# Tableau MCP server

Source: https://agentic-cmms.klementine.xyz/integrations/tableau

> Connect Tableau to Claude, ChatGPT and other AI agents to read maintenance dashboards, published data sources and KPIs such as PM compliance and MTTR.

Tableau is the analytics platform many large operations teams use to publish dashboards on maintenance performance. Workbooks show PM compliance by site, MTTR by asset class, the age of the backlog and the split between planned and reactive work. Published data sources feed them from the CMMS, the ERP and the historian.

Dashboards answer the questions someone thought of in advance. The ones that come up in a Tuesday meeting, such as "which three assets drove the rise in reactive work at Plant B?", still end in a filter-click-export routine. With Agentic CMMS, an AI agent reads the same published data and answers that follow-up directly, so the meeting can move on.

## What your AI agent can do with Tableau

- **Pull a metric.** "What was PM compliance at each site last month?" The agent reads the view or data source behind your compliance dashboard and lists the values.
- **Follow up on a chart.** "Which assets drove the rise in reactive work at Plant B?" It filters the published data source by site and work type and ranks the assets.
- **Compare periods.** "How does MTTR for conveyors this quarter compare with the same quarter last year?"
- **Search your content.** "Is there a dashboard for backlog by craft?" The agent searches projects and workbooks and returns the view name and owner.
- **Check freshness.** "When did the maintenance extract last refresh?" You know whether to trust the numbers before the meeting.
- **Summarise a dashboard.** "Give me the three biggest changes on the reliability dashboard since last week." The agent reads the view data and writes a short summary.
- **Export for a colleague.** "Send me the rows behind the overdue PM view for the Leeds site." It returns the data as a list you can share.

## Tableau with your other systems

- When a Tableau view shows overdue PMs rising, ask the agent to list the actual overdue jobs in [IBM Maximo](/integrations/ibm-maximo).
- Compare first-time fix rates in Tableau with the visit notes in [Salesforce Field Service](/integrations/salesforce-field-service) to find the cause of repeat visits.
- Run a deeper query on the raw work order history in [Google BigQuery](/integrations/google-bigquery) when a dashboard's level of detail is not enough.
- Post a Monday summary of the reliability dashboard to a [Slack](/integrations/slack) channel for shift leads.

## From dashboard to decision

Tableau is very good at showing that something changed. People still have to work out why and what to do about it. An agent that reads Tableau and your operational systems together shortens that path: from a red tile on a dashboard to the five work orders behind it, in one conversation.

Each person approves their own agent, owners and admins choose which roles may use which tools, and every query the agent runs is written to the tamper-evident activity log.

## Questions and answers

### Can ChatGPT read the numbers behind a Tableau dashboard?

Yes. The agent can query the views and published data sources you have access to and return the underlying figures, not just a picture of the chart.

### Can Claude find which Tableau workbook has our PM compliance report?

Yes. It searches the projects and workbooks you can see and tells you which view holds the metric you asked about, with its owner.

### Does an AI agent respect Tableau permissions?

The agent works through the Tableau connection your admins set up, so it only sees the content that connection allows. On top of that, Agentic CMMS lets owners limit each tool to certain roles.