Skip to main content

How the company runs

You stay the CEO. Flolah is an AI Company OS: you give outcomes in plain language, the COO runs the company, named AI employees do the work, and you supervise on Kanban, Digest, and the bell.

The same diagrams appear in the product: a compact strip above the Home dashboard, and the full view at the top of My Org.

How the company runs on Flolah — AI-native operating system for the modern enterprise

Platform map: who uses Flolah, the capabilities that get work done, governance, observability, and systems you can connect. Click for full size.

You
Stay in charge
Human CEO
Decide, approve, teach the company
Direction
Plain language
Home chat
COO selected by default
Scheduled goals
Recurring outcomes
Broadcast
Same ask to many
Company setup
Who we are
COO
Chief of staff
Plan · match purpose · Kanban · track
Routes specialty work; keeps goals, standups, and status itself
AI employees
Named roles
Platform Help
How-to from the guide
Workflow Builder
Build or repair graphs
CRM / ERP SMEs
Maker then Checker
Specialists
Research, content, discovery
Systems
Only what you granted
Knowledge
Tables, docs, search
Tools & Connectors
Catalog + business apps
Workflows
Repeatable capability
CRM · ERP · Browser · Channels
Live systems
You see it
Supervise
Kanban
Work and artifacts
Digest / Workspace
Pulse and open tasks
Bell
Results and alerts
Approvals
Gates on Kanban

Who you talk to

NeedAskWhat happens
Get this outcome doneCOO (Home chat, default)Plans, hands off to a specialist, opens a Kanban card, notifies you when ready
How do I use Flolah?Platform HelpAnswers from the product guide. Does not run the company for you
Build or repair a workflowWorkflow BuilderTurns plain English into a visual workflow. Secrets stay in API Keys
Leads and pipelineCRM Maker / CheckerMaker proposes; Checker (and you, when gated) confirm
Orders, invoices, booksERP Maker / CheckerQuote → order → invoice under Maker/Checker, with CEO gates on spend
Same ask every weekdayScheduled goals → COOCOO drafts a plan; you approve once; it fires on cadence

Prefer one clear outcome per message. You do not need to pick a specialist first.

A typical journey

  1. Hire the company — Company setup (or Hire AI employee) gives you a COO, departments, and specialists with roles and memory.
  2. Give direction — Home chat, scheduled goals, or Broadcast.
  3. COO routes work — Matches each employee’s purpose, starts Kanban, keeps goals and standups itself.
  4. Work uses your systems — Knowledge, Tools, Workflows, CRM/ERP, Browser Session, channels — only what you granted.
  5. You supervise — Kanban, Digest, the bell, and approval cards. Policies set what can run alone vs what needs you.

What the layers mean

LayerMeaning
ExperienceScreens you open: Home, Digest, Workspace, My Org, Kanban
Direct & superviseCOO routing, goals, standups, Maker/Checker, approvals, bell
OS primitivesPeople, Employees, Departments, Knowledge, Tools, Policies, Workflows, Tasks, Approvals, Memory
CapabilitiesTools, Connectors, MCP, Browser, CRM/ERP, channels
Compose & productizeVisual workflows, AgentExchange, Windows package
GovernanceEntitlements, API Keys, budgets, IP allowlists, Deny-all published services
RuntimeAgentSystem, Flolah APIs, schedulers, your isolated company data

The unit of value is a named AI employee in your org, not a workflow node. Workflows are how you compose a capability; AgentExchange is how you offer it outward — still under your policy.

Next: Human-directed agent work · Your first 15 minutes · Chat with the COO