Before you start
You need:- An AI client that supports remote MCP servers and OAuth
- A HireOtto account with GA4 access and available usage
- A Google login that can access the GA4 property you want to analyze
Connect the GA4 server
The labels differ by AI client, but the sequence is the same:- Open your AI client’s connectors, integrations, or MCP settings.
- Add a custom remote MCP server.
- Enter
https://ga4.hireotto.com/mcp. - Complete the HireOtto sign-in when prompted.
- Return to the conversation after the server connects.
Authorize Google Analytics
Ask your AI client:Connect my Google Analytics account. Use the default profile.
- Choose the Google login that has access to the required GA4 property.
- Review and grant the requested read-only Analytics permission.
- Accept HireOtto’s Terms and Privacy Policy to complete the connection.
- Return to your AI client.
default. It is the right choice for most users and is available on every plan that includes GA4 beta access.
Verify accounts and properties
Before requesting a report, list the accessible accounts and properties:List every Google Analytics account and GA4 property available to the default profile. Include the account name, account ID, property name, and property ID. Do not run a report yet.
- Account name
- Account ID
- Property display name
- Property ID
Inspect the property before reporting
A configuration read is a useful first check:For GA4 property PROPERTY_ID, summarize the property details, data streams, key events, Google Ads links, custom dimensions, and custom metrics. Read only; do not change anything.
PROPERTY_ID with the numeric ID returned by property discovery. You may also supply the full properties/PROPERTY_ID resource name.
Configuration reads can reveal whether you selected the correct web or app stream and whether the measurement objects needed for analysis exist. They do not validate that a website tag fires correctly, that events contain the intended parameters, or that imported conversions are suitable for bidding.
Run a first standard report
Start with a constrained question and an explicit date range:For GA4 property PROPERTY_ID, report sessions and active users by session default channel group for the last 30 complete days. Compare with the previous 30 days, show the raw values and percentage change, and flag incompatible fields before running the report.
- The property ID
- Exact start and end dates, or an unambiguous relative window
- Dimensions
- Metrics
- Filters
- Comparison period
- Desired output
Check recent activity
Use a realtime report when you need recent activity rather than a historical date range:For GA4 property PROPERTY_ID, show active users by country and device category from the realtime reporting window. Read only.
Profile behavior
The optionalprofile_id parameter selects which saved Google connection to use.
profile_id or use default. Agency and Enterprise users can authorize named profiles to keep client or business logins separate.
To connect a named profile:
Connect Google Analytics using the profile ID client_acme.
Using the client_acme GA4 profile, list accessible accounts and properties.
Read and write scope
Reporting defaults and limits
These defaults apply when the AI client does not supply an override.Standard reports
Realtime reports
Metadata and compatibility
- Metadata search returns up to 500 dimensions and 500 metrics by default; the limit is clamped to 1–2,000 for each group.
- Deprecated API names are excluded by default.
- Compatibility checks return compatible fields by default.
Plans, beta status, and usage
GA4 reporting is currently in beta.Common failure cases
The server connects, but GA4 is not authorized
The AI client-to-HireOtto connection succeeded, but the Google connection did not. Ask to connect Google Analytics again, open the authorization link, finish consent, and then retry property discovery.No accounts or properties appear
The authorized Google login has no visible GA4 properties, or it is the wrong login. Reconnect with the correct Google account or ask an Analytics administrator to grant that login access. Property discovery only returns resources accessible to the caller.The expected property is missing
Verify that you authorized the intended Google identity and that its access has not been removed. Search by both property display name and property ID.A named profile is rejected
Named profiles require Agency or Enterprise access. Use thedefault profile or change to a plan that supports multiple profiles.
Authorization worked before but now fails
The Google token may have expired, been revoked, or lost the required scope. Reconnect the affected profile and repeat property discovery.A report says dimensions and metrics are incompatible
Remove or replace the incompatible fields. Ask HireOtto to search property metadata and check compatibility before rerunning the report.A report returns no rows
Confirm the property, date range, filters, and field names. No rows can also be a valid result when the property has no matching data. Do not broaden the request until you have verified the property and filters.The output is incomplete
Reduce the dimensions, narrow the date range, or request a CSV within the supported export limit. If the result reached a row cap, treat it as partial rather than as a complete account view.A report is blocked by quota or usage
Retry later for a temporary Google Analytics API quota response. For a HireOtto plan or credit block, review the account’s current entitlement and usage.Practical prompts
Confirm the connection
Using the default GA4 profile, list accessible accounts and properties with names and IDs. Do not run a report.
Review measurement configuration
For property PROPERTY_ID, summarize data streams, key events, Google Ads links, custom dimensions, and custom metrics. Flag missing information but make no changes.
Review acquisition
For property PROPERTY_ID, compare sessions, active users, key events, and total revenue by session default channel group for the last 30 complete days versus the previous 30 days. Check compatibility first and preserve raw values.
Review landing pages
For property PROPERTY_ID, show landing pages by sessions, engaged sessions, engagement rate, key events, and total revenue for the last 28 complete days. Return the top 100 rows and identify pages that need investigation, not automatic changes.
Use property metadata before a custom report
Search the metadata for property PROPERTY_ID for dimensions and metrics related to campaigns, landing pages, revenue, and key events. Show API names and check whether my selected fields are compatible before running anything.
Keep profiles separate
Using the client_acme profile, list accessible properties. Confirm the account and property IDs before running a report.
Safe operating pattern
For consequential analysis, use this sequence:- Discover the live property.
- Confirm the exact property ID.
- Inspect relevant configuration.
- Check reporting metadata and compatibility.
- Run a small report with explicit dates.
- Review raw rows and any truncation notice.
- Expand or export only after the small report is correct.