Guide

How to create a monthly SEO report with ChatGPT or Claude

A monthly SEO report answers a few questions: did search traffic grow, which searches and pages moved, and did those visitors convert. This guide shows what to put in the report and the prompts that build it from your own Search Console and GA4 data.

Last updated

What a monthly SEO report includes

Use this as your template. Each section answers one question, and each comes from one data source.

SectionQuestion it answersData source
Headline numbersDid search traffic go up or down this month?Search Console: clicks, impressions, CTR, average position
Queries that movedWhich searches gained or lost the most clicks?Search Console, by query
Pages that movedWhich pages gained or lost the most clicks?Search Console, by page
Visits and resultsDid visitors from Google stay and convert?GA4: sessions, engagement rate, key events from google / organic
Index checkAre the pages that lost clicks still on Google?Search Console URL Inspection
Summary and next stepsWhat should we work on next month?Your assistant, from the rows above

Compare whole calendar months, such as October with September. Partial months and rolling 28-day windows make month-to-month changes hard to read.

Why build it in ChatGPT or Claude

The usual way is to export the Search Console Performance report and a GA4 report each month, paste them into a spreadsheet and write the summary by hand. An AI assistant can fetch the numbers, compare the months and draft the summary, if it can read your data.

QuerySail runs two hosted MCP servers, one for Google Search Console and one for Google Analytics 4. MCP, the Model Context Protocol, lets an AI app call outside tools when a question needs them. With both servers connected, one chat can build the whole report.

Connect both servers

  1. Create a QuerySail account. The trial needs no card.
  2. Open Account → Connections. Select Connect Search Console, then Connect Google Analytics, and approve read-only access on Google's consent screen each time.
  3. Add both endpoints below to your AI app as two separate connectors, and sign in with your QuerySail account when each one asks.
https://mcp.querysail.com/gsc/mcp
https://mcp.querysail.com/ga4/mcp

The steps for each AI app are in how to connect Search Console to ChatGPT and Claude and how to connect GA4 to ChatGPT and Claude. Do them once for each endpoint.

Seven prompts that build the report

The prompts use the sample site Sample Outdoors and report on October 2026. Run them in order in one chat, so each step can use the results of the steps before it. Replace the property names and dates with your own.

1. Check that both servers answer

Which Search Console properties and which Google Analytics properties can you access?

Tool: list_properties (both servers)

Each server has its own list_properties tool. Note the two IDs for the same site, for example sc-domain:sample-outdoors.example in Search Console and properties/100000001 (Sample Outdoors (web)) in Google Analytics.

2. Get the headline numbers

For sc-domain:sample-outdoors.example, compare clicks, impressions, CTR and average position for October 1 to October 31, 2026 with September 1 to September 30, 2026.

Tool: compare_performance

Give both months as explicit dates. Without them, the comparison period is the same number of days immediately before, which does not line up with calendar months. The answer gives both totals with absolute and percent changes, and splits the change in clicks into a part from impressions and a part from CTR.

3. Find the queries that moved

For the same property and months, which queries gained or lost the most clicks? Show the top 10 each way.

Tool: compare_performance

The same tool with a query breakdown returns the top gainers and losers. Ask to rank by impressions instead when you want to see searches where your pages started or stopped appearing.

4. Find the pages that moved

Now do the same by page: which pages gained or lost the most clicks between September and October?

Tool: compare_performance

A page that lost clicks while its queries held steady may have dropped in position. A page that lost impressions may have dropped out of results for some searches. Both are worth a closer look in step 6.

5. Add visits and key events from GA4

For Sample Outdoors (web), compare sessions, engagement rate and key events for October 1 to October 31, 2026 with September 1 to September 30, 2026. Only include sessions where source / medium is google / organic. Show the landing pages that gained or lost the most sessions.

Tool: compare_report

Search Console stops at the click. GA4 shows what visitors from Google did next. The sessionSourceMedium filter keeps only Google's unpaid results, which is the traffic Search Console counts. Use custom comparison dates so both tools cover the same calendar months.

6. Check the pages that lost clicks

Inspect the three pages that lost the most clicks. Are they indexed, when did Google last crawl them, and does Google use the canonical they declare?

Tool: inspect_urls

This rules out the simplest causes of a drop: a page that left the index, a noindex tag, a robots.txt block or a different canonical. If all three pages are indexed and crawled, the drop is more likely about rankings or demand.

7. Write the report

Write the October 2026 SEO report for Sample Outdoors from the results above. Use these sections: headline numbers, queries that moved, pages that moved, visits and key events, index check, and three next steps. Keep it under 400 words. Report the changes as observed and label any possible cause as a guess.

No tool call. Your assistant works with the rows it already has.

Your assistant works with the rows it already has, so no new data is fetched. Ask for a table, bullet points or an email draft, depending on who reads the report.

Example report

Made-up example. We invented these numbers to show the shape of a report. They do not come from a real site or a QuerySail customer. Search Console dates are in Pacific Time and GA4 dates are in the property's time zone.

Headline numbers for Sample Outdoors, October 2026 compared with September 2026:

MetricSeptemberOctoberChange
Clicks (Search Console)8,9109,840+10.4%
Impressions (Search Console)296,500312,000+5.2%
CTR (Search Console)0.0300.032+0.002
Average position (Search Console)15.114.20.9 better
Sessions, google / organic (GA4)7,9908,720+9.1%
Key events, google / organic (GA4)188214+13.8%

The summary your assistant writes from those rows could read like this:

Search clicks rose 10.4% to 9,840. About half of the gain came from more impressions and half from a higher CTR. "winter hiking gear" gained the most, up 310 clicks, and /blog/winter-hiking-gear was the page that gained the most. "best hiking boots" lost 140 clicks, and /blog/best-hiking-boots lost 150. That page is indexed and Google crawled it on October 27, so it did not leave the index. Its average position fell from 3.8 to 5.2. Organic sessions in GA4 rose 9.1% and key events rose 13.8%. Next steps: refresh /blog/best-hiking-boots, link to it from the winter gear page, and check next month whether its position recovers.

The data shows what changed, not why. A seasonal rise in winter searches is a likely guess for the winter gear gain, but the report should say it is a guess.

Limits to know

Other ways to make a monthly SEO report

A Looker Studio dashboard with the Search Console and GA4 connectors updates itself and suits a fixed set of charts that someone checks every month. It does not write the summary or explain which pages to look at. A spreadsheet of exports works with any tool, but you repeat the exports each month. An assistant connected to your data suits you if you want the summary written and want to ask follow-up questions, such as which queries a falling page lost.

Try it on your own site

QuerySail costs $8 a month for both servers after a 7-day free trial with no card. To dig into a page that lost clicks, read how to find content decay. To find queries that are close to the first page of results, read how to find striking-distance keywords.

Questions, answered

Something else? See support.

What should a monthly SEO report include?

Clicks, impressions, CTR and average position from Search Console compared with the month before, the queries and pages that gained or lost the most clicks, sessions and key events from Google organic traffic in GA4, the index status of pages that dropped, and a short list of next steps.

Can ChatGPT or Claude make an SEO report from my Search Console data?

Yes, when the assistant can read the data. Connect Search Console and Google Analytics through an MCP server such as QuerySail, and the assistant fetches each number it needs. You can also export CSV files and upload them, but then you export new files every month.

Should I compare with the previous month or the same month last year?

Use the previous month to see recent changes, and the same month last year when your traffic is seasonal. Both QuerySail compare tools can compare with the same weekdays 52 weeks earlier. Ask for both if you are not sure.

When in the month should I run the report?

Wait until about the fourth day of the new month. QuerySail uses finalized Search Console data by default, which ends about three days before today. GA4 can take 24 to 48 hours to process events.

Why are Search Console clicks and GA4 sessions different in the report?

They count different things. Search Console counts clicks on Google results. GA4 counts sessions from browsers that ran the Google tag, so visitors who decline cookies or block scripts are missing. Compare each tool with its own previous month rather than with each other.

Can QuerySail send the report every month on its own?

No. QuerySail answers when your assistant asks. Save the prompts from this guide, change the dates each month and run them again.

Does this cost extra?

No. Search Console and Google Analytics are on the same $8 per month plan. The 7-day free trial needs no card.