The use of artificial intelligence (AI) in Staffbase products is based on the core principles of privacy, security, and compliance. Our commitment to delivering secure and trusted communication solutions is deeply embedded in every aspect of our technology and reflected in our contractual safeguards.
As AI becomes an integral part of workplace productivity, Staffbase recognizes that regulatory clarity and transparency are more important than ever. AI-powered Staffbase features help you work more efficiently and creatively, while maintaining the trust of your employees and enterprise-grade safeguards.
Is the use of AI in Staffbase products secure?
Yes, all our AI-based features are secure. All data stays within the AI models and is not used for training models. Staffbase uses a self-run model and OpenAI Services, both of which are hosted on Microsoft Azure.
For OpenAI Services specifically, this means that data, prompts, and outputs are:
- Encrypted and securely stored within the Azure infrastructure.
- Not shared with other customers.
- Not used by any services operated by OpenAI, such as ChatGPT.
- Not used to improve or train models, or any Microsoft or third-party products.
- Not interacting with any Microsoft services
Learn more about data, privacy and security at Microsoft Azure OpenAI Service.
Learn more about Staffbase security standards.
Is customer content used to train AI models?
No. Customer content is not used to train any third-party or Staffbase AI models. All customer content remains private and protected.
What language model does Staffbase use?
Staffbase currently uses a self-run model and OpenAI’s state-of-the-art language model, both of which are hosted securely on Microsoft Azure. This model is one of OpenAI’s most efficient and accurate to date. Updates to newer models are under evaluation and may be introduced to ensure best-in-class performance. See How do Staffbase AI features work below for specifics about which features use which language model.
What are Staffbase’s AI principles?
The use of AI in Staffbase products is based on the principles of TRUST: Transparency, Responsibility, User-in-the-loop, Security, Traceability. AI in the Staffbase platform is designed to support human communication, not replace it. Our AI TRUST Principles ensure that every feature is understandable, controlled, and built with care. This allows organizations to adopt AI without compromising on ethics or employee experience.
Can AI be customized or modified by users?
No. Users cannot modify the AI models, exclude specific inputs, or apply custom generation rules yet.
How does Staffbase ensure outputs are fair and unbiased?
Staffbase AI uses state-of-the-art language models from OpenAI, which are developed following industry best practices to reduce bias and ensure high-quality results. However, generative AI is generated based on likelihoods, not guaranteed facts. Therefore, it is not guaranteed that outputs will be fully accurate, exhaustive, or free from bias.
To help ensure fairness and reliability Staffbase:
- Selects only trusted, enterprise-grade models reviewed by our Security and Legal teams.
- Encourages customers to independently review and validate AI-generated content before sharing or publishing it, especially in critical use cases.
This approach balances the benefits of generative AI with the safeguards necessary for professional communication.
Can customers opt out of AI features?
Yes. All features with AI are optional so you do not have to use any features that are AI-powered. Reach out to Staffbase Support or your Customer Success Manager for more information.
Are AI outputs limited by content permissions?
Yes. Any Staffbase features using AI respect existing content permissions. Users can only generate content based on information they already have access to within their environment.
What legal safeguards are in place?
Use of AI does not alter any existing Staffbase contracts. AI-specific clauses are included in our Product-Specific Terms. Personal data processed through AI features is handled in accordance with your existing Data Processing Addendum.
How does Staffbase handle personal data in AI?
AI features are developed with privacy in mind. If personal data is processed, it follows the terms of your Data Processing Addendum (DPA) and is handled by approved sub-processors listed at staffbase.com/en/legal/subprocessors.
Does AI at Staffbase comply with the EU AI Act?
All Staffbase AI-powered features are reviewed regarding the EU AI Act. We consider our current AI-powered features to qualify as either limited or minimal risks under the EU AI Act. We recommend that customers assess the impact for their own industry and use case.
Staffbase does not have any high risk area AI features, for example biometrics, educational training, or law enforcement, as defined in Annex 3 in the EU AI Act.
We support our customers in ensuring the use of AI features by users is transparent. Ongoing compliance in relation to the EU AI Act and other AI regulations is reviewed regularly in collaboration with our Legal and Security teams.
How can I see in the Support Portal documentation if a feature uses AI?
Articles about features that use AI have an AI label. The label is situated either at the top of the entire article or at the top of the section where the AI is relevant.
Is generative AI usage data shared with customers?
Currently, we do not share usage metrics for generative AI features in customer reporting. However, a one-time manual report can be provided upon request.
How do Staffbase AI features work?
Content and On-Demand Translation: Staffbase provides on-demand content translation in collaboration with the third-party service Azure AI Translator. AI supports both editors and users to easily translate content into any of the languages added to the platform. This allows them to quickly create or read content in their preferred language, increasing speed and accessibility inside the platform. Service: Azure AI Translator.
Companion Editor: The Companion Editor assists communicators in drafting and improving internal news articles. It uses generative AI to suggest phrasing, headlines, summaries, and tone adjustments. The model helps streamline content creation while respecting brand voice. The Companion Editor generates suggestions based only on the editor's input. It does not learn from previous articles or access other content in the platform. Service: Microsoft OpenAI Services.
Smart Impact: Smart Impact uses AI to analyze how employees interact with your campaigns. It identifies communication patterns and surfaces actionable insights, such as how specific content types affect engagement. Smart Impact leverages analytics combined with AI logic to measure campaign effectiveness and optimize communication strategy, without using generative text. Service: Microsoft OpenAI Services.
AI in Staffbase Email: Staffbase Email offers AI-powered tools to help users write and refine. It can generate subject lines and summarize long messages. The AI suggestions are based on only the content within the email draft and are not trained on past communications or platform data. Service: Microsoft OpenAI Services.
Sentiment Analysis for Comments: Sentiment Analysis for Comments uses AI to detect the emotional tone of employee comments, labeling them as positive, neutral, or negative. It helps communicators quickly assess overall sentiment and identify potential issues. The feature uses non-generative AI techniques and respects content access controls. Service: self-hosted.
Alignment Surveys: AI supports Alignment Surveys by suggesting clear and relevant survey questions tailored to your communication goals. It helps save time by generating well-phrased items that reflect common engagement themes. Suggestions are based on best practices and customer input, but are fully editable by the survey creator. Service: Microsoft OpenAI Services.
News Drafting in Briefings: In Briefings, AI uses the information in the briefing, including the purpose, audience, tone, and attachments, to generate a news post draft. The AI-generated draft is created and pre-filled depending on the selected channel type. Inside the draft, users can further manually fine-tune or use the Companion Editor. Service: Microsoft OpenAI Services.
Video Subtitles: AI supports the automatic generation of subtitles for videos uploaded into the Staffbase platform. Subtitle generation occurs without additional user action. Text can then be reviewed and edited in the File Manager directly. Subtitles are available in different languages. Service: Microsoft Azure Video Indexer API
AI Podcast: The AI Podcast provides personalized audio podcast episodes to users or user groups based on published news posts. After selecting specific news channels and posts to include in the content, Administrators inform the AI about the general podcast and individual episode contexts. Users then receive summarized audio content generated by AI. Service: Microsoft OpenAI Services.
Content Governance: AI supports the Content Governance feature with the Duplicate Pages and Contradictory Pages issue types. It reviews pages to check for similar or conflicting content and alerts editors of potential need for review. Service: Microsoft OpenAI Services.
AI Page Creation: AI supports page creation in pages created with the Content Designer. Managing Editors can prompt the AI in natural language to create the structure and text in multiple languages. They can also upload files for the AI to review to help support the creation process. Service: Microsoft OpenAI Services.
Navigator: The Navigator allows employees to ask questions using voice or text in any language and instantly receive answers grounded in company knowledge. AI powers the Navigator by finding and responding to employee queries based on content found within the app and intranet. Service: Microsoft OpenAI Services.
AI Summary Block: In draft emails, the summary block compiles the email information into an easy-to-read overview. With one click, you can add an editable and adjustable block that can be dragged and dropped anywhere within the email. Service: Microsoft OpenAI Services.
Comments
0 comments
Please sign in to leave a comment.