Skip to main content

​
What are Agents?

Agents are AI assistants you shape for a specific job in your organization — not open-ended chatbots, and not traditional software that follows the same path every time. You give an agent a goal and guardrails (instructions, context, and capabilities). It reads the situation, decides the next step, uses the tools you allow, and adapts when things change. That is fundamentally different from a fixed automation: same flexibility, less guarantee — which is why thoughtful setup matters.
New to the idea of agents altogether? Start with What agents are — and aren’t. It explains the mental model in plain language before you configure anything in XSpada.
Screenshot placeholder Agent Manager landing page. Show the hub grid: Chat, My Agents, Chat Review, Studio, and Artifacts — with short descriptions under each card.

​
What you can do

  • Build specialists tuned to roles in your business — onboarding, research, drafting, triage, and more.
  • Equip them with workspace knowledge (linked files, registry data, assigned Actions), domain Focuses, and importable Skills.
  • Promote finalized agents to team members so they can work on Evolve tasks alongside your staff.
  • Test safely in the Lab and version configuration so you can iterate without breaking production use.
  • Chat directly, review past conversations across the workspace, and reuse outputs saved as artifacts.
  • Connect agents to Actions and Tracks so AI participates in real workflows — not just one-off chats.

​
The big picture

Every agent in XSpada is built from four layers you will hear about throughout this guide:
  1. Identity — name and description; who this agent is for your team.
  2. Instructions — system prompt, variables, and workspace intelligence snippets.
  3. Brain — which model powers the agent and how creative or concise it should be.
  4. Capabilities & knowledge — Focuses, assigned Actions, linked context files, conversation attachments, and Skills.
Think of it like hiring a specialist: the instructions are the job description, the brain is how fast or careful they think, and capabilities are the systems they are allowed to touch. Skills are playbooks you attach for specific procedures. Team promotion is when a finalized agent gets a seat to work on Evolve. The harness — versions, the Lab, Chat Review, guardrails in Actions and Tracks — is what turns a capable demo into something your organization can trust.
Diagram placeholder Four stacked layers (Identity → Instructions → Brain → Capabilities). A side arrow labeled “Harness” wraps versions, Lab, and review. User-facing labels only — no technical terms.

​
How Agents fit with the rest of XSpada

​
Where to find Agents

Open Agents from your workspace sidebar. The Agent Manager hub routes you to My Agents (build and manage), Chat (talk to any agent), Chat Review (workspace conversation history), Studio (combined build-and-test layout), and Artifacts (saved outputs from agent work).
Screenshot placeholder Sidebar with Agents highlighted. Include sub-nav when viewing an agent: Details, Chat, Studio, Artifacts.

​
Where to go next

What agents are — and aren't

The mental model: drivers vs. rails, loops, harnesses, and realistic expectations.

System prompt & variables

Instructions, placeholders, and the three levers: prompt, context, and tools.

Focus & capabilities

What your agent is allowed to do in the workspace.

Agent skills

Import playbooks and assign them to agents.

Promote to team member

Add finalized agents to your team for Evolve work.

Create an agent

Step-by-step: manual or AI-assisted creation.

Glossary

Quick definitions for every Agents term.
If you would like something unique to your business or have a unique issue, use XSpada Evolve or ask a XSpada rep.