How to Use the Power BI Connector

The LLM Pulse Power BI connector lets you load your AI visibility data into Power BI Desktop. Use it to build dashboards for mentions, citations, visibility, sentiment, and competitor benchmarks.

With the Power BI connector, you can:

• Load your AI visibility metrics into Power BI Desktop
• Build and refresh dashboards for mentions, citations, visibility, sentiment, and competitors
• Write custom queries against your LLM Pulse data
• Schedule automatic refresh so reports stay current

This connector is currently a private beta custom connector. That means you install a connector file in Power BI Desktop rather than selecting LLM Pulse from Microsoft's built-in connector list.

Before you start

You need:

• A Scale plan or above. The connector authenticates with an API key, which is available on Scale and Enterprise plans.
• An LLM Pulse API key. Open the user menu in the bottom left of the app, go to API Keys, and create one if you have not already. Keys start with llmpulse_.
• Power BI Desktop on Windows. Custom connector setup and report authoring happen in Power BI Desktop.
• The LLMPulse.mez connector file. Download it from the Integrations page in the app (Tools > Integrations > Power BI) or directly from the connector download link.

Install the connector

1. Close Power BI Desktop.
2. Copy LLMPulse.mez to:

   Documents\Power BI Desktop\Custom Connectors

1. Open Power BI Desktop.
2. Go to File > Options and settings > Options > Security.
3. Under Data Extensions, allow custom extensions for beta testing.
4. Restart Power BI Desktop.

Connect your data

1. In Power BI Desktop, open Get data.
2. Search for LLM Pulse.
3. Select the connector and click Connect.
4. Paste your LLM Pulse API key.
5. Leave the optional fields blank to browse your projects.
6. Pick a project.
7. Choose Timeseries - last 30 days or one of the helper tables.

Available tables

The current beta includes:

• Timeseries - last 30 days: Daily metrics by date and actor (your project or a competitor)
• Metrics: Summary metrics, prompt summary, Share of Voice, top sources, AI crawler traffic, and AI referral traffic
• Dimensions: Competitors, tags/collections, prompts, prompt executions, sources, mentions, citations, sentiment categories, models, locales, and AI bot catalog
• Search Console: Summary, time series, top queries, and top pages when your project has a connected property
• Citation intelligence: Citation groups by URL, domain, or host
• AI Model Insights: Model summary, position distribution, and AI Overview result availability
• Content and Workflow: AI responses, sentiment records, recommendations, and GEO Writer tasks
• Webhooks: Active webhook subscriptions
• API Endpoint Catalog: A reference table for every REST API endpoint available to the connector

Available metrics

The time series table includes:

Field	What it measures
mentions	AI responses that mention the actor
citations	AI responses that cite the actor
responses	Total prompt executions
visibility	Share of queries where the actor appears
aivisibilityscore	Position-weighted visibility score
avg_position	Average citation position
avgmentionposition	Average mention position
citation_rate	Share of responses that cite the actor
sentimentverypositive	Very positive sentiment share
sentiment_positive	Positive sentiment share
sentiment_neutral	Neutral sentiment share
sentiment_negative	Negative sentiment share
sentimentverynegative	Very negative sentiment share
net_sentiment	Overall sentiment balance

Custom date ranges and filters

For a custom range or filter, create a blank query and call the connector directly:

let
    Source = LLMPulse.Contents(
        123,
        90,
        "week",
        "mentions,visibility,citation_rate,ai_visibility_score"
    )
in
    Source

Replace 123 with your project ID. The second value is the number of days, and the third value is the granularity: "day", "week", or "month".

For endpoints that are not exposed as a default navigator table, use LLMPulse.Table(...):

let
    Source = LLMPulse.Table(
        "/metrics/top_sources",
        [project_id = 123, #"range" = 90, per_page = 100],
        "data"
    )
in
    Source

Advanced users can call GET and POST endpoints with LLMPulse.Api(...):

let
    Source = LLMPulse.Api(
        "/prompts",
        "POST",
        [],
        [project_id = 123, prompts = {"best CRM software"}]
    )
in
    Source

Good to know

• Percentages use a 0 to 1 scale in Power BI. A visibility value of 45% appears as 0.45, so Power BI percentage formatting works correctly.
• Paginated navigator tables load the first 100 rows by default. Use LLMPulse.Table(...) with page and per_page for deeper paging.
• DELETE operations are not run from the Power BI connector. Use the REST API directly for destructive operations such as deleting a webhook subscription.
• The connector is private beta. You may need to approve custom connector loading in Power BI Desktop.
• Scheduled refresh needs a gateway. Because this is not a certified built-in connector yet, Power BI Service refresh requires the on-premises data gateway with the connector installed.

Scheduled refresh

To refresh in Power BI Service:

1. Install the on-premises data gateway on a Windows machine that stays online.
2. Copy LLMPulse.mez to a folder the gateway service account can read.
3. In the gateway app, enable custom connectors and choose that folder.
4. In Power BI Service, enable custom connectors for the gateway cluster.
5. Create a data source using the LLM Pulse connector and your API key.
6. Configure scheduled refresh on the semantic model.

Troubleshooting

The connector does not appear
Confirm the file is in Documents\Power BI Desktop\Custom Connectors, custom extensions are enabled, and Power BI Desktop has been restarted.

Authentication fails
Check that your API key starts with llmpulse_, has not been revoked, and belongs to an account on Scale or above.

The chart is empty
Try a longer range, switch from daily to weekly granularity, and remove restrictive filters.

Scheduled refresh fails
Confirm the gateway machine is online, the connector file is installed in the gateway custom connector folder, and the API key is set in Power BI Service credentials.