If you’ve been hearing about HighLevel Managed Agents but haven’t yet built one, this guide will show you a practical place to start.
We’re going to create an AI sales assistant that reviews a new opportunity, examines information already held in your CRM, identifies potential problems and recommends what the salesperson should do next.
It will also save its findings as a note and create a follow-up task.
And here’s the interesting part. HighLevel already provides a template that does this. You don’t have to build the agent from scratch.
I’ve tested the standard template myself in a HighLevel account, without changing its instructions or tools. I’ll explain what happened, including an initial unsuccessful test and what we learned from it.
What Is a HighLevel Managed Agent?
A Managed Agent is an AI-powered assistant that can be given a job to perform using information and tools available within HighLevel.
Traditional HighLevel workflows are excellent when you know exactly what should happen. For example, when a prospect submits a form, create an opportunity, send a confirmation email and notify the salesperson.
A Managed Agent can take things further. Rather than simply moving information from one place to another, it can review information, interpret what it means and recommend or take permitted actions using its instructions and tools.
For more background, read HighLevel Managed Agents vs n8n: From Robotic Automation to AI Assistants.
What We’re Going to Build
Our first agent is called New Deal Brief. It’s one of the ready-made use cases available in HighLevel Managed Agents.
When a new opportunity is created in a selected pipeline, the template is designed to:
- Gather information about the contact from the CRM.
- Review available form answers, conversations, appointments, notes and previous deals.
- Produce a short briefing explaining the opportunity.
- Identify missing information and potential risks.
- Recommend the next step.
- Save the briefing as a CRM note.
- Create a follow-up task for the deal owner.
Think about the practical value. A salesperson receives a new enquiry and gets a concise briefing and suggested next action, without manually searching through several screens.
Step 1: Open Managed Agents in HighLevel
In your HighLevel account, navigate to the Managed Agents area within AI Agents and Agent Studio.
Look for the screen headed Build your own agent. You’ll see a box inviting you to describe what you want your agent to do, followed by ready-made use cases.
At the time of my test, these included Social Media Posting, Lead Research & Scoring, Review Request, New Deal Brief, Email Campaign, Weekly Creative Studio, Prospect Finder, Competitor Watch and Marketing Performance Report.
Select New Deal Brief.
Note: HighLevel may change its interface and available templates over time.
Step 2: Select the New Deal Brief Template
HighLevel describes New Deal Brief as an agent that runs when an opportunity is created in a pipeline you choose. It gathers CRM information, prepares a briefing, saves it as a note and creates a task for the deal owner.
The template uses an Opportunity created trigger.
This matters. The agent is designed to run in response to an event, rather than continually scanning every opportunity in the CRM.
Step 3: Choose the Pipeline
The template is designed to work with selected pipelines. Choose the pipeline where you want new opportunities reviewed, and check the trigger configuration before enabling it for live activity.
For testing, I recommend using a test pipeline or a pipeline where you can safely create test opportunities.
Step 4: Understand What the Agent Can Access
One useful aspect of the template is that it’s designed to work with information already held in HighLevel, subject to the tools and permissions available to the agent.
- Contact details
- Answers collected through forms or quizzes
- Previous conversations
- Appointments
- Notes
- Previous opportunities
- Information about the new deal
The quality of the briefing depends partly on the quality of the CRM data. If your CRM contains little more than a name and email address, the agent has limited information to work with.
But if you’ve been using HighLevel to collect enquiry details, run marketing campaigns and manage conversations, it may have much more useful context.
Step 5: Create a Test Opportunity
Open Opportunities in HighLevel and create a new opportunity in the pipeline you’ve selected. Use a test contact and a recognisable opportunity name, such as TEST – New Website Enquiry.
Save the opportunity, then inspect the agent’s execution activity. You can also use the Managed Agent testing facilities, but there is an important distinction between simulating an event and creating a genuine CRM record.
I discovered this during my own testing.
Step 6: What Happened During My First Test
I used Test Fire with a contact in my own HighLevel account.
The agent ran, but it reported that it couldn’t find the opportunity it had been asked to review. It searched for relevant opportunities associated with the contact, but nothing matched.
So it did not create a briefing note or task.
That was a sensible result. Rather than inventing a pipeline, opportunity value or sales history, the agent explained that it didn’t have enough information to complete the job.
It was still able to identify some information about the contact, including missing details and an email address flagged as invalid. But without a matching opportunity, it stopped short of creating CRM records.
What did we learn?
An agent can execute successfully without completing its intended task. In this case, the missing opportunity prevented the briefing from being produced.
The test didn’t establish why the opportunity was missing. A simulated trigger, a timing issue or an incorrect opportunity reference were all possible explanations.
The lesson: Check the underlying CRM record. Don’t assume that pressing a test button proves the entire process works.
Step 7: The Second Test Worked
For the next Test Fire run, there was a genuine open opportunity associated with the contact.
This time, the agent found the deal and completed its work. I hadn’t changed the template’s instructions or tools. This was the standard New Deal Brief template supplied by HighLevel.
It reviewed the contact’s background
The contact had previously completed a quiz on my website. The saved answers indicated that the prospect:
- Was still planning their business.
- Wanted to generate more leads.
- Had no customers or sales yet.
- Had fewer than 200 contacts.
- Was relying on word of mouth and referrals.
- Used Post-it notes, Excel and MailChimp.
The agent used those answers to build a picture of the prospect’s situation. It also recognised that the quiz answers were stored without their original question labels, so it didn’t pretend to know more than the information established.
It examined the opportunity
The agent identified an opportunity in my MarketerM8 Consulting pipeline, including its stage, source and recorded value.
It checked for previous conversations, appointments, notes, tasks and deals. There wasn’t much history available, and the agent said so.
It spotted potential problems
The briefing identified several points worth checking:
- The email address was flagged as invalid.
- The contact’s company and telephone number were missing.
- Budget and timescale weren’t recorded.
- The prospect appeared to be at a very early stage of business.
- The opportunity had a recorded value of zero.
These observations don’t necessarily mean the lead is unsuitable. They help establish what needs to be clarified.
It recommended a next action
Rather than suggesting an elaborate sales pitch, the agent recommended confirming that the lead was genuine, checking contact details and starting a conversation about the prospect’s plans.
It even drafted a possible opening line based on the quiz answers. Importantly, it did not send that message.
It created a note and a task
Finally, the agent saved the briefing as a CRM note and created a follow-up task assigned to the deal owner.
In my test, the task was scheduled for the next business day. The agent explained its scheduling decision in the execution result.
That’s where the example becomes particularly interesting. The agent didn’t merely summarise information. It used CRM data, made recommendations and completed two administrative actions.
Step 8: Check the Results in Your CRM
After running your test, open the relevant contact record and check the Notes and Tasks areas. Confirm that the briefing and follow-up task actually exist.
Then compare the agent’s statements with the information recorded in the CRM.
- Has it found the correct opportunity?
- Has it interpreted the contact’s information accurately?
- Has it distinguished recorded facts from missing information?
- Is the recommended next action sensible?
- Has it created the note and task against the intended records?
- Is the task assigned to the correct person?
- Is the due date appropriate?
Don’t judge an agent solely by how convincing its written response sounds. A plausible-looking briefing isn’t necessarily an accurate briefing.
In my successful test, the agent found the only open deal for the contact, even though the test trigger used a generic opportunity name. That worked here, but it doesn’t prove the agent will always select the correct deal when several exist.
I’d test that situation before relying on the agent across a busy sales pipeline.
Step 9: Should You Let the Agent Create Tasks Automatically?
Automatic task creation is part of the standard template. That’s useful, but it deserves some thought.
If you create 50 new opportunities in a day, do you want 50 new tasks? Possibly. But you should first decide how those tasks fit into your existing sales process.
Consider who owns opportunities, how tasks are assigned, what due dates are appropriate and whether other workflows already create similar follow-up tasks.
Otherwise, you could end up with duplication.
Just because you can automate an action doesn’t mean you should automate it without considering the wider process.
Start with a small number of test records, inspect the results and then decide how you want the agent to operate in everyday use.
Step 10: What Could You Build Next?
Once you’ve successfully tested the New Deal Brief agent, you can think about more advanced applications.
For example, you might want an agent that reviews opportunities that haven’t progressed recently, identifies missing information or prepares a briefing before a sales appointment.
Or you might want an agent that reviews a quotation and suggests an appropriate follow-up based on previous conversations.
These are possible next projects, not functions demonstrated in this tutorial. Each would need its own tools, instructions, testing and safeguards.
For more ideas, see 10 Practical HighLevel Managed Agent Ideas for Small Businesses.
What This Tells Us About Managed Agents
I’ve been working with marketing automation since 2009, and one thing I’ve learned is that the technology is rarely the biggest challenge.
The real challenge is deciding what you want the system to achieve and making sure it has reliable information to work with.
That’s still true with AI agents.
What impressed me about this test wasn’t simply that the agent could produce a well-written briefing. It reviewed several pieces of CRM information, recognised what was missing, suggested a sensible next action and completed administrative work without me writing a complicated set of instructions.
The standard template was enough to get started.
That’s encouraging for small businesses that don’t have the time or technical resources to build complicated AI systems.
However, I wouldn’t switch an agent on across every sales pipeline without testing its behaviour first. Start small, check the output and make sure it fits your business process.
Do You Need the Advanced HighLevel Features?
Managed Agents are part of HighLevel’s expanding set of advanced capabilities, and access depends on the account and plan you’re using.
At MarketerM8, our Professional plan at Β£197 + VAT per month includes the advanced HighLevel features available through our Founder Agency offering, including Managed Agents.
Our Starter plan at Β£97 + VAT per month is the Lite option and does not include Managed Agents.
Both plans include unlimited contacts and users.
You can find out more at MarketerM8. Check applicable AI usage charges before enabling agents for routine use.
Want Help Building Your First HighLevel Managed Agent?
If you’re interested in using HighLevel Managed Agents but aren’t sure where to start, I can help.
I’m a UK-based HighLevel consultant with more than 25 years’ experience in direct-response marketing and hands-on marketing automation experience dating back to 2009.
My approach is to start with the business problem, then work out whether a workflow, an AI agent or a combination of the two is the most sensible solution.
Book a 20-minute discovery call
You can also explore our HighLevel workshops for practical, hands-on training.
Frequently Asked Questions
Do I need coding skills to build a HighLevel Managed Agent?
Not necessarily. HighLevel provides ready-made Managed Agent templates that can be used as starting points. More advanced agents involving specialist integrations or custom tools may require additional technical knowledge.
Can a Managed Agent create tasks in HighLevel?
Yes. In my test, the standard New Deal Brief template successfully created a follow-up task and assigned it to the deal owner. I verified the task in the CRM.
Can a Managed Agent read contact information?
The New Deal Brief template is designed to retrieve relevant CRM information. In my test, it successfully read contact details, quiz answers, opportunity information and available activity history.
What happens if the agent cannot find the opportunity?
In my first test, the agent couldn’t find the opportunity and did not create a note or task. It explained what information was missing rather than inventing a briefing.
Does the New Deal Brief agent automatically email the prospect?
It didn’t in my test. The template prepared a suggested opening message but did not send it to the contact.
Should I use a Managed Agent instead of a HighLevel workflow?
Not always. Conventional workflows are usually preferable for predictable, rule-based actions. Managed Agents are more interesting when the task involves interpreting information, assessing context or preparing a recommendation.
For a deeper comparison, read HighLevel Managed Agents vs n8n: From Robotic Automation to AI Assistants.
About the Author
Julian Mills
HighLevel Consultant & Marketing Automation Strategist
Julian Mills helps business owners automate lead generation, sales follow-up, customer communication and business processes using HighLevel and MarketerM8. Since 2009 he has helped businesses implement CRM, marketing automation and AI-powered systems that save time, improve customer experience and generate more sales.
About Julian
Book a Discovery Call
Join a Free HighLevel Workshop
Learn About HighLevel with MarketerM8



