Guide
Fibery has AI built into many places - and that can give both power and confusion. This guide gives you a map: what each AI capability does, where to find it, and one example to make it concrete. For deeper details, each section links to a dedicated guide.
AI Builder - build workspace structure from a prompt
Describe what you need, and AI builds/adjusts your workspace structure: databases, fields, relations, views, even automation rules — all from plain language. Also useful for bulk data operations, restructuring, and even pulling in data from external APIs.
Beyond Builder, you'll also run into AI assistance in smaller contextual spots — for example, writing formula syntax or generating markdown content inside an automation step. These aren't Builder — just lightweight helpers that show up where you need them.
Check Fibery AI Agent for more details.
AI Ask Mode - ask questions, get answers
This is your interactive AI Smart agent. You ask a question, it reasons over your data and gives a structured answer. Works at three levels of focus:
Workspace level — sees almost everything. Best for open-ended or cross-system questions.
Example: "What changed while I was on vacation last week?"
View level — works within a specific view and its filters. Best when you want to analyze a slice of data or understand a view someone else built.
Example: "Explain how this report was built and what it's showing."
Entity level — goes deep on one record: its fields, comments, history, and connected data. Best for understanding what happened around a specific item.
Example:
“For this customer, who promised them a discount and when?”
AI in Automations - AI that works without you
Instead of asking AI manually, you define rules and let it run in the background.
AI in Automations - lightweight, repeatable steps: summarize, classify, extract. Best for high-frequency, well-defined tasks.
Example: Assign a sentiment score to every incoming customer feedback automatically.
AI Agent in automations - more complex, multi-step reasoning. Best when the task requires judgment, not just text processing.
Example: "When a customer is lost, make an analysis."
Supporting tools - everyday helpers
There are also many small features that reduce friction without driving big workflows:
External AI (MCP, n8n) — connect Fibery to outside systems
Fibery MCP Server - use Fibery data as context for external AI agents, or trigger actions in other tools based on Fibery changes. For teams with advanced setups beyond what's built in.
Where to go from here
Fibery AI works at every layer - from building your workspace structure to analyzing a single record to automating repetitive work in the background. The best way to start is to pick one area that matches something you already do, and try it there.
Not sure where to begin? Start with AI Ask Mode at the workspace level - ask it something you'd normally dig through manually. That's where it can click.