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

# Kowalah Agent

> Kowalah, right where you work — an AI adoption partner in your Slack threads, Teams channels, and Google Chat spaces, at the moment of need.

The Kowalah Agent is the front end of the Kowalah experience — delivered right where your team already works. Anyone in your organization, from the CEO to the intern, can `@mention` Kowalah in Slack, Teams, or Google Chat the moment they're doing some work and think *"I wonder if AI could help us with this."*

A finance team in the middle of month-end? `@Kowalah` in their thread to figure out how AI could speed up the variance commentary. An HR lead writing job descriptions? `@Kowalah` to design an AI workflow that drafts them properly. A sales director with 30 customer interview transcripts to synthesise? `@Kowalah` for a working approach.

No new tool to learn, no login to remember. Same Kowalah that lives in the [platform](/platform/overview) and [Claude Desktop](/integrations/mcp/overview) — just wherever the moment of need actually happens.

<Info>
  The Kowalah Agent is available to Kowalah Managed Services customers, and during a paid Kowalah programme. [Learn more about how it works](/get-started/how-it-works).
</Info>

## What it does

The Agent exists to drive AI adoption inside your organization. It helps employees apply AI to the work they're already doing, in three ways.

### Answer "could AI help us with this?"

Bring the Agent any real task you're working on, and it'll help you figure out how AI could fit. It draws on patterns from across Kowalah's customer base — what's worked, where the friction usually is — and gives you a concrete answer, not a generic feature list.

```text theme={null}
@Kowalah How could we use AI to speed up our month-end
variance commentary?
```

### Design an AI workflow

For recurring tasks — board packs, weekly reviews, customer summaries, hiring rounds — the Agent designs an AI workflow you can run on demand, or scopes it as proper work for Kowalah to build. For analysis, synthesis, and drafting, it often does the work right there in the thread.

```text theme={null}
@Kowalah Help me design an AI workflow for our weekly KPI
report. It takes three hours every Friday.
```

### Find where AI fits in your role

When you don't know where to start, the Agent walks you through what's on your plate and surfaces 2-3 high-value AI use cases — grounded in your actual work, not a generic role template.

```text theme={null}
@Kowalah Where could AI make the biggest difference in my
week as Head of People at a 1,500-person manufacturer?
```

## How the Agent helps you act

Once you know *what* you want to do with AI, the Agent has three ways to help you act on it:

* **Find an accelerator** — search Kowalah's global library plus your organization's private library for an existing prompt, GPT, training, or workflow that already solves your problem
* **Run it together in the thread** — for analysis, synthesis, and drafting the Agent can do directly, it'll do the work live and iterate with you
* **Raise an Opportunity** — when the work needs proper scoping, ongoing tracking, or specialist help, the Agent captures it as an [Opportunity](/platform/opportunities) for the Kowalah team to triage into a [Deliverable](/platform/projects/deliverables), an [Expert Request](/platform/expert-requests), or a new [Project](/platform/projects/overview)

The Agent always asks before raising anything and describes exactly what it will write.

## How it works

Mention `@Kowalah` followed by your request in any channel, thread, or direct message:

```text theme={null}
@Kowalah How could we use AI to triage our 400 daily
support tickets?
```

The Agent acknowledges within seconds, reads your workspace context, and responds in the same thread. Follow-up messages continue the conversation — no need to repeat context.

The shape of every conversation: **understand the work → suggest an AI approach → find an accelerator or design a workflow → (sometimes) raise an Opportunity**.

## Example interactions

<CardGroup cols={2}>
  <Card title="Speed up month-end" icon="calculator">
    "@Kowalah How could AI help us write month-end variance commentary faster?"
  </Card>

  <Card title="Design a board-pack workflow" icon="workflow">
    "@Kowalah Help me automate the leadership pack I build every Monday"
  </Card>

  <Card title="Triage customer feedback" icon="messages-square">
    "@Kowalah What's a good AI approach for synthesising 30 customer interviews?"
  </Card>

  <Card title="Draft job descriptions" icon="briefcase">
    "@Kowalah Help me build an AI workflow for drafting sales job descriptions"
  </Card>

  <Card title="Find where AI fits" icon="lightbulb">
    "@Kowalah Where would AI make the biggest difference in my week?"
  </Card>

  <Card title="Find an accelerator" icon="rocket">
    "@Kowalah Is there an accelerator for board reporting?"
  </Card>
</CardGroup>

## What your role changes

Your [organization role](/platform/organization/roles-and-permissions) affects what the Agent can see for you, not what it can do. Anyone in the organization can ask AI questions, design workflows, and raise an Opportunity.

| Role          | What you see                                                                                      |
| ------------- | ------------------------------------------------------------------------------------------------- |
| **Admin**     | Everything in your organization                                                                   |
| **Core team** | Everything in your organization                                                                   |
| **Member**    | Opportunities you raised, and projects, deliverables, and Expert Requests you're a stakeholder on |

Pending discoveries are personal — you always see your own invites.

## Response times

| Scenario                      | Typical time  |
| ----------------------------- | ------------- |
| First message in a new thread | \~20 seconds  |
| Follow-up in the same thread  | 3-5 seconds   |
| Simple lookups                | 5-10 seconds  |
| Complex analysis              | 15-30 seconds |

<Tip>
  Use threads for multi-step conversations. The Agent remembers thread context, so you can iterate on an AI workflow design across several messages without repeating yourself.
</Tip>

## Set up the Agent

The Agent needs to be installed in your Slack, Teams, or Google Chat workspace before you can use it.

<CardGroup cols={3}>
  <Card title="Slack" icon="hash" href="/integrations/agent/setup-slack">
    Installation guide
  </Card>

  <Card title="Microsoft Teams" icon="users" href="/integrations/agent/setup-teams">
    Installation guide
  </Card>

  <Card title="Google Chat" icon="message-square" href="/integrations/agent/setup-google-chat">
    Installation guide
  </Card>
</CardGroup>
