> ## Documentation Index
> Fetch the complete documentation index at: https://workllm.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# AI Agents | WorkLLM

> WorkLLM gives you ready-made AI agents for every department, sales, marketing, HR, support, and more, so your team can delegate real work to AI.

AI Agents are WorkLLM's ready-made, purpose-built workflows for specific tasks and questions. As soon as your team joins WorkLLM, they have access to agents for Sales, Marketing, HR, Product, Engineering, and more, with no setup, no prompting expertise, and no waiting.

<Note>
  A full directory of ready-made AI Agents will be added to these docs as a separate page. This page covers how agents work.
</Note>

## Three types of AI Agents

Not every task is the same. WorkLLM gives you three types of AI Agents so your team always has the right tool for what they need to get done.

<AccordionGroup>
  <Accordion title="Knowledge Agents">
    Ask anything, get an answer grounded in your company's actual knowledge. Knowledge Agents connect to your organization memory, docs, and apps, so your team gets reliable, context-aware answers instead of searching through files or asking a teammate.

    **Great for:** product Q\&A, policy questions, client context, competitive research, onboarding
  </Accordion>

  <Accordion title="Task Agents">
    Run a specific piece of work with one click, no prompting required. Task Agents handle repeatable, structured work your team does daily: fill in a few details and get a finished output that anyone can run, no AI experience needed.

    **Great for:** writing, summarizing, drafting, researching, reporting, and anything done more than once a week
  </Accordion>

  <Accordion title="Workflow Agents">
    Agents connected to your work apps that run on their own. Set a trigger, an app event, or a schedule, and the agent handles the rest in the background, with no one needing to remember to start it.

    **Great for:** multi-step automations, scheduled reports, event-triggered tasks
  </Accordion>
</AccordionGroup>

## Every agent is a starting point, not a ceiling

Customize any ready-made agent to fit your exact workflow, or build a new one from scratch, no technical skills needed.

<CardGroup cols={2}>
  <Card title="Customize a ready-made agent" icon="sliders">
    Change the inputs, adjust the output format, update the tone, or connect it to a different data source. It takes minutes, not days.
  </Card>

  <Card title="Build from scratch" icon="hammer">
    Have a workflow that doesn't fit any of the ready-made agents? Define what it needs, what it does, and who on the team can run it. No coding or prompting expertise required.
  </Card>
</CardGroup>

## Powered by organizational memory

WorkLLM agents draw from your org memory automatically. No context to paste in, no setup required.

* **No copy-pasting context** — Your products, tone, and client details are already in org memory. Every agent pulls them automatically before anyone types a thing.
* **Consistent across your whole team** — Everyone runs agents grounded in the same company knowledge, instead of five different versions of the same email or proposal.
* **Gets smarter over time** — Every great output saved to org memory makes every future agent run better.

## Security & privacy

Every WorkLLM workspace is isolated. What your agents know and do stays inside your workspace.

<CardGroup cols={2}>
  <Card title="Dedicated tenant per customer" icon="building-shield">
    Your workspace runs in its own isolated environment, not shared infrastructure.
  </Card>

  <Card title="Zero data retention at the model layer" icon="ban">
    The AI models your agents use never store or train on your prompts or outputs.
  </Card>

  <Card title="Role-based access controls" icon="key">
    Admins control who can run agents, who can build them, and who can access which data.
  </Card>

  <Card title="Full audit logs" icon="file-lines">
    Every agent run is logged, so admins can see who ran what, when, and what the output was.
  </Card>
</CardGroup>

## FAQs

<AccordionGroup>
  <Accordion title="What is an AI Agent in WorkLLM?">
    An agent is a purpose-built AI workflow for a specific task or question. Instead of figuring out what to ask and how to ask it, your team picks the right agent, fills in a few details, and gets a finished output. No prompting knowledge needed.
  </Accordion>

  <Accordion title="What's the difference between a Knowledge Agent and a Task Agent?">
    A Knowledge Agent answers questions by pulling from your org memory and connected apps, things like product details, client context, or past decisions. A Task Agent handles a specific repeatable job, like drafting a proposal, summarizing a call, or writing a job description. Both are ready to use from day one.
  </Accordion>

  <Accordion title="Do my team members need any AI or prompting skills to use agents?">
    No. Agents are designed so anyone on your team can run them without knowing how to write prompts or pick the right model. They fill in the details the agent asks for and get the output.
  </Accordion>

  <Accordion title="Can I customize the ready-made AI Agents?">
    Yes. Every AI Agent can be customized to fit your exact workflow. You can change inputs, adjust the output format, update the tone, or connect it to a different data source. Changes take minutes, not days.
  </Accordion>

  <Accordion title="Can I build my own AI Agents from scratch?">
    Yes. If none of the ready-made agents fit a specific workflow your team needs, you can build one from scratch. Define what it needs, what it does, and who can run it. No coding or technical skills required.
  </Accordion>

  <Accordion title="How do agents use organization memory?">
    Agents automatically have access to org memory and pull relevant context from it, so you don't need to specify the same context in multiple places. If you'd rather run an agent without memory context for a specific task, you can do that too.
  </Accordion>

  <Accordion title="Which apps do AI Agents connect to?">
    AI Agents can connect to Google Workspace, Slack, Jira, HubSpot, Notion, Salesforce, and more. Once connected, agents pull relevant context from your apps automatically so your team doesn't have to copy-paste anything in.
  </Accordion>

  <Accordion title="Is my data safe when running AI Agents?">
    Yes. Every WorkLLM workspace runs in a dedicated, isolated environment. Your data is never shared with other organizations, and the underlying AI models never retain or train on your prompts or outputs.
  </Accordion>

  <Accordion title="What are Workflow Agents?">
    Workflow Agents run automatically, triggered by a specific app event or a schedule, without anyone needing to start them.
  </Accordion>
</AccordionGroup>

## Next steps

<CardGroup cols={2}>
  <Card title="Request a Demo" icon="calendar" href="https://calendar.app.google/Muovg96CzondTeYw6">
    Talk to the WorkLLM team and see agents set up around your workflows.
  </Card>

  <Card title="Contact Us" icon="envelope" href="https://workllm.io/company/contact">
    Have a question or want to talk to the team? Reach out and we'll get back to you.
  </Card>
</CardGroup>