
AI dashboard examples built from live marketing data
Each example is a live dashboard that Claude built from one Porter prompt, while the numbers are demo data from a fictional outdoor brand. Open a demo, then its template page for the prompt to run on your own accounts. All 8 cover SEO and local search for now.
8 examples








How to build an AI dashboard from marketing data
An AI dashboard is a page of KPI cards, trends and tables that an AI assistant builds when you ask for one in plain language. To build one on marketing data, connect your accounts in Porter, then add Porter to Claude or ChatGPT and paste a prompt. After that, the assistant queries your data and lays out the dashboard.
Connect your sources in Porter, free.
After you sign up, connect GA4, Search Console, Google Ads, Meta Ads or any other source. Also, every account you connect works in Claude and ChatGPT, with no second setup.
Add Porter to Claude or ChatGPT.
In Claude, open Customize › Connectors › Add custom connector, paste
https://mcp.portermetrics.com/mcp, then sign in to Porter and allow access. In ChatGPT, open + › More › Add sources and connect the Porter Metrics app. Both have step-by-step guides: Porter for Claude and Porter for ChatGPT.Paste a prompt and open the dashboard.
Copy a prompt from the examples above, paste it into a new chat and replace [site] with your site. Then the assistant reads your data through Porter and builds the dashboard.
What you need
- Porter: the Free plan reads the last 15 days in AI chats. Paid plans, however, read full history.
- Claude: Dashboards is in beta on the Pro, Max, Team and Enterprise plans.
- ChatGPT: connectors need a paid plan: Plus, Pro, Business or Enterprise.
Watch the build · 8:12 · Build a Google Search Console dashboard with Claude · Porter Metrics on YouTube ↗
In the video
- 0:57 Connect Search Console to Claude through Porter
- 2:04 Plan the dashboard and write the prompt
- 5:35 Turn it into a live Porter dashboard
Build it in Claude or ChatGPT
Because both assistants read the same Porter connection, the choice comes down to where the dashboard should live and who needs to see it. Claude has a Dashboards feature with charts that refresh on request, while in ChatGPT the dashboard is built inside the chat.
| Question | ||
|---|---|---|
| Where the dashboard lives | Claude Dashboards (beta), saved with your other artifacts | Charts, tables and a summary inside the chat |
| Plan needed | Pro, Max, Team or Enterprise for Dashboards, although Porter connects on every Claude plan | A paid plan: Plus, Pro, Business or Enterprise |
| Refresh | Claude pulls fresh numbers from Porter when you ask it to update, and each chart shows when it was refreshed | When you ask again, ChatGPT runs a new query through Porter |
| Sharing | Each viewer needs their own Claude and their own data connection. Otherwise, publish it as a Porter dashboard for people without Claude | Not tested by Porter yet |
| How to add Porter | Customize › Connectors › Add custom connector, then paste Porter’s MCP URL | + › More › Add sources › Porter Metrics › Connect |
| Porter’s 8 example prompts | Tested | Not tested yet |
Checked October 28, 2026 against Anthropic’s guide to Claude Dashboards, Anthropic’s guide to sharing artifacts and Porter’s ChatGPT tutorial.
How to check an AI dashboard’s numbers
An AI dashboard is only useful when its numbers match the source. Before you share one, run four quick checks. These checks catch the usual mistakes before a client does.
Ask which query it ran.
Before anything else, ask the assistant which source, fields, filters and dates each chart uses. Also, in Claude Dashboards each chart shows the query that produced it (Anthropic).
Compare one KPI with the source.
Open Search Console, GA4 or the ad platform for the same dates and check one total, such as clicks or sessions. When the gap is big, it points to a filter or the wrong account.
Confirm the date range and time zone.
“Last 7 days” can end today or yesterday, and each platform reports in its own account time zone. So ask the assistant to state both.
Keep conversions apart.
Google Ads conversions, Meta Ads results and GA4 key events count different things, so keep them in separate columns rather than adding them up.
Porter’s limits: reporting is read-only, each query returns up to 1,000 rows, and some platforms report with up to a 48-hour delay. Also, Search Console days fill in over 2 to 3 days.
What an AI marketing dashboard should include
A good AI marketing dashboard answers one question for one reader, on data that keeps updating. Every example above has these six parts, so they also make a fair test for any tool.
Live connectors
Data comes from GA4, Search Console or the ad platform through their APIs, instead of a CSV export that goes stale.
KPI cards with the change
Three to five numbers, each compared with the previous period or a goal, so changes stand out.
A trend
Daily or weekly lines, so a drop shows when it started.
A top-items table
Top pages, queries or campaigns, with the change next to each one.
One breakdown
A split by device, country, channel or location, because a total alone hides where a change came from.
An “explain this change” prompt
A follow-up question the reader can ask, so the dashboard leads to the next step.
AI dashboard tools compared by type
Most AI dashboard tools fall into four types, and they differ in where the data comes from, how it refreshes and where the dashboard lives. So the right one depends more on your data than on the AI model.
| Type | Data | Refresh | Where it lives | Best for |
|---|---|---|---|---|
| File-upload generators | CSV, Excel or Sheets files you upload | When you upload a new file | The generator’s own app | One-off charts from a spreadsheet |
| BI tools with AI | Company databases and warehouses | Scheduled, set up by a data team | The BI tool | Teams with an analyst and a warehouse |
| Reporting apps with AI | Marketing platforms, through the app’s connectors | Scheduled in the app | The reporting app | Agencies happy to work in one more app |
| Connectors into your AI assistant (Porter) | Marketing platforms, through live connectors | Whenever you ask | Claude, ChatGPT or a shared Porter dashboard | Teams that already use Claude or ChatGPT |
AI report generators are close cousins: they write a fixed report for one period, often as a document or slides, while an AI dashboard stays live and can be filtered.
FAQ
ChatGPT, Claude, cost, refresh, safety and reports.
Can ChatGPT create a dashboard?
Yes, with Porter connected, ChatGPT pulls live numbers from your marketing accounts and turns them into charts, tables and a summary in the chat. However, connectors need a paid ChatGPT plan: Plus, Pro, Business or Enterprise. Porter hasn’t tested its 8 example prompts in ChatGPT yet, so check the numbers on your first one.
Can Claude make dashboards?
Yes, Claude Dashboards, in beta on the Pro, Max, Team and Enterprise plans, builds a dashboard as an artifact and refreshes it from Porter when you ask. Also, each of Porter’s Claude dashboard templates comes with a demo and a tested prompt.
Is it free?
Porter’s Free plan never expires and needs no card: unlimited connected accounts, up to 30 days of history, and 15 days in AI chats. After a 14-day trial with full access, paid plans start at $12.50 per account a month billed annually (Porter pricing). Your Claude or ChatGPT plan, however, is billed separately.
Does an AI dashboard update itself?
Not on its own schedule: in Claude, the dashboard pulls fresh numbers from Porter when you ask it to update, and each chart shows when it was refreshed. In ChatGPT, similarly, you ask again in the chat. So for viewers without an AI tool, publish it as a Porter dashboard, which keeps updating.
Is my data safe?
Porter’s reporting tools, the ones dashboards use, are read-only: the assistant can list accounts and fields and query data, but it can’t change campaigns or settings. Also, sign-in uses OAuth, the assistant only sees the rows it asks for, and you can revoke access at any time. Ad and social changes are a separate Automation feature that asks before it writes anything.
What is the difference between an AI dashboard and an AI report generator?
An AI dashboard stays live: you open it, filter it and ask it to refresh. An AI report generator, however, writes a fixed report for one period, usually a document or slides with a summary. Many teams use both: a dashboard for weekly tracking and a report at month end.
Build your first AI dashboard
Connect one source in Porter, then add Porter to your AI assistant and paste a prompt. Also, the Free plan needs no card.