Ask Zia - Agentic AI in BI
Generative AI is redefining how business users interact with data. In Zoho Analytics, every stage of the analysis, from data preparation and visualization to building powerful machine learning models, is seamlessly infused with AI. At the core of these capabilities is Ask Zia - Agentic AI (human led and AI assisted), a semi-autonomous system that not only understands the analytical intent but also provides intelligent recommendations and actions.
With a natural, conversational interface, it empowers users of different personas to build pipelines, complex queries, generate reports, insights, and perform actions, all through simple, intuitive prompts. This AI-driven assistant bridges the gap between data complexity and business decision-making with a human-in-the-loop.
Access to GenAI Capabilities
Users with Administrator Privileges can access the GenAI capabilities. The Account Admin can grant access to other Organization and Workspace Admins. Refer to the GenAI configuration article to learn more.
Ask Zia for Data Engineers
- Transform by Example: It helps you transform the data in any column to your needs just by providing examples of the output.
- Formula Suggestion: Finding the correct formula for analyzing data can be tedious and time consuming. With the Formula suggester, you can create formulas by asking questions in simple natural language and get intelligent suggestions.
- Query Suggestion: Get accurate SQL queries by defining the logic in simple terms. Ask Zia interprets the request and generates a valid SQL query statement based on the data tables.
- Data Pipeline: Build and automate data pipelines with ease. Instead of manually configuring each data transformation step, describe your data preparation goals in plain language; Ask Zia will create the complete pipeline for you.
Ask Zia for Data Analysts and Business Users
- Report Creation: Create insightful reports by describing your requirements. Ask Zia understands the analytical intent, generates suitable reports, applies relevant filters or groupings, and recommends follow-up views or questions, thereby making report building faster, smarter, and accessible to all users.
- Dashboard Creation: Whether you’re building dashboards from scratch or enhancing existing dashboards, Ask Zia understands your layout preferences, content needs, and visual goals and delivers a personalized experience every step of the way. It allows you to add or remove views and change themes.
- Insight Generation: Deciphering insights in the reports or visualizations could be tedious and time intensive. Zia Insights provides contextual narratives - Descriptive, Predictive, Diagnostic, and Prescriptive insights to understand the What, Why, and How behind any outcome.
- Sharing and Collaboration: Instantly access, schedule emails, and export reports and dashboards by just typing the prompts with the action skills.
- Data Stories: Convert reports and dashboards into compelling, narrative-driven slideshows with Ask Zia.
Embed Ask Zia
Embedding Ask Zia, the AI-powered analytics assistant, brings advanced analytical capabilities like conversational insights and Key Driver Analysis (KDA) directly into the applications users already work with. Users can ask business questions, explore data, detect patterns or anomalies, and identify the key factors driving their metrics without switching platforms. This makes data interaction more seamless, accessible, and insight-driven across the organization.
The GenAI Architecture
To power the AI capabilities, Zoho Analytics predominantly provides two LLMs - the Zoho LLM for report creation, and OpenAI LLM for capabilities such as data pipeline, report and dashboard creation, and insights generation.The administrators can choose which LLM to use based on business requirements and privacy.
Security, Privacy and Governance
At Zoho, we prioritize privacy from the ground up. For AI-powered features, the information shared with large language models (LLMs) is strictly limited based on the task type. In most cases, only prompts and metadata such as column names, data types, and the data model, are shared. Columns marked as Personally Identifiable Information (PII) are never shared under any circumstance.
For insight recommendations, only aggregated data is sent to the LLM to maintain both accuracy and confidentiality. You can manage and customize these data-sharing preferences in the GenAI Configuration settings page.