Use of Google Workspace APIs and AI/ML Model Development Policy

Compliance with Google Workspace API Usage: Spekit affirms that the Google Workspace APIs accessed through the Services are not used for developing, improving, or training generalized artificial intelligence (AI) or machine learning (ML) models.

Purpose of API Usage: Spekit’s use of Google Workspace APIs is strictly limited to enhancing the user experience and functionality of our Services as explicitly described in our service documentation and does not extend to any form of AI or ML model training.

Data Handling and Processing: All data accessed via Google Workspace APIs is processed in accordance with Spekit’s privacy policy, strictly adhering to the intended purpose as specified and not for the training of generalized AI or ML technologies.

Restrictions on Data Use: Spekit ensures that no data obtained from Google Workspace APIs will be used to train, calibrate, or improve the capabilities of AI or ML algorithms or models.

Transparency and Disclosure: Spekit is committed to maintaining full transparency about our use of data obtained from Google Workspace APIs and will continuously provide clear disclosures about the specific purposes for which this data is used.

Privacy and Security Measures: Robust privacy and security measures are implemented to protect the data accessed via Google Workspace APIs, ensuring it is handled in compliance with all applicable laws and regulations, including GDPR, CRPA, and others.

Third-Party Data Sharing and Access: There is no sharing of data obtained through Google Workspace APIs with any third parties for the purpose of AI or ML model training. 

‍AI Assistant Connectors‍

Spekit offers connectors that let an authorized Spekit user link their Spekit account to a third-party AI assistant they have chosen (for example, Anthropic’s Claude) (each, an “AI Assistant”). Once connected, the user can ask the AI Assistant questions, and the AI Assistant can retrieve Spekit content the user already has permission to access, such as Speks, topic content, and deal room information, and bring it into the user’s conversation with that AI Assistant. This section describes how data is handled when a user enables and uses an AI Assistant connector. The AI Assistant is a Third-Party Service the user selects and controls; it is not a Spekit service provider or sub-processor, and the user’s use of the AI Assistant is governed by that provider’s own terms and privacy policy.

Data collection

•    Authentication credential. When a user connects an AI Assistant, Spekit uses a standard authorization (OAuth) flow. The user authenticates as themselves and the AI Assistant receives a temporary access credential scoped to that user’s existing Spekit permissions. No separate password or persistent credential is created, and nothing new is stored about the user as a result of connecting.

•   Requests. When the user asks the AI Assistant a question, the request is transmitted through the connector to Spekit’s backend using the user’s credential, exactly as if the user had made the request in the Spekit web app or browser extension.

•   Content retrieved. Spekit returns only the content the user is permitted to access under the same role and permission rules that apply in the Spekit application.

•   Operational metadata. The connector logs limited operational metadata for debugging and security (for example, which tool was called, success or failure, and timing). The connector does not log the content of the questions asked or the answers returned.

Use and storage

•   The access credential is used solely to authenticate and authorize the user’s requests to Spekit; the same permission rules are enforced as in the Spekit application.

•   Spekit does not store the content of requests or the content returned to the AI Assistant through the connector.

•   Operational metadata is stored for debugging, security, and service-reliability purposes.

•   Spekit does not use data accessed through the connector to train, calibrate, or improve generalized AI or ML models.

Third-party sharing

•   At the user’s direction, the Spekit content responsive to the user’s request is transmitted to the AI Assistant the user has connected. Once that content reaches the AI Assistant, it is handled under the AI Assistant provider’s own terms and privacy practices, including any retention or model-training practices, which Spekit does not control or monitor.

•   Spekit does not sell this content and does not share it with any other third party for advertising or AI/ML model-training purposes. The connector service operates on Spekit’s infrastructure and any subprocessors are governed by the Data Processing Agreement between Spekit and Customer.

Data retention

•   The access credential is temporary and is revoked when the user disconnects the AI Assistant or the credential expires under the applicable OAuth lifecycle. The validity period for access credentials is limited to 24 hours, while the lifespan of refresh credentials extends to a 30-day duration.

•   Request and response content is not retained by the connector.

•   System logs are maintained for a period of 100 days, while performance traces are kept for 7 days and operational metrics are stored for a duration of 180 days.

Your choices and control‍

Enabling an AI Assistant connector is optional and user-initiated. A user (or their administrator, subject to account settings) may disconnect the AI Assistant at any time, which revokes the connector’s access. Access is always limited to the content the user is already permitted to see in Spekit.

Contact

If you have questions about this disclosure or how data is handled through an AI Assistant connector, contact Spekit at privacy@spekit.com, or Spekit Inc., 3301 Lawrence Street Suite #1, Denver, CO 80205.

‍