Studio Assistant FAQ
Frequently asked questions about Visier's AI-powered support layer for Studio and our approach to implementing AI technologies.
Overview
What is Studio Assistant?
Studio Assistant is an AI-powered support layer for Studio, Visier’s interface for managing the people analytics lifecycle. The following Q&A is intended to help customers understand how Visier is currently employing AI technologies.
What specific tasks can the Studio Assistant help with?
It acts as a navigator for Studio tasks, such as assisting with data model management, configuring security settings, and understanding how to customize the look and behavior of your analytics solution.
How does the Studio Assistant generate accurate responses?
The system uses Retrieval Augmented Generation (RAG) to identify your question, retrieve the most relevant sections from official Visier documentation, and provide that specific context to the AI to generate a grounded answer.
What is the Translation Layer in the architecture?
The Studio Assistant acts as a bridge, using the natural language capabilities of LLMs to effectively translate your conversational questions into precise technical prompts that the Studio system can process.
Large language models (LLMs)
Which LLMs power the Studio Assistant and where are they hosted?
The solution utilizes Claude Haiku 4.5, Sonnet 4.5 and Opus 4.5 models from Anthropic. These models are hosted within AWS to ensure enterprise-grade reliability and security.
Is my data used to train these global models?
No. Our LLM partners are SOC 2 certified and have contractually committed to not persisting or using any data sent by Studio Assistant to train their own models.
Training and knowledge base
How is the Studio Assistant trained to know about Visier?
We don't train the AI in the classical sense. Instead, we provide it with a curated library of Visier University content and public documentation (docs.visier.com) to reference whenever a query is made.
How often is the knowledge base updated?
To ensure the Studio Assistant stays current with new product features and documentation, the knowledge base is refreshed on a weekly basis.
Transparency and bias
How does Visier address potential AI bias?
Visier utilizes LLMs that employ specific mitigation techniques, including careful data acquisition and algorithmic fairness. We also perform continuous monitoring and evaluation to ensure the system remains transparent and reliable.
Where can I find more information on Visier’s ethical AI stance?
For deeper insights into our processes, see Visier’s Bias Prevention and Transparency Statement.
Customer control and privacy
Does the Studio Assistant process personal employee data?
While Studio Assistant primarily uses documentation to answer how-to questions, other AI-driven features within the Studio environment—like AI Automap—do interact with your organization's actual data to perform their functions. Any data processed by Studio Assistant or AI Automap remains strictly isolated within your specific tenant scope. This data is never shared between customers, and Visier does not use your organizational data for training or fine-tuning.
How long does Visier keep my query history?
Queries and feedback are stored for 30 days before being purged from our system to maintain data privacy.
How can I provide feedback on the AI’s performance?
Users can upvote or downvote specific responses. This feedback is tracked and investigated by Visier to improve Studio Assistant’s accuracy in future updates.
Under the hood: Data modeling
What is data modeling in the context of Visier?
Data modeling is the process of organizing and standardizing disparate workforce data into a structured format. In Studio, this means taking disparate data from various sources using an automated ETL (Extract, Transfer, and Load) process and mapping it into a unified framework so it can be used for accurate, cross-platform analytics.
What is AI Automap and how does it work?
AI Automap is a specialized feature that uses machine learning to accelerate data ingestion. It analyzes your source data values—such as column headers and sample records—to suggest the best fit for the Visier data model. While it processes data attributes to make these suggestions, it operates under Visier’s strict security protocols and does not use your data to train global AI models.
What is the Event-Based analytic model?
Unlike traditional static databases, Visier uses an event-based model built to handle the constant flux of people data. This allows the system to accurately track time-based changes in the employee journey, such as promotions and transfers.
How are security rules enforced within Studio?
Admins define role-based permissions that control access to specific populations or attributes (like gender and ethnicity). These security rules are applied consistently across the application, ensuring users only see data they are authorized to view.
How does data modeling ensure data integrity?
The model acts as a single source of truth, ensuring that everyone in the organization is using the same definitions for metrics like Headcount or Turnover, regardless of how the data was named in the original source system.