# Ask an Agent Team

> The Ask an agent team step sends a prompt to one of your AI Teams. The step returns varied responses, based on how you set it up.

- Source: https://www.taskade.com/learn/automation/agent-team
- Section: Automation
- Updated: 2026-11-09
- Index of all pages: https://www.taskade.com/learn/llms.txt

## Overview

With the **Ask an agent team** step, you can send prompts to your existing [AI Teams](https://www.taskade.com/learn/agents/ai-teams). The action returns a set of varied responses, shaped by how you set it up. You can reference that output in later actions.

> **TL;DR:** **Ask an agent team** is an automation step. It sends a prompt to one of your [AI Teams](https://www.taskade.com/learn/agents/ai-teams). The team's response goes back into your automation, ready for later steps. Pick an execution mode and a processing type to shape the output.

```mermaid
flowchart LR
  A[Trigger fires] --> B[Automation runs]
  B --> C[Ask an agent team action]
  C --> D[Team output in next steps]
  style A fill:#0d1b2a,color:#7dd3fc,stroke:#38bdf8,stroke-width:1.5px
  style B fill:#2a1225,color:#ff8fa3,stroke:#ff2d60,stroke-width:2px
  style C fill:#2a1225,color:#ff8fa3,stroke:#ff2d60,stroke-width:2px
  style D fill:#0d1a0e,color:#86efac,stroke:#22c55e,stroke-width:1.5px
```

> 🧬 **Connect the dots:** An Agent Team puts several AI agents on one prompt together, right inside an automation. Each agent can run on a different model, picked from **frontier models from top AI labs**. Each agent can also use [custom tools](https://www.taskade.com/learn/automation/agent-tool) and pull from [project knowledge](https://www.taskade.com/learn/automation/agent-knowledge). See [Custom AI Agents / GPTs](https://www.taskade.com/blog/custom-ai-agents-gpts) to go deeper.

![](https://www.taskade.com/static_images/learn/automation/agent-team-44068535.png)

---

### How to Use

1. Click **Add step** and choose the **Ask an agent team** step.
2. Select the **Agent Team** you want to use.
3. Select the **Execution Mode.**
4. Enter the input you want the Agent Team to process.

![](https://www.taskade.com/static_images/learn/automation/agent-team-058a977e.png)
5. Select the **Processing Type**.
6. Enter the input for the Additional Prompt. (Custom Processing Type)
7. Add the appropriate fields to be extracted from output. (Custom Processing Type)

![](https://www.taskade.com/static_images/learn/automation/agent-team-28087627.png)
8. The output from this action can then be referenced in the actions after.

---

### Action Settings

Set up these fields for this action.

| 🏷️ Field | 🔤 Description |
| --- | --- |
| Agent Team | The Agent Team that will process the prompt. |
| Execution Mode | The execution mode the Agent Team will use. |
| Input | The prompt you want the Agent Team to process. |
| Processing Type | Configure what to extract from the conversation. Choose from custom or pre-built types. |
| Additional Prompt (Custom Processing Type) | Give instructions about what you want to be extracted from the Team conversation. |

---

> **Note:** An **Ask an agent team** step runs once and has up to 15 minutes to finish. Taskade does not start it again on its own. If a team task is long, split it into two or more steps.

### Execution Modes

Here are the available Execution Modes in this action. Each one will determine how the Agent Team processes the prompt.

![](https://www.taskade.com/static_images/learn/automation/agent-team-9a84bbbb.png)

| Execution Mode | Function |
| --- | --- |
| Auto | The action picks the single best agent on the team to process your prompt. |
| Everyone | Every agent on the team processes your prompt, one after another. |
| Orchestrate | A lead plans the work, gives each step to an agent on the team, and adds a review step. Use it for multi-stage work. |

---

### Processing Types

Here are the different processing types you can choose from. This will determine what is extracted from the conversation.

#### Summary

This summarizes the AI conversation as if it were writing notes for a teammate who was not present. The AI conversation is categorized into: a general summary, main objective, key suggestions and follow up actions.

![](https://www.taskade.com/static_images/learn/automation/agent-team-9b6964da.png)

#### Lead Qualification

This analyzes and scores sales leads based on their potential. The AI conversation is categorized into: Qualification, Fit Score, Budget Match, Timeline, and Next action.

![](https://www.taskade.com/static_images/learn/automation/agent-team-23468b31.png)

#### Custom

Add your own prompt to give instructions about what you want to be extracted from the Team conversation. Add the appropriate fields that you want to be extracted.

![](https://www.taskade.com/static_images/learn/automation/agent-team-1748c744.png)

### Related guides

- [AI Teams](https://www.taskade.com/learn/agents/ai-teams). Create the teams this action calls
- [Team Generator](https://www.taskade.com/learn/agents/team-generator). Text-to-team in one step
- [Multi-Agents](https://www.taskade.com/learn/agents/multi-agents). Multi-agent collaboration patterns
- [Agent Action](https://www.taskade.com/learn/automation/agent-action). Single-agent version
- [Structured Output](https://www.taskade.com/learn/automation/structured-output). Typed team output
- [Branch Action](https://www.taskade.com/learn/automation/branch-action). Route on team response

---
