What is AI Governance?
AI Governance is RecordPoint’s tool designed to ensure well-governed and well-secured data is used in AI models while enforcing compliance and security controls. It allows organizations to control, filter, and protect the data being sent to large language models (LLMs) like Microsoft Copilot, GPT-based models, retrieval-augmented generation (RAG) systems, or AI agents.
Early Access
AI Governance is currently in early access. If you’re interested in joining the program, please contact product@recordpoint.com for access.
How to Access AI Governance
You can find AI Governance in the RecordPoint portal by clicking the AI Governance icon in the navigation menu.
Understanding AI Governance Structure
The core organizational unit in AI Governance is the Project.
A Project represents a subset of your data that you want to expose to an AI model while maintaining compliance and governance policies.
Creating a New Project
- Click the New Project button.
- Provide a title and description for your project.
- Save your project to begin configuring it.
Adding Data to a Project
Once a project is created, you need to add data to it. Navigate to the Data tab, where you have two options:
- All Data: Exposes all data from your RecordPoint portal while enforcing applicable attribute-based access control (ABAC) security rules.
- Saved Search: Choose a predefined saved search created in RecordPoint’s Advanced Search to expose a specific subset of your data to the AI model.
Ensuring AI Governance & Compliance
To maintain compliance and fairness in AI usage, it’s important to log actions taken when defining your dataset.
Review & Compliance Logging
- Navigate to the Review page to document actions taken to mitigate bias and integrate fairness into your dataset selection.
- You can download this review report to support compliance efforts for regulatory frameworks like the EU AI Act or NIST AI Risk Management Framework (RMF).
Exposing Data to AI Models
Once your data is governed and validated, you can expose it to AI models through four different channels:
- Microsoft Copilot – Enables Copilot Studio to access your RecordPoint-managed dataset.
- RecordPoint AI Assitant – Allows RecordPoint’s internal large language model to use the curated data.
- AI Export – Enables you to send data to be processed by internal AI models or third-party systems.
- Other Agents – Connects to retrieval-augmented generation (RAG) AI systems and custom AI agents.
Connecting to Microsoft Copilot
If you choose to expose your data to Microsoft Copilot, you will need to configure it through Copilot Studio.
- Setup requires a 1:1 session with the RecordPoint team to establish the connection.
- Once configured, your governed dataset can be securely accessed within the Microsoft Copilot ecosystem.
With Copilot you can publish to their supported ecosystem, like Microsoft Teams.
Need Help?
For assistance with AI Governance setup and configuration, contact RecordPoint Support or email product@recordpoint.com.