahrefs Agree. The SOP work goes to agents, but someone still has to decide which buyer questions matter for a client and explain why they're missing from the answers. That judgement is what agencies will get paid for.

[Martin Williams @martincoderra]

@aleyda @SERanking Brand searches are where clients feel this first. If an AI Overview sits on top of their own name, it's worth checking what it actually says about them before a prospect does. Plenty haven't looked.

[Martin Williams @martincoderra]

Your client ranks first on Google. Ask ChatGPT the same buyer question and they might not get a mention.

The engines disagree with each other far more than most agencies assume.

Report AI visibility engine by engine. A blended score hides where the client is missing.

[Martin Williams @martincoderra]

@randfish A waste of time if you sell it as a traffic channel. Not if the client's real question is why ChatGPT recommends someone else. Being named in the answer is a brand problem, and it shows up in branded search and sales calls, not GA4 referrals.

[Martin Williams @martincoderra]

@lilyraynyc Love this. The About page is the one page where a brand explains plainly who it is, what it does and why you should trust it. That's exactly what an AI answer needs. It's also the page most agencies never put in the brief.

[Martin Williams @martincoderra]

@gaetano_nyc Agree. The AI answer often builds the shortlist and Google just catches the branded search afterwards. That's why last-click reporting undersells AI visibility. Worth tracking branded search lift alongside citations.

[Martin Williams @martincoderra]

@DynamicBusiness The demand here is real. The next question Aussie businesses should ask is how it gets measured: which prompts, which assistants, and what the baseline looked like before any work started. Without that, nobody can prove it moved.