Steadybase
Core Concepts

AI Workers

Specialized AI agents that autonomously execute GTM tasks with their own roles, memory, and workflows.

AI Workers

Workers are the core execution units of Steadybase. Each worker is a specialized AI agent with a defined role, persistent memory, and one or more active Temporal workflows.

What is a Worker?

A worker is not a generic chatbot. Each worker has:

  • Role — A specific GTM function (Account Executive, BDR, Content Writer, etc.)
  • Memory — Persistent durable memory that accumulates over time
  • Workflows — Active Temporal workflows assigned to the worker
  • Status — Real-time execution state (running, waiting, stopped)
  • Current Task — What the worker is actively doing right now

The 8 Workers

Sarah AE West

Account Executive covering West region. Prepares call briefs, manages client onboarding workflows.

Mike BDR East

Business Development for East region. Scores calls, qualifies leads, routes prospects.

Deal Desk Bot

Automated deal desk. Processes discount requests, validates deal terms, enforces pricing rules.

Content Writer

Content generation agent. Creates QBR decks, call briefs, proposals, and research reports.

Call Scorer

Call analysis specialist. Scores Gong recordings, identifies coaching moments, tracks talk patterns.

Health Monitor

Customer health tracker. Monitors signals across accounts, flags at-risk customers, tracks champion changes.

Forecast Engine

Revenue forecasting. Calculates consumption forecasts, tracks EBAM/EAR, projects quarterly numbers.

Onboarding Bot

Client onboarding automation. Manages day-by-day onboarding tasks, tracks completion, escalates blockers.

Worker Lifecycle

Created → Running → [Waiting] → Running → ... → Stopped
              ↑         │
              └─────────┘
           (signal received)
  • Running — Actively executing a workflow or processing a task
  • Waiting — Paused, awaiting a human signal (approval, input) or external event
  • Stopped — No active workflows; idle until started

Worker Memory

Each worker maintains its own memory scope within the hierarchical memory system. Worker memory persists across workflow executions, allowing workers to learn and improve over time.

// Example: Reading worker memory
GET /api/memory/worker/sarah-ae-west
// Returns all memory entries for Sarah's worker scope

Starting a Worker

Workers are started by triggering a Temporal workflow:

POST /api/workers/sarah-ae-west/start
Content-Type: application/json
Authorization: Bearer <token>
 
{
  "workflowType": "client-onboarding",
  "params": {
    "accountId": "acme-corp",
    "accountName": "Acme Corp"
  }
}

The worker's status changes to running and a new Temporal workflow execution begins.

On this page