An agent at work · Ranking monitor
Every week the same measurement,so you see what really moves.
The Ranking monitor is one of the agents in Zaela. Every week it measures where your website stands in Google and Bing, and how visible your brand is in ChatGPT, Claude, Gemini, Perplexity and Google's AI Overview. It compares each week with the ones before and prepares the report as a Gmail draft.
The film
Ranking monitor
The method
This is how the Ranking monitor works.
-
01Setup
A fixed set, agreed once.
At the start it proposes 10 to 20 keywords: a few brand terms, the main commercial terms and some within reach, in positions 5 to 20. Next to them come 8 to 12 AI prompts, written the way a real customer asks, across the funnel. Once the team approves, the set is fixed, with location, language and pinned AI models in one configuration. A scheduled task then runs it every Monday morning.
-
02Google and Bing
Positions 1 to 100, and the AI Overview.
For every keyword it reads the top 100 in Google and in Bing for your location and language. It notes your position and the exact URL, whether Google shows an AI Overview, whether your domain is cited in it, and which domains are. Bing matters too: ChatGPT leans on Bing's index.
-
03AI assistants
Named, in which order, cited as a source.
Every prompt goes to ChatGPT, Claude, Gemini and Perplexity, with web search on. Per answer it scores three things: are you named, in which order, and is your site cited as a source. It also notes which competitors appear, and any wrong information about your price, location or offer.
-
04Snapshot
A history that is never rewritten.
Each week is added to one file in your memory, in a fixed format with date, time, location and language. Earlier weeks are never changed or deleted: the history is where the value lies.
-
05Evolution
Week on week, and the trend.
Per keyword: up, down, new or out of the top 100. Per platform: mention rate and citation rate against last week. Five key figures sit at the top, each with its change, all calculated exactly with a calculator, never by heart.
-
06Report and draft
Ready to check, never sent by itself.
An HTML report and an A4 PDF with the five key figures, tables for Google and Bing, a matrix of prompts by platform, the trend, and what stands out with two or three actions. The PDF goes into a Gmail draft to your contact, with a short, human summary. The team checks it and sends it.
This page is based on the agent's instructions and knowledge in Zaela. The drawings are illustrations.
What it pays attention to
What it pays attention to.
- 01
Same protocol every week
Same settings, same set, same scoring. Without that, evolution is noise.
- 02
Models pinned
The model per AI platform is fixed and noted with every measurement. If a model changes, the report says so, because the trend may break there.
- 03
The trend counts
AI answers vary. One missed week is not a drop; three weeks in a row is.
- 04
Wrong is worse than missing
Being cited with wrong information is worse than not being cited. Every error about price, location or offer is flagged.
- 05
Always the context
Every report states location, language, measurement date and the models used.
- 06
Signals to the team
Big drops, wrong AI information or a competitor suddenly showing up everywhere are flagged internally.
Teamwork
Who it works with.
SEO & GEO specialist
For strategic advice when something drops, and for the actions in the report.
Report builder
For the multichannel monthly report.
AI visibility analyst
Does the full, in-depth AI measurement. The Ranking monitor follows up between two of them.
Agent builder
Updates its knowledge when a tool action or an AI model changes.
Tools it uses
- DataForSEO for Google and Bing results
- DataForSEO for ChatGPT, Claude, Gemini and Perplexity
- Search Console (read-only, at setup)
- Calculator
- Gmail (drafts only)
What it never does.
- Invent or estimate a figure. A failed call is reported as a failed measurement, never filled in with last week's value.
- Change the keyword or prompt set on its own. Proposals go to the team, and a replaced keyword keeps its history.
- Send an email to a client. It only prepares drafts.
- Present one AI measurement as hard fact. The report says the trend leads.
- Measure with a made-up set. Without an approved configuration, it starts with the setup.
What you get
What you get.
report · .html .pdfThe weekly report
Key figures with their change, Google and Bing tables, the AI matrix, the trend and what stands out, in HTML and as an A4 PDF.
Gmail draftA draft to your contact
Three to five plain sentences and the PDF attached, waiting for the team to send it.
ranking-monitor-data.mdThe history
Every week's snapshot in your memory, never rewritten.
ranking-monitor-config.mdThe configuration
Domain, keywords, prompts, location, language and pinned models, fixed after approval.
Curious where your organisation stands, week after week?
Agents at work
All agents
Report builder
Builds your recurring marketing report in a fixed format, read-only and open about every number.
See how it works Film · 1:10Content writer
Researches, writes and illustrates a complete blog post, with an SEO package and a cover that moves.
See how it works Film · 1:12AI visibility analyst
Measures how often, where and how you are named in AI answers, and which websites the AI relies on.
See how it works Film · 1:04
Stay informed
Curious what Zaela could become for you?
Leave your details and we’ll get in touch to tell you more about Zaela.
Thank you.
We’ll be in touch soon.