Studio

Studio is where teams build, test and run the AI agents and automations that do work in Figena. In Figena it appears as Cilila Manage.

Why Studio

Useful AI work repeats: the same report every Monday, the same follow-up when a deal moves. Studio turns that work into agents a team can build once, test before they go live, and run with the same permissions and records as everything else in Figena.

Studio is shown in Figena as Cilila Manage, the control panel behind Cilila.

What Studio does

  1. Knowledge with sources

    Give agents knowledge bases they can answer from, with each answer citing where it came from.

    • Knowledge bases with access control.
    • Answers that cite their sources.
    • Results filtered to what each person may see.
  2. Build visually

    Lay out an agent's steps on a canvas instead of writing code.

    • A visual builder for multi-step agents.
    • Versioned definitions, published when they are ready.
    • The same building blocks shared across teams.
  3. Tools and triggers

    Agents act through tools across Figena and connected services, and start on their own when something happens.

    • Tools for finance, people, CRM, eSign, IT and more.
    • Actions in connected services through governed orchestration.
    • Triggers on a schedule, from a signed webhook, or from an event in Figena.
  4. Choice of models

    Use the approved third-party model that suits each task, with automatic failover.

    • Models from providers such as OpenAI, DeepSeek, Qwen and NVIDIA.
    • Failover to another approved model when one is unavailable.
    • Data classified before it reaches any model.
  5. Tested before release

    Agents pass their own tests, and a person approves them, before they reach the team.

    • Test cases with expected results and a minimum pass rate.
    • Release approval by a named person.
    • Run history with cost, time and outcome for each run.

How Studio works with Cilila

  • One runtime

    Agents built in Studio run on the same engine as Cilila, with the same tools and limits.

  • Same governance

    Every agent action goes through Guard and OpenPort, as drafts where it matters.

  • Visible runs

    Each run is recorded, so a team can see what an agent did and why.

Where Studio helps

Team knowledge
An assistant that answers from the team's own documents, with sources.
Recurring work
Weekly summaries and follow-ups that run on a schedule.
Reacting to events
An agent that starts when a record changes in Figena or a webhook arrives.
Operations briefings
A daily brief of what changed and what needs attention in the workspace.

From idea to running agent

  1. Lay out the agent's steps in the visual builder.
  2. Give it knowledge bases and the tools it may use.
  3. Choose when it runs: on a schedule, from a webhook or on an event.
  4. Test it against its cases, and have it approved for release.
  5. Run it, and follow each run's history and results.

Research behind the platform

All research