Skills

Here is your updated Skills page with Agentic Hyperflow added naturally (no fluff, clean insertion):

---

Skills

Skills define what your AI Agent can do.

They connect your AI Agent to external systems and enable it to perform real actions such as creating incidents, updating records, retrieving data, or triggering workflows.

Without skills, an AI Agent can only respond conversationally. With skills, it becomes an execution layer.

---

How Skills Work

When a user interacts with an AI Agent:

• The AI understands the request

• It determines whether a Skill is required

• It selects the appropriate Skill based on its configuration

• The Skill executes an action in a connected system

• The result is returned to the AI Agent

• The AI responds to the user

This allows AI Agents to move from answering to acting.

---

Add a Skill

Step 1 — Open Skills

Navigate to your AI Agent and go to the Skills tab. Click Add Skill.

---

Step 2 — Select a Library

Choose an Autom Mate Library.

Libraries represent integrations or systems such as:

• ServiceNow

• Xurrent

• Salesforce

• Stripe

• Custom integrations

Each library contains a set of available actions.

---

Step 3 — Select an Action

After selecting a library, choose an Action.

Actions represent specific operations, for example:

• Create Incident

• Add Note

• Create Request

• Get Records

---

Step 4 — Configure Credential

Select or create a Credential to connect the action.

Credentials store authentication details required to interact with the selected system.

The AI Agent will use this credential when executing the action.

---

Step 5 — Configure the Skill

Define how the AI Agent should use this action.

Skill Name

A clear name for the skill.

---

When to Use

This acts as instruction to the AI Agent.

Describe:

• What the action does

• When it should be triggered

• Example user intents

👉 This is one of the most critical fields — it directly impacts agent behavior.

---

Inputs

Define the input parameters required for the action.

Examples:

• incident title

• description

• user ID

• priority

These inputs will be dynamically filled by the AI Agent based on user requests.

---

Run Settings

Define how the action should be executed.

Supervised (Approval Required)

• The action will require approval before execution

• Used for sensitive or high-impact operations

Unsupervised

• The action runs automatically

• Used for safe, repeatable operations

---

Step 6 — Save

Save the configuration to add the skill to your AI Agent.

---

How the AI Agent Uses Skills

The AI Agent does not randomly trigger actions.

It uses:

• The “When to Use” description

• The user’s request

• The available inputs

to decide:

• Whether to call a skill

• Which skill to use

• What data to send

👉 Well-defined skills = accurate automation.

---

Extending Skills with Agentic Hyperflow

If you cannot find the required skill in the available libraries, you can extend your agent using Agentic Hyperflow.

• Open the agent’s Agentic Hyperflow

• Create or modify workflow steps

• Define custom actions

• Map those actions to your AI Agent

This allows you to:

• Build custom integrations

• Create multi-step logic

• Connect to any external system

• Go beyond out-of-the-box capabilities

---

Best Practices

• Keep Skill Names clear and specific

• Write precise When to Use instructions

• Avoid overlapping skills with similar purposes

• Use Supervised mode for risky operations

• Test skills in real scenarios

---

Typical Use Cases

Skills enable AI Agents to:

• Create and manage incidents

• Update tickets and records

• Retrieve system data

• Trigger workflows

• Integrate across multiple systems

---

Next Steps

Once skills are configured, you can expose your AI Agent through channels such as:

• WhatsApp

• Microsoft Teams

• Webhook

• Mate Chat