Home/HighLevel/How to/Set up Managed Agents
How to · AI agents
How to set up Managed Agents in GoHighLevel. Describe the job, it builds the agent.
Managed Agents in GoHighLevel are AI agents you build by describing the work in plain language. They can use knowledge bases, take CRM actions, run on schedules or triggers, remember context and connect to outside tools through MCP connectors.
Describe the job, HighLevel builds the agent.
The short answer
To set up Managed Agents in GoHighLevel, open AI Agents, Agent Studio and the Managed Agents tab, click Create Agent, describe the agent's goal, channels, knowledge bases and allowed CRM actions or pick a prebuilt use case, then review the draft, refine it in the chat sidebar, set triggers or schedules, test it and publish. Give each agent one job.
Michael is a sponsored HighLevel affiliate. Links to HighLevel use his partner link at no extra cost to you. HighLevel prices and product facts on this page come from HighLevel's own pricing page and help center, checked on the date above.
Build
How do you set up Managed Agents in HighLevel?
- In the subaccount, click AI Agents, then Agent Studio, then the Managed Agents tab.
- Click Create Agent.
- Describe the goal, channels, knowledge bases and the CRM actions it may take, or choose a prebuilt use case.
- Review the generated agent and refine tone and rules in the chat editing sidebar.
- Set what starts it: a CRM trigger or a schedule.
- Connect tools if needed, such as MCP connectors.
- Test with real examples, then publish.
Ideas
Jobs Managed Agents do well.
| Agent | What it does |
|---|---|
| Lead researcher | Looks up new leads and adds notes |
| Follow up writer | Drafts follow up based on the conversation |
| Daily summary | Summarizes yesterday's pipeline every morning |
| Data hygiene | Flags incomplete or duplicate records |
Before you start
What do you need before you set up Managed Agents?
An agent that can change records needs clear limits. Before you create one in HighLevel (GoHighLevel), write these down.
- One job for the agent, in a sentence
- The CRM actions it may take, and the ones it may not
- A knowledge base if it needs business facts
- What starts it: a CRM trigger or a schedule
- An understanding of how runs are billed. See AI usage costs
Test it
How do you test a Managed Agent?
Test on records you can afford to get wrong.
- Run the agent on a test contact or a small sample.
- Read what it produced, line by line.
- Check every record it changed.
- Give it an awkward example, such as a contact with missing details.
- Let a scheduled agent run once and review the result before you leave it on.
Common mistakes
What mistakes should you avoid when you set up Managed Agents?
- One agent with five jobs. Give each agent one job. Small agents are easier to trust.
- No limits on what it can change. State the allowed CRM actions in the description.
- Publishing the first draft. Refine tone and rules in the chat sidebar first.
- Forgetting usage. Runs count toward your AI plan. Schedule only what earns its place.
- No review habit. Check its output each week at first.

Give each agent one job and clear limits on what it may change. Small agents are easier to trust.
Questions
Common questions.
What are Managed Agents?
Managed Agents are AI agents in Agent Studio that you build by describing the work in plain language. They can use knowledge bases, take CRM actions, run on schedules or triggers, remember context and connect to outside tools through MCP connectors. Give each agent one job, because small agents are easier to trust.
Where do I find them?
In the subaccount, click AI Agents, then Agent Studio, then the Managed Agents tab. Click Create Agent and describe the goal, channels, knowledge bases and the CRM actions it may take, or choose a prebuilt use case. You can open Agent Studio and build a first agent during a 30 day GoHighLevel trial through Michael's link.
Can they run on a schedule?
Yes. When you set what starts the agent, choose a CRM trigger or a schedule. A daily summary agent, for example, can summarize yesterday's pipeline every morning. Let a scheduled agent run once and review the result before you leave it on, and schedule only what earns its place, because runs count toward your AI plan.
Can they use outside tools?
Yes. Managed Agents connect to outside tools through MCP connectors, which you add when you set the agent up. State the CRM actions the agent may take, and the ones it may not, in the description. Test on a small sample and check every record it changed before you publish.
What is the difference between Managed Agents and Conversation AI?
Conversation AI replies to customer messages. Managed Agents do work you describe, such as research, summaries or data checks, on a trigger or a schedule. One talks with customers, the other works in the background on your records. You can use both for different jobs.
What jobs suit a Managed Agent?
Researching new leads and adding notes, drafting follow up based on the conversation, summarizing the pipeline each morning and flagging incomplete or duplicate records. Each of those is one clear job, which is what a Managed Agent handles best. Write the job in a sentence before you create the agent.
Where can I learn more?
See the Managed Agents complete guide, the Agent Studio complete guide and HighLevel AI agents. Managed Agent runs count toward your AI plan, so read up on AI usage costs before you schedule anything. If AI is the reason you are considering the platform, read whether GoHighLevel is worth it.
Next step
See it in your own account. Thirty days, every feature.
Start a trial, build the first system and test it before the first invoice. Usage fees for phone, SMS, email and AI can apply during the trial.
HighLevel has three plans at a flat monthly price, each with unlimited users and contacts. See what each one includes and what it costs.
Compare the PlansStarter, Unlimited and Founder