We put together a digest of 8 search developments from recent weeks.
3 shifts:
β Web data: new ways for AI tools to search the web
β Search experience: Google AI Mode adds a new model and topic tracking
β Commerce: shopping inside AI tools
One has a deadline π
Our default API rate limit just went from 2,000 to 10,000 requests per minute.
5Γ more headroom for rank trackers, SERP collection at scale and parallel jobs.
Nothing to change: it applies automatically.
Custom limits stay as they are.
2 things to know before you scale up π
β The 10K is shared across APIs, not per API
β Labs, Backlinks, OnPage, Content Analysis, AI Optimization (not LLM Scraper) and Trends: 30 simultaneous requests each
Tip: one Standard task_post can carry up to 100 tasks and count as 1 request.
Details: l.dataforseo.com/4BnGRy0
Your SEO report can look fine and be wrong.
5 checks for silent failures:
β Request: same search settings?
β Task: did it really succeed?
β Results: is the data complete?
β Parser: can it read the response?
β Pipeline: did the data survive?
Check 2 can cost you money π
1. Were branded queries included?
2. Which SERP features were on those pages?
3. Which sites were sampled, and how?
Rather see your own numbers than an average? SEOInsider is free and runs on DataForSEO APIs: l.dataforseo.com/4s49Brn
π¬ Which benchmark do you still trust?
Everyone says Reddit is the way into AI answers. We checked 40 SEO brands.
β Reddit was cited in 17.1% of their Google AI Overview answers
β Reddit footprint vs AI visibility: 0.66 correlation Looks convincing.
Then we controlled for brand size π
One domain had 2.7Γ more backlinks and still ranked lower.
We compared 5 domain pairs. Backlink volume alone didn't explain who ranked higher.
So what did? π
$1.20.
That's the API cost for 1,000 AI-answer tasks a month on ChatGPT + Gemini (Standard mode).
It's our API, so we priced DIY too, 500 queries/month per platform:
β browser hosting, server, ChatGPT plan: $44β49/mo
β building the pipeline: ?
β maintaining it: ?
π
Track which sites Google's AI Overview cites for your keywords, on autopilot.
Free @n8n_io workflow:
β runs on a schedule (weekly by default)
β pulls the AI Overview via DataForSEO SERP API
β extracts each cited source: title, URL, domain
β saves it to Google Sheets
π
Prefer a BI dashboard? Swap Google Sheets for it. You'll need an n8n instance and a DataForSEO account.
Template: l.dataforseo.com/4hVBf4W
Which keywords would you track first?
A competitor outranks you. Don't jump to backlinks.
Check in this order:
1.Technical: can Google index your page?
2.Intent: does their page match the search?
3.SERP: what sits above organic?
4.Keywords: what do they rank for that you don't?
5.Links: only now
More below β
Links go last because they're the most expensive to fix.
Check the cheap things first.
For step 5, our free template finds sites that link to your competitors but not to you: l.dataforseo.com/4jNkFqz
Which of the 5 goes wrong most often for you?
What's your brand's AI visibility?
We measured one brand across 146 AI answers:
β named in the answer: 58%
β site cited as a source: 36%
β share of all cited sources: 5.7%
Same brand. Same answers. Three different numbersπ



















