# How to connect Power BI to Plex Smart Manufacturing Platform > Source: https://netray.co/integrations/plex-power-bi > Publisher: Netray (netray.co), AI and integration engineering for ERP and manufacturing systems. > Section: Integrations ## Short answer Connect Power BI to Plex by pulling data through Plex APIs or scheduled extracts into a warehouse you control, then building a model with Power BI on top. Plex is cloud hosted, so there is no direct database connection or on-prem gateway to Plex itself. ## System Plex Smart Manufacturing Platform ## Capability power-bi ## Intro Many Plex customers want Power BI because finance, executives and customers already live in the Microsoft stack, and because Plex's own analytics may not cover cross-system views such as combining Plex production data with payroll, CRM or a legacy ERP at another site. The first thing to accept is the architecture. On classic on-prem ERPs you can point Power BI at a database. With Plex you generally cannot, so the real project is a data pipeline: extract from Plex on a schedule, land it somewhere durable, model it, then report. The quality of that middle layer decides whether people trust the dashboards. Once the data is modelled, a natural-language layer on top lets managers ask follow-up questions without waiting for a new report, which is where the extra value beyond ordinary dashboards appears. ## Use Cases ### OEE and downtime by workcenter **Description:** Combine production history and downtime events into availability, performance and quality views by shift, with reason code Pareto charts. ### On-time delivery and backlog **Description:** Track promised versus shipped dates by customer and part family, with an aged backlog view for customer service. ### Scrap and first-pass yield **Description:** Trend scrap quantity and cost by operation, defect code and supplier lot to point quality work at the biggest losses. ### Inventory and material shortages **Description:** Show on-hand, allocated and incoming material against upcoming demand to flag shortages ahead of the schedule. ### Multi-site and multi-system rollups **Description:** Join Plex plant data with another ERP or finance ledger for a group-level view that neither system provides alone. ### Customer scorecards **Description:** Publish a controlled report for a key customer showing their quality and delivery performance from your data. ## System Surface - Plex REST and data source APIs, callable from Power Query, Azure Data Factory, Fabric pipelines or a script, with API key authentication. Confirm endpoint details, paging and rate limits in your developer portal. - Plex scheduled exports and analytics extracts, which can be dropped to file storage and ingested without custom code if API access is limited. - Plex Connect or similar integration tooling that can push data to your storage on a schedule or by event. - Plex's own analytics and reporting capabilities, which are worth checking before you duplicate a view in Power BI. - Power BI Desktop and the Power BI service, with scheduled refresh against your warehouse. An on-premises data gateway is only needed if the warehouse itself sits on-prem. ## How It Works ### List the five dashboards that matter **Detail:** Write the questions, the audience and the refresh frequency for each. This determines which Plex data you extract, and at what grain. ### Choose the landing zone **Detail:** Use a SQL database, Azure Synapse or Fabric lakehouse, or another warehouse. Do not hit the Plex API from every report refresh. ### Build extraction jobs **Detail:** Call the Plex APIs or ingest scheduled exports incrementally, using modified dates where available, with retries and a log of rows loaded. ### Model for reporting **Detail:** Create dimensions for part, customer, workcenter, shift and calendar, and facts for production, downtime, scrap, shipments and inventory. Keep keys stable. ### Build reports and measures **Detail:** Define OEE, on-time delivery and yield measures once in the semantic model, with documented formulas the plant agrees on. ### Schedule refresh and monitor **Detail:** Set refresh windows that respect API limits, alert on failed loads, and show a last-updated timestamp on every page. ### Add natural-language Q&A **Detail:** Layer Power BI's built-in Copilot or Q&A, or a private assistant over the semantic model, so managers can ask follow-ups in plain language. ## Example Flows ### Nightly OEE refresh **Trigger:** Scheduled pipeline at 01:00 plant time. **Steps:** - Job pulls the previous day's production and downtime records from Plex. - Rows are upserted into the warehouse fact tables. - Power BI semantic model refreshes at 03:00. - Supervisors open the OEE report at shift start with yesterday's numbers. **Result:** Consistent morning metrics without anyone exporting spreadsheets. ### Ask a follow-up about a bad week **Trigger:** A plant manager sees on-time delivery drop and types a question. **Steps:** - The question goes to a Q&A layer over the semantic model. - It generates a query against the model, not the live Plex API. - It returns late orders by customer and cause with the query shown. - Manager clicks through to the underlying orders in the report. **Result:** A same-minute answer that is traceable to the governed model. ## On Prem Plex is a cloud platform, so there is no on-premises Plex database to connect to and no gateway for Plex itself. What you can keep on-prem or in a private cloud is the warehouse and the model: land the extracts in a SQL Server in your data center and let Power BI reach it through an on-premises data gateway, or keep everything in your own Azure tenant. For natural-language questions without sending data to a hosted AI service, run an open-weight model on your own GPUs against the warehouse, with generated SQL shown to the user and run through a read-only account. Be aware that Power BI service itself is a Microsoft cloud product, so fully air-gapped reporting would need Power BI Report Server or a different front end. ## Security - Store the Plex API key in a managed secret store used by the pipeline, never in a report file. - Apply row-level security in the semantic model so plants, sites and customers see only their data. - Mask or exclude pricing and personal data from the warehouse unless a dashboard needs it. - Control workspace and sharing settings, and avoid publishing to the web for plant data. - Document refresh credentials and lineage so audits can trace a number back to its Plex source. ## Options ### Plex built-in analytics **Fit:** Enough when plant-only views from a single system meet the need. Evaluate before building a pipeline. **Effort:** Configuration. ### Pipeline into a warehouse plus Power BI built in-house **Fit:** The standard route for cross-system reporting and governed metrics. **Effort:** Four to ten weeks for a first set of dashboards, depending on data quality. ### Partner build with Netray **Fit:** Useful when you also want a private natural-language layer over the model, with generated queries visible and read-only. **Effort:** Scoped over several weeks, built on your warehouse. ## Pitfalls - Refreshing reports directly against the Plex API, which hits rate limits and gives inconsistent snapshots. - Defining OEE differently on each dashboard. Put the definition in the semantic model once. - Ignoring time zones and shift calendars, which skews daily numbers. - Loading full tables nightly instead of incrementally, which gets slow as history grows. - Releasing a natural-language feature before the model has clean names and descriptions, which makes answers unreliable. ## Faq ### Can Power BI connect directly to Plex? **Answer:** Not in the way it connects to an on-prem database. Check your tenant for supported connectors, but the dependable approach is to extract via API or scheduled export into a warehouse and connect Power BI to that. ### Do we need an on-premises data gateway? **Answer:** Only if your warehouse is on-premises. The gateway connects the Power BI service to your own data sources. It does not connect to Plex, which is hosted by Rockwell Automation. ### How often can data refresh? **Answer:** That depends on API limits and your extraction design. Nightly is simplest, and many plants run hourly for production views. Near-real-time needs event-driven pushes and is rarely worth the complexity for reporting. ### Can we ask questions in plain English? **Answer:** Yes, through Power BI's own Q&A or Copilot features if licensed, or a private assistant over your semantic model. Results are only as good as the model's naming, relationships and measure definitions. ### Should we use Plex analytics instead? **Answer:** If the views you need are already there, yes, it is less to maintain. Choose Power BI when you need to combine Plex with other systems, apply your own governance, or share with people outside Plex. ## Search terms plex power bi, power bi plex integration, connect power bi to plex, plex smart manufacturing power bi, plex oee dashboard power bi, plex data warehouse power bi, plex api power bi, power bi for plex erp reporting, plex power bi natural language, plex manufacturing dashboards --- Talk to an engineer: https://netray.co/contact?intent=erp_ai | Machine catalog: https://netray.co/agent/catalog.json