AI & Integrations Guide#

Decide what Reportworq connects to and on what terms: source report providers, data sources and connectors, authentication providers, AI providers, Microsoft 365 Copilot and MCP with a content-security strategy, the REST API, and the Script Runner.

Integration strategy#

• Plan your integrations - The decisions an integration owner makes before wiring anything up, where reports come from, which data sources to connect, how people sign in, whether to expose AI and MCP, and how to keep content safe once you do.

Source report providers#

• Source report providers - The configured, trusted locations Reportworq is allowed to read report source files from, what each provider type can and cannot do, and the parameter-driven dynamic paths that let one job read a different workbook per run.

Data sources & connectors#

• The integrations hub - Settings, Integrations is the single place to add, test, enable, and edit every external connection Reportworq uses, grouped into six families covering authentication, the MCP platform tool, AI providers, datasources, distributors, and report providers.
• Connect a data source - Add a datasource connection in the integrations hub, fill its connector-specific settings, and validate it with Test connection before anyone models or reports against it.
• Connector catalog - Every datasource connector Reportworq ships, grouped by category, with the authentication approach and the per-connector traps that most often block a connection.
• Connect IBM Planning Analytics - Connect Reportworq to IBM Planning Analytics (TM1), choose the right authentication type for self-hosted, Cloud (IBM SoftLayer), or Cloud (MCSP), and handle the deployment traps that support is most often asked about.
• Connect SQL, OLE DB, and ODBC - Connect Reportworq to relational databases with the SQL Server, OLE DB, ODBC, and SQLite connectors, and to API and SaaS systems through CData drivers, keeping credentials out of the connection string and avoiding the common ODBC DSN and timeout traps.
• Connect Power BI and Analysis Services - Configure the native ADOMD connection to a Power BI semantic model, Azure Analysis Services, or SQL Server Analysis Services, choose an authentication mode, and query it with DAX, DMV, or MDX.
• Data models - What a data model is, why it sits between a source system and report authors, and how to build one, curate its fields and filters, and publish a view for authors to use in the Excel add-in.

Identity & authentication#

• Authentication providers - Which identity providers Reportworq supports, what each one needs from you, and the cutover facts that decide whether a switch goes smoothly or locks everyone out.
• Microsoft 365 OAuth setup - Register one Azure (Entra ID) application and authenticate Reportworq to it, so Entra sign-in, the SharePoint provider, the Email, SharePoint, and Teams distributors, and email data collection all work from a single credential.

AI in Reportworq#

• AI in Reportworq - How Reportworq applies AI to finance-approved, governed content instead of raw databases, so answers are trusted, auditable, and scoped to what each user is allowed to see.
• Connect an AI provider - Add and enable an AI provider connection (OpenAI, Azure OpenAI, Amazon Bedrock, or Reportworq AI) so every downstream AI feature has a model to call.
• The AI assistant - The in-app AI assistant is present in the product but not yet switched on in this build. This page explains its current state and points to the AI features you can use today.
• AI-generated insights in reports - Attach model-written narrative to a distributed report, either as a distribution-time AI Insight variable or inline in the workbook with the RWAIINSIGHT function.
• The AI prompt store - The global store of the internal prompts Reportworq sends to your AI provider, editable in Settings ▸ AI Prompts, with variable tokens, factory defaults, and a one-click reset.

Copilot & MCP#

• Copilot, MCP, and trust paths - The three ways an AI client or external system reaches Reportworq, the Microsoft 365 Copilot agent, the MCP shim, and the automation API, and the two trust paths that broker their traffic.
• Secure content before you expose it to AI - The content-security decisions to make before enabling the MCP server, which path carries a user identity and which does not, what a reader can reach, and how to verify it rather than assume it.
• Microsoft 365 Copilot agent - Generate the Reportworq Copilot agent package, deploy it into Microsoft 365, and let Copilot users query their reports in natural language, scoped to their own permissions.
• The MCP shim for desktop AI clients - Download the Reportworq MCP connector, paste a per-client snippet into a desktop AI client such as Claude Desktop, and confirm the connection reaches your reports.
• MCP audit - The MCP Server Log records every MCP tool call served to an AI client and every request rejected at the gate, with a full request-and-result drill-in, under Settings ▸ Auditing.

Automation & extensibility#

• The automation API - Reportworq's HTTP API for running and monitoring jobs from outside the product, delivered two ways, the Local REST API on the LAN and the Cloud API over CloudHub, with opposite key models.
• Run a job by API - Call the /api/v1 commands to list, run, monitor, and cancel a job, and filter a single run to one output type and one destination so different callers get different deliveries from the same job.
• Turbo Integrator, Workato, and Task Scheduler - Launch Reportworq jobs from an IBM Planning Analytics Turbo Integrator process, from a Workato recipe, or from Windows Task Scheduler, with the PowerShell hang fix that keeps a script from stalling.
• Deploy the Office add-in - Get the Reportworq Office add-in onto users' Excel and PowerPoint, by Microsoft 365 Centralized Deployment or by a shared-folder Trusted Add-in Catalog sideload, with SSL as a hard prerequisite.
• The Script Runner - Attach C# script files to a distribution job so they run at six points in the execution pipeline, what the six hooks can reach, how to configure and validate a script, and the full-trust security posture to understand before enabling it.
• Script API reference - Everything a Reportworq .csx script can call, the globals available at each hook, the shared job state dictionary, logging and halt semantics, the AfterSend job result, and the compilation environment.
• Example scripts - A library of 29 working .csx examples covering every Script Runner hook, downloadable individually or as a single zip, each one compile-checked against the live script API on every build.