> For the complete documentation index, see [llms.txt](https://converso.gitbook.io/product-documentation/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://converso.gitbook.io/product-documentation/ai-agents/agent-set-up.md).

# Agent set up

### Deploy

To set up your agent, create a new workspace and click on the Agents tab. You will then be able to choose an agent template from the agent store.

Once you have deployed your agent, give it a name and a description (the description is for your internal reference). \
\
It's presently only possible to deploy one agent per workspace. If you want to change your agent, you can remove it from your workspace, and select another template from the agent store. 

Your agent also shows here its activation state - this is set in the configure tab, against a specific configuration (see below).\ <br>

### Configure

Once you have deployed, you can start configurating your agent, personalising it to your brand, business and tone.&#x20;

You can create multiple configurations via the "+ New config" button, and then give that configuration a name.

#### Activate

You can test which config produces the optimum responses from your agent via the Test section (see below).&#x20;

Once you have decided which configuration works best, you can activate the agent via clicking the activate button alongside the configuration name.&#x20;

<figure><img src="/files/Sgx1r959yNDlhfAAFUoX" alt="" width="273"><figcaption><p>Activate selected configuration</p></figcaption></figure>

Once the agent is activated it is then "on" and live, and will answer inbound messages from your customers over the channels you have connected to.

When activated, the agent will be added as a new user to that workspace's team, making it available like a human agent to be assigned to conversations to respond to customers' questions. &#x20;

You can deactivate the agent and turn it "off" through the same setting. The agent will no longer answer customer messages until activated again.

The following describe the different configuration settings:

#### Business name

The name of your business as it will be referenced by the AI agent during conversations. This ensures the agent identifies itself correctly and represents your brand consistently across all customer interactions.

#### Product type

Defines the product or service category the AI agent is supporting. This helps the agent contextualise its responses and stay relevant to your specific offering when answering customer queries.

#### Tone

Sets the overall conversational style the agent uses when responding. Options like "Formal" produce professional, polished language, while more casual settings create a friendlier, conversational feel better suited to informal channels.

#### Language variant

Selects the language the AI agent will use when communicating with customers. Choose the variant that matches your customer base to ensure responses feel natural and locally appropriate (presently English is only available).

#### Empathy level

Adjusts how emotionally attuned the agent's tone is when responding to customers. Set to "High" for warm, understanding language that acknowledges the customer's situation, or lower it for a more matter-of-fact, efficient style.

#### Max sentences

Sets the maximum number of sentences the AI agent will use in a single reply. Keep this low for concise, chat-style responses, or increase it when your use case requires more detailed explanations.

#### Technical level

Controls how much technical detail and terminology the agent includes in its responses. "High" suits audiences comfortable with domain-specific language, while lower settings produce simpler, more accessible explanations.

#### Support team reference

Defines the name or label the AI agent uses when referring to your human support team. For example, entering "our Customer Care team" means the agent will use that phrase when offering to escalate or hand over a conversation.

You can then save the configuration of your agent, and proceed to test its responses in the Test area (see below).

#### URL

URL is your business website address. Your agent uses it to generate a short summary of what your business does, so it can answer broad questions like "what do you do?" directly and consistently without searching your knowledge base. &#x20;

### Test

Test is a safe environment where you can evaluate your AI agent’s responses using different configurations. Switch between configurations to compare their performance and determine which best meets your priorities.

You can also activate the agent using your preferred configuration, just as you can from the Configure section.\
\ <br>