Docs

Emma

Emma

Emma is the AI analyst built into datagab. She investigates, builds and reports across all four products, and people approve every real action.

What Emma does

  • Investigates signals: gathers related logs and findings, writes up what happened, says how sure she is and suggests a next step.
  • Proposes actions: prepares the fix, such as revoking sessions or blocking public access, and asks a person to approve it.
  • Builds from a sentence: drafts dashboards, workflows and forms from a description.
  • Fixes code: opens pull requests for vulnerable libraries and unsafe code.
  • Finds owners: works out who owns a finding and asks them directly.
  • Reports: sends daily and weekly summaries on a schedule.

Approvals

Emma can read and suggest, but anything that changes a system waits for a person. You can approve in the console, in Slack or in Teams. Every proposal and approval is written to the audit trail.

Automations

An Emma automation is a job with instructions, the tools she may use and a trigger, such as a schedule. Create automations under Automation › Emma. You can pause, edit or run any automation at any time, and see what each run did.

What Emma knows

Emma reads across signals, cases, cloud and code findings, costs and logs. Under SOC › Emma Settings you can add a company profile, such as your offices, VPN ranges and key systems, so her answers fit how your company works. Agent Memory shows what she has learned.

Custom agents

Under Settings › Custom Agents you can create your own agents with their own name, instructions and tools. Custom agents use private models by default, or your organization's own OpenAI or Anthropic key if you choose.

Usage and cost

Every call Emma makes is metered. The Agent Console shows what Emma and custom agents did and what it cost, and AI budgets keep spend in check.