Skip to main content
AI Copilot helps you design pipelines, write SQL, and troubleshoot runs without leaving Planasonix. You describe intent in natural language; the assistant proposes nodes, edges, queries, and explanations grounded in your workspace context where your plan allows it.

What AI Copilot does

Chat

Ask questions, iterate on designs, and get step-by-step guidance in the sidebar.

Pipeline generation

Generate or extend pipeline graphs from a short description of the data flow.

SQL assistance

Draft, refactor, and debug SQL against your connected sources.

Knowledge base

On Enterprise plans, ground answers in your organization’s own documents.
Copilot does not replace data governance: you still review generated SQL, credentials, and schedules before production promotion.

Availability

AI Copilot is feature-gated by subscription tier and admin policy. If you do not see the Copilot icon or entry points, your organization may have disabled AI features or your role may lack permission. Workspace owners can confirm entitlements under Plans and pricing and security settings.
Some regions or industries restrict use of third-party language models. Your legal team may require you to keep certain projects on manual authoring only.

Platform vs user API keys

Model calls can be billed and authenticated in two ways:
Platform-managed keys use Planasonix’s provider agreement. You do not paste your own vendor key; usage may count toward platform quotas or your contract’s AI add-on.
When BYO keys are required by policy, ensure rotation and least-privilege API tokens follow your internal standards.

Responsible use

  • Review generated pipelines in a dev environment before promoting.
  • Avoid pasting secrets, personal data, or regulated payloads into chat unless your security review approves that workflow.
  • Use feedback controls (thumbs up/down or comments) so your admins can tune quality over time.

Next steps

1

Enable Copilot

Confirm the feature is on for your workspace and choose platform or BYO keys per policy.
2

Open chat on a pipeline

Open a draft pipeline and start with a concrete goal (“ingest S3 JSON, flatten nested fields, load to Snowflake”).
3

Iterate with previews

Run previews and fix issues with SQL assistance before scheduling.