Agents work from shared context
Agents can use connected inboxes, company knowledge, schedules, reusable skills, and business tools instead of depending on context pasted into each conversation.
Manor AI is an AI business workspace—also called a business AI workspace—for ongoing small-business operations. Goals, agents, company context, inboxes, schedules, approvals, citations, and work history stay connected across days and months, without turning every process into code.
An AI business workspace for small business is a persistent operating layer where agents keep working from durable goals, connected inboxes, company knowledge, schedules, approvals, citations, and activity history. Unlike a standalone chatbot or builder, the workspace preserves context and keeps recurring work moving between sessions.
Manor AI is designed around the operating loop of a small business, not a blank chat window. The workspace keeps the agent, the evidence it used, the action it prepared, and the human decision in one reviewable system.
Agents can use connected inboxes, company knowledge, schedules, reusable skills, and business tools instead of depending on context pasted into each conversation.
Approval gates, citations, escalation rules, and activity logs keep sensitive sends and decisions visible before an agent changes customer or business state.
Teams can turn a successful process into an agent skill, scheduled check, or repeatable workflow rather than rebuilding the same instructions in separate chats.
Manor's source-available self-hosted edition lets teams inspect the architecture and operate their own workspace. It uses the Sustainable Use License 1.0, not an OSI-approved open-source license.
Agents need more than a prompt. They need trusted documents, inbox history, schedules, customer context, policies, project notes, and the current state of the business.
Skills turn repeatable judgment into reusable capabilities: draft in the company voice, cite a source, check a policy, summarize a thread, prepare a report, or create a follow-up.
The workspace should show what ran, which source was used, where the agent needs approval, and which actions were logged after the workflow finished.
These tools can work together, but they solve different parts of the job. A chatbot handles a conversation. An automation builder follows explicit trigger-and-action logic. A business AI workspace keeps agents, company context, tools, review rules, and work history together so the business can supervise multi-step operational work.
| Capability | Chatbot | Automation builder | AI business workspace |
|---|---|---|---|
| Primary job | Answer questions and draft text. | Run predefined steps when a trigger fires. | Coordinate agents, knowledge, tools, approvals, and follow-up work. |
| Business context | Current conversation and pasted files. | Inputs explicitly passed into each step. | Connected sources, inbox history, schedules, skills, and current work state. |
| Decision model | Generate the most useful response. | Follow deterministic branches and conditions. | Use agent judgment inside defined permissions and escalation rules. |
| Human control | Review the answer manually. | Inspect runs and handle error paths. | Review cited evidence, approve sensitive actions, and inspect activity logs. |
| Best fit | Ad hoc research, writing, and questions. | Stable, repeatable system-to-system processes. | Operational work that combines context, preparation, routing, action, and review. |
The strongest workspace is not the one with the longest feature list. Test whether it can complete one real business job with the right evidence, boundaries, and handoff. Use these six checks before expanding to more agents or teams.
Can the workspace use the approved inboxes, documents, customer history, schedules, and tools needed for the job without making the operator rebuild context every time?
Can a reviewer inspect the evidence behind a claim? A useful AI knowledge base with citations should reveal when the source is missing or insufficient.
Can the business separate safe preparation from sensitive sends, price changes, refunds, access changes, and other actions that need approval?
Can a successful process become a reusable skill or workflow instead of remaining a one-off chat that disappears when the conversation ends?
Can the team see what ran, which sources and tools were used, what is waiting for review, and which follow-up still remains open?
Can you measure time-to-review, accepted drafts, source gaps, missed escalations, or completed follow-ups for one narrow workflow?
Start with one operating loop that a person can review in minutes. The goal of the first version is not maximum autonomy; it is a reliable path from incoming work to an evidenced next action.
Pick a repeated task such as inbox triage, lead follow-up, customer support routing, a weekly report, or a scheduled operations check. Define the trigger, desired output, owner, and success metric.
Add only the policies, SOPs, message history, customer records, schedules, and tools the first workflow needs. Mark the approved source of truth for each important decision.
Define the goal, required sources, reusable skills, allowed outputs, and failure behavior. The AI agent builder guide shows how those pieces fit together.
Let the agent prepare low-risk work, but require review for money, external commitments, policy conflicts, missing evidence, customer emotion, or access changes. Keep the reason in the approval and activity log.
Inspect accepted drafts, source gaps, incorrect routes, missed follow-ups, and time saved. Expand to another action or workflow only after the first loop is predictable.
An AI business workspace does not have to start as an empty builder. Manor users can install a complete, editable workspace Blueprint based on a tested operating method. Through the AI Business Workspace Creator Marketplace, creators can package useful systems and buyers can put them to work in their own account. Buyers and creators can also browse the live AI Workspace Marketplace.
Install a workspace that already defines its goals, agents, skills, knowledge structure, workflows, schedules, approvals, safeguards, and setup guidance instead of starting from a blank prompt.
The installed Blueprint becomes an editable copy in the buyer's Manor account. Buyers connect their own approved data and tools; a creator's private workspace, credentials, and customer data are not transferred.
After setup, the workspace keeps goals, context, schedules, open work, approvals, and activity history connected across sessions. It is an ongoing operating environment, not a one-time AI output.
Yes. Both phrases describe a shared operating layer for agents, skills, company knowledge, connected tools, workflows, approvals, citations, and activity logs.
It is a shared operating layer for agents, skills, sources, tools, schedules, workflow runs, approval rules, citations, and logs.
Yes. Manor AI is a long-running AI business workspace for small businesses, solopreneurs, and one-person businesses that need goals, agents, context, schedules, approvals, and recurring operations to continue across days and months.
Yes. A long-running workspace preserves goals, company context, schedules, open follow-ups, approval state, and activity history so recurring operations do not restart from a blank chat.
Yes. A Manor workspace Blueprint packages a complete operating design that you can install as an editable copy, connect to your own approved context and tools, and keep running over time.
Yes. Creators can turn a tested operating method into an installable workspace Blueprint with agents, skills, knowledge structure, workflows, schedules, approvals, and setup guidance, then publish it through the Manor Workspace Marketplace.
No. Manor AI is meant to work at the agent and business-context layer. Its platform direction includes compatibility with n8n-style workflows so deterministic automations can stay useful.
Approvals let agents prepare useful work while keeping risky decisions, customer-facing sends, refunds, pricing, contracts, and account changes under human review.
Look for connected company knowledge, source citations, permissioned tool access, reusable workflows, approval gates, activity logs, and a clear way to measure whether the first workflow saves time without hiding mistakes.
Choose one measurable operating job, connect only the trusted sources and tools it needs, define the agent's allowed actions, place approval gates around sensitive decisions, review citations and logs, and expand only after the workflow is reliable.
It is for small business owners, solopreneurs, indie founders, agencies, property managers, and lean teams that need AI to help with real operating work.
For a small service team, use an AI business workspace for client delivery to keep the approved brief, draft tasks, review decisions, and launch package together. Our illustrative studio workflow includes task acceptance criteria and a pilot checklist.
Keep goals, context, agents, schedules, workflows, approvals, citations, and activity history connected as the business keeps moving.