Agent Builder
Build and deploy AI agents with full control
Create analytics agents grounded in your data and governed by your rules, for every user, team, and customer.
Launch and scale analytics agents
in minutes
Easy development with or without code
Start via a template or describe your agent in plain language, and configure everything in the UI or via the API.
Context-aware and grounded by default
Automatically connect every agent to your business data, definitions, and rules, with accurate, consistent, and secure outputs.
Observable pre-built agentic framework
Enable your agents to plan, execute, and self-correct across multi-step tasks with a built-in reasoning framework.
Define your agent’s role, behavior,
and capabilities in one place
Skills
Give any agent the right skills for the task
- Select all available skills or configure a custom set per agent.
- Choose from skills like anomaly detection, executive narratives, and more.
- Get a detailed overview of each skill’s configuration.
Personality
Define how your agent behaves and remembers
- Set instructions for tone, role, and communication style.
- Store key facts such as company context or project details.
- Configure how memory is used during conversations.
Knowledge
Enrich your agent with your company knowledge
- Connect to documents, playbooks, and internal content.
- Enable semantic search across your information sources.
- Keep your agent aligned with your company knowledge and context automatically.
Role Permissions
Control who can use and access each agent
- Enable or disable agents across workspaces and user groups.
- Update configurations and track changes over time.
- Configure and scale across as many customers or tenants as needed.
Observability
Control and monitor everything you build
- Inspect interactions with traces and logs.
- Evaluate outputs and monitor performance over time.
- Track adoption and usage across agents and workspaces.
Build and integrate industry-specific AI agents into your workflows
Root cause analysis for finserv and banking
Sales optimization for e-commerce
Inventory optimization for Hospitality
Operational overview for healthcare
Campaign analysis for marketing
AI agents for every user and team
that works with data
Developers
Embed agents directly into products and trigger via API, scoped exactly to customer needs.
Data Analysts
Investigate KPI changes, surface anomalies, run root cause analysis, and explain what drove them.
Data Engineers
Handle data modeling work, and metric definitions consistency to free the team to focus on architecture.
Business Users
Answer questions in plain language and deliver clear, reliable outcomes grounded in real company data.
Common Questions
An AI agent is a configurable reasoning unit that can understand a goal, decide which approved capabilities to use, and complete a task in context. In GoodData, agents work over governed analytics capabilities rather than acting as standalone chatbots. They can interpret requests, use the right skills, retrieve relevant business context, and return results that reflect the user’s data, definitions, and permissions.
No. GoodData is designed so teams can configure and manage agents without starting from custom code. Builders can define behavior, control access to skills and knowledge, and test how agents respond in different scenarios through a governed management layer. Technical teams can go deeper when needed, but getting started does not depend on building everything from scratch.
This is where architecture matters. In GoodData, agents operate on a governed foundation: semantic definitions provide business meaning, approved skills define what actions are allowed, and controlled platform access ensures agents work within existing permission boundaries. AI can also be grounded with organizational knowledge and managed centrally, so responses stay aligned with the right metrics, context, and access rules.
A practical starting point is to define the use case first: what the agent should help users do, which knowledge it should use, and which skills it should be allowed to access. From there, configure the agent’s behavior, assign the right grounding and permissions, and test it against realistic questions and workflows. The goal is not just to make the agent respond, but to make sure it behaves consistently, uses the right context, and delivers useful results within governed boundaries.
