Lesson 2 · 17 min
In 2026, the first official, free dashboards on AI appeared: the Search Console generative AI report and Bing's AI Performance, now with the Topics Portfolio. Each one measures something different, and none measures everything. Knowing what each number is keeps you from drawing the wrong conclusion.
Start the lesson
Goal: open the Search Console AI report and Bing's AI Performance, and note what each one measures and what it doesn't.
Click each block.
Google counts impressions, Bing counts citations, your spreadsheet joins the two with a date.
Click each step. You need the site verified in both dashboards (course 1, lesson 3).
Open the site's property and find the generative AI report. Note the total impressions for the last 28 days and the pages with the most impressions.
If little or nothing shows up, note "no data yet": that's a baseline too.
In Bing Webmaster Tools, open AI Performance. Write down the total citations, the 5 most-cited pages, and 5 grounding queries.
Export the CSV: there is no API for this data (a third-party report says it is "on the backlog, no timeline").
If the beta already shows up for your site, write down the 3 topics with the most demand and the least competition. They become your content plan, along with course 2.
An AI impression is not a click; Bing citations are a sample; neither one measures ChatGPT, Claude, or Gemini directly. For those, lesson 3 teaches the question battery.
Adding Google impressions and Bing citations into a single number. They are different measures: keep two columns.
The first block is the spreadsheet. The second is the bing-ia template, which cross-checks the Bing CSV with the Search Console history (detailed in course 7).
mes;gsc_impressoes_ia_28d;bing_citacoes;bing_paginas_citadas;top5_paginas_bing;top5_grounding_queries;observacao 2026-10;<n ou sem dado>;<n>;<n>;<url1|url2|...>;<consulta1|consulta2|...>;<o que mudou no site este mês>
git clone https://github.com/inematds/inemaseo-sistemas && cd inemaseo-sistemas/ferramentas/bing-ia # 1. import the CSVs exported from AI Performance (pages and grounding queries) python3 bing_ia.py importar ~/Downloads/ai-performance-*.csv --db dados/bing.sqlite # 2. cross-check with the Search Console history (gsc-historico) python3 bing_ia.py relatorio --db dados/bing.sqlite --gsc-db ../gsc-historico/dados/gsc.sqlite --site seusite.com --saida saida/seusite
bing-ia splits your pages into four groups:
The example in the repository uses fictitious CSVs (made-up numbers, real URLs) to show the tool. Microsoft does not document the export columns: the names the tool recognizes are a guess to confirm with a real CSV.
It takes about 30 minutes if your site is already verified.
Fill in the first row of the spreadsheet with the AI impressions from Search Console and the Bing citations. No data yet? Write "sem dado" and the date: in three months you will have something to compare against.
Done when: linha-de-base-ia.csv has the row for the month with both columns filled in (a number or "sem dado") and the date.
In AI Performance, write down 5 grounding queries and the page cited for each one. Mark which of those phrases also show up in your site's Search Console.
Done when: a list with 5 grounding queries, the cited page, and "shows up on Google: yes/no" (or a note that Bing doesn't show citations yet).
The explanation appears after your answer.
Question 1 of 4
What does the generative AI report in Search Console show?skill
Question 2 of 4
Which is the only official dashboard that shows AI citations?
Question 3 of 4
What is a grounding query?
Question 4 of 4
How do you combine Google and Bing in the monthly report?
Choose an answer. The explanation appears after you try.