Setting Up AI Marketing Agents

AI Marketing Agents automate repetitive tasks like lead routing, follow-ups, and workflow creation using natural language instructions. Each agent operates autonomously once configured.

Available agent types

Lead Routing Agent — Distributes leads to sales reps based on territory, product interest, or company attributes. Handles round-robin, account-based, and custom routing logic.

Follow-up Agent — Sends personalized emails to leads based on form responses or engagement. Adjusts messaging and timing based on lead score and behavior.

Ad Operations Agent — Manages ad campaigns by updating audiences, pausing underperforming ads, and creating lookalike segments from converting leads.

CRO Agent — Runs A/B tests on forms and landing pages, analyzes results, and automatically implements winning variations.

Reporting Agent — Generates weekly summaries of pipeline metrics, form performance, and conversion trends. Sends reports to Slack or email.

Create an agent

1. Navigate to AI Agents in your dashboard

2. Click Create Agent

3. Choose the agent type that matches your goal

4. Describe what you want the agent to do in plain English

5. Set guardrails like approval thresholds or excluded segments

6. Enable the agent to start automation

Natural language configuration

Agents understand instructions like:

• "Send a follow-up email 10 minutes after form submission if the lead score is above 70"

• "Assign leads from healthcare companies to the healthcare sales team"

• "Pause Facebook ads when cost per lead exceeds $150"

• "Run an A/B test on the pricing page form with a shorter version"

• "Send a Slack summary every Monday with the previous week's demo bookings"

The agent translates your instruction into executable workflows.

Set guardrails and approvals

Control what agents can do autonomously:

Approval required. Make agents ask before taking high-impact actions like pausing campaigns or assigning leads above a certain value.

Spending limits. For ad operations, set daily or weekly budget caps the agent cannot exceed.

Excluded segments. Prevent agents from acting on VIP accounts or specific customer types.

Test mode. Run agents in simulation mode first to preview actions before enabling full automation.

Monitor agent activity

Each agent maintains an activity log:

• Actions taken and when

• Decisions made and the reasoning

• Performance metrics like time saved or conversion impact

• Errors or edge cases requiring attention

Review logs weekly to verify agents are meeting your goals and adjust instructions as needed.

Connect agents to tools

Agents integrate with your existing stack:

• CRM — Create, update, and assign records in HubSpot or Salesforce

• Email — Send automated follow-ups via your connected email account

• Slack — Post notifications, summaries, or alerts to channels

• Ads platforms — Update audiences and campaigns in Google Ads, Facebook, LinkedIn

• Calendars — Book meetings using Calendly or Chili Piper links

Authorize integrations in Settings > Integrations before creating agents that need them.