Skip to main content

Impact Intelligence Security and Privacy

Scott O'Reilly
  • Updated

Impact Intelligence protects your information while providing AI-powered analysis. Only the information required by each Impact Intelligence feature is sent to our AI services. Your data is not used to train AI models.

How Your Data Stays Secure

Your data remains securely stored on Spider Impact servers. Impact Intelligence sends only the information needed to complete a specific AI-powered request.

What Information May Be Sent

Depending on the feature, this information may include:

  • Dataset names and descriptions
  • Field names and field types
  • Aggregate or derived statistics
  • Selected individual values, such as minimum and maximum values
  • KPI names, values, and related context

For example, Scorecard Insights may send the KPI names, values, goals, and trends needed for its analysis. Impact Assistant may send dataset and field names, along with the summary statistics needed to answer a question. Dataset Insights may send derived values and selected individual values required to generate an insight.

This information is sent to the AI service only for the duration of the request-response cycle. It is not stored by the AI service after the response is complete, and it is never used to train AI models.

Impact Intelligence does not send entire datasets or provide AI services with unrestricted access to your database.

Example

Suppose you open the “Insights” tab on a scorecard. Spider Impact may temporarily send the relevant KPI names, values, dates, and supporting context needed for the analysis.

The AI service processes that information and returns a response. The information exists within the AI service only while that request is being processed. It is not retained after the response is complete or used for model training.

Your complete scorecard, datasets, and database remain securely stored on Spider Impact servers.

You Control Whether AI is Enabled

Organizations that do not want to use AI features can disable all AI functionality in the application administration settings. When AI functionality is disabled, Spider Impact does not send information to AI services. Connections from AI tools such as ChatGPT and Claude have their own setting, described in the next section.

AI Tools You Connect to Spider Impact

Administrators can also let people connect ChatGPT, Claude, Microsoft 365 Copilot, and other AI tools to Spider Impact. These connections work differently from Impact Intelligence. The AI tool reads Spider Impact data on a person's behalf, and the data it reads goes to the company that provides the AI tool, under your organization's own agreement with that company. The protections described above cover Spider Impact's own AI services, not AI tools you connect.

  • Connections are off until an administrator turns on “Allow MCP Connections From ChatGPT, Claude, Etc.” in the “AI” panel of Application Administration.
  • Each person approves their own connection, and the AI tool can see only what that person can see.
  • AI tools can only read unless an administrator also allows changes and the person agrees when connecting.
  • The “API Access” page lists every connection. People can disconnect their own AI tools there, and administrators can disconnect anyone's.
  • A connection that goes unused for 30 days ends on its own.

See our Connecting ChatGPT, Claude, and Copilot to Spider Impact article for more information.

Cloud Security

Government Cloud (FedRAMP) Security

For Spider Impact government customers, all infrastructure operates within Amazon GovCloud behind our secure FedRAMP-authorized boundary. The AI services used by Impact Intelligence are also within Amazon GovCloud and are part of this secure boundary.

Information sent to these services is used only to process the current request. It is not retained after the response is complete or used to train AI models.

Commercial Cloud Security

For commercially hosted customers, Spider Impact uses Amazon Web Services with enterprise-grade security. Information remains securely stored on Spider Impact servers and is sent to AI services only when required to complete an AI-powered request.

Information sent to the AI service is used only for the duration of the request-response cycle. It is not retained after the response is complete or used to train AI models.

Key Security Benefits

  • Your data is securely stored by Spider Impact — Your scorecards, datasets, and other organizational information remain securely stored on our servers.
  • Only necessary information is temporarily sent — AI services receive only the information needed to complete the current request.
  • No storage by AI services — Information is not retained after the request-response cycle is complete.
  • No AI training on your data — Your information is never used to train AI models.
  • AI can be completely disabled — Administrators can turn off all AI functionality in the application administration settings.
  • Data values come from Spider Impact — Values presented as organizational data are supplied by Spider Impact rather than generated by the AI model.