Five connected AI agent tasks surrounding a CRM contact review panel in teal and copper

5 AI Managed Agents for Follow-Up and Better Leads

In AI, Automation, GoHighLevel by KathrineFLeave a Comment

You can spend months improving lead generation and still overlook the people already sitting in your CRM. Someone downloaded a resource, asked a question, or booked a consultation. Then the next step got buried under everything else you had to do.

GoHighLevel Managed Agents are AI-powered tools that can help automate internal business tasks, from identifying missed follow-ups to preparing consultation research and drafting marketing content.

In this tutorial, Marketecs founder Kathrine Farris walks through five Managed Agents she built and tested inside the real Marketecs account, including a follow-up agent that identified eight suggested messages. Three immediately reminded her of people worth contacting.

That result is a useful starting point: better visibility and relevant recommendations. It doesn’t mean the agent generated three sales. Kathrine still decides which opportunities deserve attention and what gets sent.

Kathrine has since continued experimenting with Managed Agents for content creation, marketing reporting, and SEO. You can see those additional examples in our follow-up video, but we’ll start here with the first five agents and what we learned from testing them.

What You’ll Learn About GoHighLevel Managed Agents

  • How to describe an agent’s job and review the configuration it creates.
  • Five uses for CRM follow-up, content, meeting preparation, contact review, and lead prioritization.
  • Why approval, delivery instructions, and testing matter as much as the initial prompt.

Start your free 30-day HighLevel trial to explore the platform. Marketecs may earn a commission if you sign up through our link, at no additional cost to you.

Video Time Stamps

What Are GoHighLevel Managed Agents and How Do They Work?

GoHighLevel Managed Agents are AI agents you can configure to complete specific business tasks using instructions, connected tools, and available account information. Depending on their setup, they can run on demand, on a schedule, or in response to a trigger.

A Managed Agent is useful when you can explain the job, identify the information it needs, and recognize a good result. “Help with sales” is too vague. “Find recent leads who have not replied for three days and draft a message for my approval” gives it a concrete assignment.

Our tutorial starts in AI Agents and Agent Studio, where the builder can create a Managed Agent through conversation. The resulting setup includes instructions, triggers, connected tools, and optional knowledge bases. Review each part rather than assuming the builder chose the schedule or permissions you intended.

These examples focus on internal business work. They serve a different purpose from the voice and conversation agents that answer customers directly.

Update: Managed Agents have usage limits and billing considerations. HighLevel currently lists pay-per-use token pricing and monthly run allowances within its AI Employee plans. Check HighLevel’s current AI pricing and your account’s available features before choosing how often an agent should run.

1. Follow Up With Leads Who Stopped Responding

The first agent tackles an ordinary sales problem: a lead takes action, a conversation starts, and nobody follows up when the reply stops.

Define the Review and the Exceptions

In the video, the starting request is to follow up with leads who have not replied in three days. The builder proposes a daily schedule at 9 a.m., connects LeadConnector, and creates a decision process. Kathrine then adds an essential instruction: send communications to her for approval before sending anything to a lead.

A practical starting brief for your own account would include:

  • The contact history and time window to review.
  • Exclusions for opt-outs, team members, clients, spam, and people who already booked.
  • A check against duplicate follow-up messages.
  • A draft message and a reason for recommending each contact.
  • The person who approves it and where that person receives the notification.

Test Agent Recommendations and Approval Notifications

The manual test reviewed recent contacts over roughly 30 days. It returned eight suggested follow-ups and excluded several types of contacts that did not belong in the list. Three recommendations stood out to Kathrine as people she should revisit.

The test also exposed a gap: the recommendations were waiting inside the agent. Requiring approval did not automatically mean an email would arrive. She added an instruction to email her when recommendations needed approval. Run another test after that change to confirm that both the approval hold and the notification work.

2. Draft Social Media Content for Review

The social agent creates seven posts across seven days, starting a week after the run date. That leaves time for the social media manager to review the copy and visuals before anything goes live.

This agent runs on demand in the demonstration. A scheduled version could prepare next week’s drafts every Monday. Another possible use is a project-submission form: a team member adds photos and project notes, and an agent prepares a post for review.

During Kathrine’s test, image-folder placement, consistent logo application, and access to other saved brand resources did not behave as she wanted. Those are reasons to test your own media process carefully. They do not establish a permanent limitation for every account. Check the resulting image, watermark space, claims, tone, and scheduling status before approving a post.

3. Prepare for Consultations Before They Happen

A consultation-preparation agent can do the first pass of research you normally squeeze in before a call. The request in the video asks for an email two days before an appointment, covering the business, its services, its audience, its history, and likely areas of interest.

The builder asks which calendars count as consultations, which is an important detail. A prospect consultation and an existing client’s routine check-in may need very different preparation.

Connect the relevant business knowledge base and make web research available where needed. Ask the agent to distinguish what it found from what it inferred, and check the source website yourself before relying on the brief. For the timing test, use an upcoming appointment on the selected calendar. A manual run against older meetings can test the research output, but it does not prove that the pre-meeting schedule fires correctly.

4. Identify Potential Spam Contacts in Your CRM

The spam-review agent examines a newly created contact and applies a potential-spam tag when the record looks questionable. Tagging makes the result easy to filter and review without immediately deleting a record.

Kathrine improves the instructions after a test by explaining her actual data rules: her forms require a name and email, and certain existing tags identify people who should be treated as legitimate contacts. Those details help the agent interpret her CRM.

Your criteria may be different. A legitimate first-time phone caller may have no email on file. Missing information is a reason to investigate, not universal proof of spam. Also keep contact validity separate from communication consent: a known customer who opted out should still have that preference respected.

Start with a review tag and compare the agent’s decisions with real records before allowing any further action. Our CRM spam-filter automation tutorial covers a complementary workflow for form and message spam.

5. Prioritize High-Scoring Leads Against Your Ideal Client Profile

The “boiling lead advisor” builds on an existing lead-scoring process. Marketecs uses stages such as cold, lukewarm, warm, hot, and boiling to organize engagement. The agent reviews contacts reaching the boiling stage, compares them with the ideal client profile, and explains who may deserve a direct conversation.

The useful distinction is fit. Someone can be very active without being the right client for your services. A knowledge base describing your offers and ideal clients gives the agent more context than a score alone.

The request calls for a weekly summary. After testing, Kathrine adds “email” explicitly because the first result stayed inside the agent. Review the pipeline stage, event trigger, scheduled interval, recipient, and account time zone. Then test which contacts appear in the report and when it arrives. A generated configuration still needs that check.

More GoHighLevel Managed Agent Examples

After building these five agents, Kathrine continued testing how Managed Agents could support other parts of our marketing work.

In a follow-up video, we share five more GoHighLevel Managed Agents, including examples for content repurposing, newsletter drafting, marketing performance reporting, and SEO, AEO, and GEO content analysis.

Those examples build on the same approach covered here: start with a specific task, give the agent useful context, and test the results before relying on them.

Choose One Agent and Give It a Clear Finish Line

You do not need five agents to get started. Pick the task you repeatedly postpone and define the output you would actually use: three researched contacts, a meeting brief, or a reviewed batch of social drafts.

For each test, check the selected records, exclusions, factual accuracy, delivery method, and approval behavior. Refine the instructions when a result misses the mark. Once the task works reliably, decide whether it should remain on demand or run on a schedule.

GoHighLevel Managed Agents FAQs

Do I need to write code to build these AI agents?

The examples use conversational instructions inside HighLevel’s builder. You still need to understand the task well enough to review the tools, triggers, instructions, and results it creates.

Can an AI agent send lead follow-ups automatically?

The workflow in this demonstration requires approval first. Keep that hold in your instructions and test that no lead message leaves before approval. A notification to the reviewer is a separate requirement.

Which AI agent should a service business build first?

A consultation brief or a follow-up recommendation list is a practical starting point because the output is easy to inspect. Choose the task that addresses a real gap in your current process.

Are Managed Agents included without additional usage costs?

Do not assume that. Check your enabled features, plan allowance, and current AI usage pricing before running an agent frequently, especially across a large contact list.

Key Takeaways on GoHighLevel Managed Agents

The strongest use case is the one that makes your next business action clearer. Start with a specific task, supply the right CRM and business context, and test how the result reaches the person responsible. Human review turns a promising recommendation into a deliberate next step.

Read the full transcript

Every business owner would like a crystal ball that tells them exactly who to reach out to, when to reach out to them, and who the best fit is. And today, I’m going to show you exactly how to find it I’ve been working with small business owners, entrepreneurs, and thought leaders for 17 years now And I can tell you that no matter what system I’ve ever worked with, one of the biggest questions is always return on investment or how do I actually make money at this? Business owners are constantly running promotions, webinars, events, creating lead magnets, improving their search engine optimization, running Google Ads or Facebook Ads.

There’s all kinds of marketing methods to bring in new leads, to bring in people that might want to hire them. But then the question becomes, great, so I know my cost per click, I know my cost per lead, I’m bringing in potential new customers, but for service-based businesses, it’s how does that convert? How do I actually turn that into paying customers? Because marketing is just part of the puzzle. Sales is the other part. Today, what I wanna show you is how our favorite CRM actually has a solution for this, and it’s extremely affordable, very easy to use, and fully built in, no additional services required.

My name is Kathrine Farris. I am the founder of Marketecs and Marketecs Engine, and I’ve been helping small businesses since 2009 with their marketing. And most recently, in addition to our services, we started using a very specific platform to accomplish these goals, and that platform is our favorite CRM, HighLevel or GoHighLevel. We are an affiliate. I do have a link down below. If you’ve never heard of it, you can get a 30-day free trial and follow along with exactly what I’m gonna show you here today, but if you are already using it, I do have a consultation.

If you need help with it, certainly book an appointment. But in addition to actually being an affiliate, we built all of our systems, all of our processes out in HighLevel, and we resell that as Marketecs Engine, and there’s a lot more that goes into it than just that. I’m very much oversimplifying this. So if you have questions, again, book a consult down in the description below. But let’s get into the meat and potatoes of today’s topic, which is actually finding the gold in your CRM, finding the contacts that matter, finding that low-hanging fruit so at the end of the day, you can make money.

Now, if you have been around any of my content before, whether it be LinkedIn, my blog posts, or other YouTube videos, you know that I’ve talked a lot about lead magnets and lead generation, as well as lead scoring, and all of those are basic features within HighLevel. But recently, HighLevel has been incorporating a lot more AI, and this is not connecting outside programs. This is built-in, baked-in AI inside of HighLevel, which is what I like about it, is that there’s no tech to set up, and there’s no extra steps to take.

And I did a full walkthrough of all the different places where AI was available just earlier this year, and at that point, I covered over 10 different items. But today I have two newer areas which are actually way more interesting than the other 10 And one in particular that I really wanna dive a little bit deeper in today and show you how to build some awesome resources that go way beyond just creating content. So let’s jump in and take a look. All right, so here we are inside of HighLevel, and we are in our account right now, our Marketecs account, and I’m not even showing you a demo account in this instance like I normally do because these are actual agents that I’m gonna show you that I’ve built myself for my business.

And yes, they’re actually agents built using AI inside of HighLevel. That’s the exciting part. But before I dive into that, I just wanna touch really briefly on some of the places where you can find AI. So the first thing we see here on the left is Ask AI. Ask AI is like having ChatGPT or Claude right in your CRM, where you can ask it anything about your system, about contacts, about recent campaigns. You can ask it to give you summaries of things, and it has full access to your system. It also has access to your brand guidelines and your brand boards if you built those out.

So those are very helpful and important as you start working with AI inside of HighLevel. You also have AI showing up in places like content AI in the social planner. Under emails, you’re going to have email AI showing up when you go into create a campaign. Under prospecting, and we covered a whole prospecting piece as well, you will have prospect AI shows up in the system here. Then under workflows, you can build workflows with AI as well, and you can create here, building using AI. Under sites, you can build websites and funnels using AI.

There is AI under blogs. There is AI basically all over the system. Under reputation management, you have an AI feature built in there as well. Like I said, I will link to the video where I dig a little bit deeper into each one of those, and you can see the ins and outs of them, and they’ve only improved since I did that last video. And I also wanna show you that there is basically now a replacement for vibe coding that also lives here in HighLevel. I did do a little test a few months ago where I compared the AI studio vibe coder against a very popular option called Lovable.

I also compared using ChatGPT versus Claude for getting your prompts. And again, these have really gotten a lot better, the visibility, the search engine optimization, all the backend stuff has gotten a lot better on these. So one thing I would mention is if you’ve already explored the vibe coding here, but you haven’t gone into your settings in the last few months, go into settings, and you’ll notice migrate to SSR right here. This is where you can uplevel the project that you already built, and it will do it for you in a matter of fifteen, twenty minutes at most and just give you all the improvements that they’ve seen over the last little bit.

But it is an action that you have to take. There’s my overview of a lot of the different places where AI shows up inside of HighLevel But I haven’t hit on the most important one yet that I’ve really been enjoying playing with, and this is what they were calling When they first came out, they were calling these super agents. And you’ll notice here in AI agents, there are voice agents, conversation agents. I have separate videos on all of those. Again, we have content AI in here, but Agent Studio, which was originally called Super Agents, these are the ones that I wanted to point out today.

These are the real focus of today’s tutorial. And what I’m gonna do is I’m gonna go and I’m gonna create a managed agent, but then I’m gonna come back and show you these ones that I’ve created, what went into the prompting, and even how things have changed and improved since the first one I built in August to the ones that I’ve built more recently. You can put these in folders. I’m a big fan of folders. You can see that they are in beta, so if you do run into any bugs, just keep that in mind that they are still experimenting and improving and updating these.

They’re not considering these as being wrapped up in a pretty little package quite yet, although I’d say they’re pretty darn close. All right. When I go to create a managed agent, they give you a couple ideas here, which is really great because one of the biggest difficulties around AI and around a new system is the use cases. Okay, there’s this brand new thing, but aha, what do I do with it? What … How can I actually use this in my business? So there’s, you know, a knowledge base assistant example, a community moderator example, email campaign, and social media posting.

I feel like the social media posting is and was the number one go-to for most people when they would come in here and start wanting to build one of these managed agents. It was like, “Oh, great, I can get it to post on my social media account every day. It will actually draft the post, draft the image, and schedule the post.” And so I also did this as my first agent, and I will show you that. But what I realized is something that I think a lot of people are coming back to right now is that AI created content always needs that human touch.

You cannot just set it on autopilot, and so that makes this okay, maybe not fantastic yet. The good thing is, is that even with the social media posting, you can actually set it to put everything in review, and that’s how I’ve set mine up. So every post goes automatically on the calendar, but it’s in review for a human to edit, look at, edit, and then actually schedule. So it’s been helpful, but the other agents I’ve built have actually been way better. Uh, but of course, I would encourage you to play around with some of these starting use case.

Like they said, they’re pre-wired tools, and you can change and edit anything after you pick what you want. The other thing you can do is you can work here within the AI builder to actually build your agent through conversation. This is just like talking with [unclear] or with Claude, and they give you a couple use cases for this as well This one here, build me an agent that follows up with leads who haven’t replied in three days. This, I do not have anything doing this yet, so I’m just gonna click that.

Super simple. I could have typed in anything that I wanted, and I’m gonna let this build here, and you can actually see what it’s doing. It’s gonna describe all the things that it’s doing here. And then I’m gonna show you the options for exactly what you can do, ’cause you’ll notice here it, it’s even pointing at what it’s doing. It’s setting up the triggers, setting up apps. You know, you can turn on capabilities. All right, so daily no reply follow up. So this is going to be set at an interval, so they’ve chosen a schedule, and it’s gonna be every day starting tomorrow at 9:00 AM.

I mean, the system picked this. I didn’t put these things in here. Never ending, and that’s pretty smart right there just out of the gate. Uh, the apps, it’s using LeadConnector. That’s what that little symbol is there. But if you wanted to connect other apps in here, you could. At this point, I have nothing else connected, so just know that. You could actually make this much smarter if you wanted to connect these other apps via MCP. I only have HighLevel or LeadConnector connected right now. I do have web search capability on.

There’s no need to have image generation, audio or video with this, but it’s pretty cool that you can actually create those items And then you can add custom APIs. All right, so the other thing I can connect is my knowledge base, and this might be something, my Marketecs knowledge base, all about our programs and stuff makes sense to connect here because we’re drafting emails, we’re drafting replies. This says that it’s going to draft replies for all of this stuff, and you can see how it’s going through. Never contact someone who opted out or asked not to be contacted.

All right, decision flow. Review the leads and activities. Identify anyone who did not reply in three days. Exclude anyone that should not receive follow-up. Do not create duplicate follow-ups. It actually included a lot of information here, but I’m a bit of a control freak when it comes to my list and it doesn’t actually say what it’s gonna do with this follow-up. So I feel like there’s just a piece missing here where it doesn’t say whether it’s actually gonna send an email or whether it’s gonna send it to me, and I would prefer that it sends it to me, so that’s what I’m gonna tell it to do.

I’m gonna tell it to send all communications to me before approval, before sending. And so now it’s going to update the instructions, and it does identify me. It will actually pull out my contact ID in the system, it will know who I am, and it will pull my email in. I don’t even have to tell it what email address to use. I have tested this. This has worked. All right, so now it says it is going to require my approval before any lead communication is sent, which is great, so I’m still wondering how it’s going to tell me that I have that information, but I tell you what, I’m gonna publish, and I’m gonna go back out to the agent, because this is the agent editor.

So now when you go out and you’re actually looking at the agent, I’m in the lead follow-up agent. This is set to run every day at nine AM, but I kind of wanna run a test. Now, there was a test option right in here. You can test the agent here, but I’ve also really liked the option of just saying Run and seeing what it does, and this is where it will actually go through the process and look for anything that fits the description of what it’s supposed to do. Now, sometimes it might just come back and say there’s nothing there for it to do.

Other times, you could say, “Run this on the last 10 contacts,” because you know that maybe there wasn’t someone in there recently. So I’m just gonna let this run for a minute and see what happens. Okay. The agent is done running, and I just wanna draw your attention to Awaiting your approval right here, that eight follow-up messages are ready to send, but none will actually go until you approve. And at the moment, it’s actually looks like it’s just storing it here in the agent, and that’s what I was curious about.

I am gonna instruct this to actually send me an email when it has someone that needs a follow-up message approved. But before that, I just wanna say that I watched this go through, and like I said, I’m in a live account, so I’m not gonna show you all the nitty-gritty details, but it excluded anyone that it recognized as a current client, a team member, as a spam contact. It identified all of these things. It spent a few minutes going through recent contacts. Basically, it took over the last 30 days because I did this as an independent run and not that it was being triggered by someone actually not communicating in three days.

And that’s why I like using that run feature in these agents. I’ve actually found that that works really well for a nice test here. And it recognized people that had already booked appointments, like I said, spam or opted out contacts. It recognized people that were past clients or current clients. It pulled out team members. And the eight that it told me, “Hey, you might actually want to follow up with these people,” you know, they took some sort of action and then went cold, they’re actually solid recommendations. Three of them I actually looked at and went, “Oh, shoot.

Yeah, I should follow up with that person.” So this was actually … This was a great test, and now I’m just gonna go in and I’m gonna edit my agent here to tell it to email me when it has recommendations. So I’m just telling it to send me an email when you have recommendations for me to approve, and it will add that to the instructions. And not everybody might need that. You, if you really want to come in and check these agents every day and you wanna keep your inbox clean, that’s perfectly fine.

You can do that, and just come in and look at the agent history. Me personally, I’m probably not gonna remember to come in here and do that, so that’s why I wanted to add that set of instructions. And so now it is supposedly going to email me, so we’re gonna go ahead and hit Publish on that. All right. And when I go back out, you can see there’s my history of what I just ran, and that’s what you can do. You can go in and look at the previous runs in order to see what happened there.

So let’s go back, and I wanna talk about the other agents that I’ve set up, and all of these are extremely useful. As I mentioned, I did create a social media agent first off, just because I wanted to give that a, a true test and give it a go. And I will show you in the edit agent, basically what I really did with this one was I used ChatGPT to produce these instructions. I actually didn’t do a lot of back and forth. This was earlier on. It was over a month ago I was building this.

And so I told ChatGPT what I wanted to do. I was using the new managed agent and how I wanted it to behave, and it gave me some instructions, and then I ran a test, and then I made some adjustments. And you can see that these are fairly long instructions here with examples. There were a few limitations. I was hoping it was gonna be able to save images in specific media folders. It was not able to do that. It was also not able to consistently add my logo, but thankfully in social media you can do a watermark, so that was okay, as long as it left room for the watermark It was also not able to pull from other brand resources that are saved in the media folders, at least not yet.

But again, it’s still in beta. So this is one that I’m not turning on as an automated. This is one that I have based on a chat started. So what I have to do is I have to go in and I have to say run on this one every time I want it to create content. And it is set up so that from the day I say run, it will schedule seven posts over seven days, but starting seven days out, and they will be in draft mode. They will need review by our social media manager.

That way I can just tell it like, “Oh, it looks like we don’t have a lot of content coming up. I’m gonna hit run on this post, on this social media agent.” It’s gonna create content for our social media manager, and then she can go in and review it and approve it and make changes and do whatever she needs to do.

But the other option would be, and there’s all these other triggers in here, you could have it trigger on a form submission. So if you were maybe a home service provider or one of those types of people, maybe you have a form where you submit photos and a description of a project and what you’re working on, and every time you submit that form, this social media agent creates a social post all around it, and you get confident in it, you have it actually post it or you have it put it in draft.

So that’s how you could use something even like a social media agent with a form. You could have it be a survey or a Facebook comment. If people are leaving you comments about how much they love your services, maybe it triggers a social media agent post. Let’s see. You could have it based on tags, based on contacts created Appointments booked. Some of these make more sense for some of the other ones I’m gonna show you. But really a schedule is the other option that you would typically see here on the social media agent example, is that somebody would say, “Every Monday, go in and create five new posts for next week, and then I have a whole week to approve them.” So that’s where you’re gonna see this one used more often than not.

All right, so that’s been a very popular example with a lot of different YouTube videos, and you can see everything that it’s connected to here. But some of the other ones that I have that I haven’t seen anyone talking about these yet, so I really wanted to talk about these other options. So a consulting prep brief. This is a great example of because HighLevel is so connected, all of my things are connected in here, my CRM, my scheduling, my contracts, everything in here is so connected that these built-in agents really have everything they need to know to help me better research prospects.

In this particular agent, and you can see I had the full conversation in here. I had no conversation outside of this agent to build this agent. And I told it, “I wanna build an agent that researches anyone who books a consultation appointment with me two days prior to their appointment, sends me an email with details about their business, what they offer, who they work with, how long they’ve been in business, and what they might be most interested in.” I don’t know about you, but that’s what I do before every consultation, is I go and I research the person that I’m going to talk to.

I look at their website, I look at who they say they work with, and, you know, do they have a lead magnet, all of this other stuff on their website. So basically what I’m doing is I’m saying, “Here’s a list of the things I normally go and research. Send me an email with all this information two days before my appointments.” And you already know when my appointments are because my appointments are booked through the same system here in HighLevel. Now, the first thing it did was come back and say, “Which calendars are you considering a consultation?

Which calendar should trigger this?” And it gave me a list of all my calendars, and I just checked the boxes. It was that simple. And then this was actually the response that it gave me after I checked the boxes, and then it said it built it, and that was it. And I went in here, and it had it… It selected the triggers. It selected the apps, the web search I may have connected the knowledge bases. That might have been the only thing I had to do, and then it set up all the instructions.

So when I came back out, you’ll notice that I hit Run on this once after I built it, and I said, “Run this agent, but if no meetings have occurred in the last two days,” because it’s only supposed to be looking two days back, “then go ahead and look at the last 10 meetings that have occurred.” Like, go further back than that. Now you can see this actually triggered today at 8:00 AM, and I got an email today about an upcoming consultation with all of the information about that person that I’m going to be talking to.

So that was actually perfect, and you can just see how the difference is there, what the difference looked like. That was initiated by a chat, and that was initiated by a trigger. So that’s my consultation prep briefing agent Then we have a spam contact reviewer because, I mean, who loves spam? Not me. So what I decided to do, and I’ve talked about spam before, I have several videos about how we have set up… I was trying to find spam, block spam, tag spam, identify spam in order to keep our [CRM] clean and hopefully just keep the health of our list really good.

And so I’m like, “I wonder what an agent could do for me here?” And that’s how I’m starting to think about these agents is everything that I’m building automations for, I’m starting to think around these processes of, “I wonder if an agent could do this different, better, or even add to the process.” Sometimes maybe an agent could come in simply to research and add certain tags or fill in certain fields, and then the automation could pick up from there, and we’ve used it in that way too for a few clients at this point.

All right. I said, “I would like to create an agent that reviews each contact created for validity and analyzes the likelihood that they are spam. If the agent believes that they are spam, apply the likely spam tag,” which already exists. It goes through, it builds this, it said… Then I did say update the notes because I did a run, and I noticed that it was trying to do a little bit too much analyzing when I actually could give it a little bit of instruction to specifically say anyone, in my system at least, that doesn’t have a name or email should be treated as spam because every form that I have people fill out requires a name and email.

I would never have a form that didn’t require that. So that to me is usually a big indicator, and then anyone who has certain tags in the system would never be spam, regardless of what you see on their contact record or whether they opted out or anything like that. So this was a little bit more indication, so I went and I added a little bit more details there, and it created these goals. Now it chose the trigger, again, it chose the app, it chose all of that, and it says it’s the new contact spam review.

And on this particular one, of course, you know, web search is very helpful because I will probably have their website, or at least their email address will have their website more often than not. We don’t need image generation, audio, video, anything like that. Knowledge base, you know, I could connect this to brand guidelines or my knowledge base about my business. It certainly wouldn’t hurt, and then it goes through, and it’s going to tag as potential spam. I did change the tag name there, by the way. Oh, and here is an example of when you need to update these, you can actually Just type.

See this tag is not quite right, but you can actually just type on these instructions. So you can change anything you want. You don’t have to go through the builder. You can just type and fix anything that you need to fix in this system. I made that one little tweak. We’re gonna hit publish. So this is now anytime a new contact gets added, they are going… It’s going to check to see if the system believes it’s spam or not. And I’ve done a lot of things to reduce the spam and the things that are coming through the system, so hopefully that one doesn’t have to trigger it too much.

Okay. And I just updated that, which is why it moved to the top. I have my boiling lead advisor. So one of the things that I highly recommend everybody turn on in their HighLevel system is lead scoring. Lead scoring is under settings. I have a couple other videos where I talk about lead scoring, and there’s lots of things you can do around lead scoring. But I build a pipeline around lead scoring. I build automations around lead scoring that kind of automates the whole process of when I consider somebody to be cold, lukewarm, warm, hot, and boiling.

And one of the things that I find is in the times where you get super busy, you may not be coming back to your pipelines as often as you should be. Or if you have someone that just doesn’t live in the CRM system quite as much as they should be, then you may need those just sort of nudges around, “Hey, you have a new boiling lead or hot lead.” And this is where really this is the gold, right? This is where it can tell you the system’s already identifying your hot leads, and now it can actually send you reports on your hot leads, as well as the whole prospecting AI features that are in here, which again, I’ve covered before.

But this one here is the one that I just recently set up, very excited about it. If we go into edit the agent, and you can see again the brief back and forth that I did to create this agent. This is how much tech savvy this takes. Very, very simple. I said, “I need an agent that can review contacts who make it to boiling status in my pipeline, analyze them against our ICP, which is ideal client profile, and send me a weekly summary of who would be worth reaching out to directly and why.” All right, so then it asks me which criteria, [unclear: criteria name] criteria should it be using And I checked off the two knowledge bases, and then you can see that it actually included those two knowledge bases.

I then didn’t have to take that action, and it marked those as that was, like, my reply. You could see here where it thought for a little while, and then it said it built a boiling lead advisor. Now, one thing I had set up here to send me a weekly summary, but I didn’t specify email me a weekly summary. After running a test, I did have to say, “Automatically send an email summary to me every week.” So that was one little tweak I had to make after running it because it was basically just leaving it in the agent, which is no better than leaving it in the pipeline stages for me to have to go review.

So for me, I want that email. All right, here is the, the trigger. So this is in the Marketecs prospecting pipeline. Somebody makes it to the boiling pipeline stage. That is the trigger. And that it is set at an interval, a scheduled interval, every Monday at 9:00 AM, and the system set all of this up. I did not touch the triggers on this. This is simply made from this brief conversation that I had. I think that’s what’s amazing me so much about this, if you can’t tell. All right, it’s set up that it’s obviously gonna look in the system here in HighLevel, and it’s gonna do a web search, so any information it has on the potential lead, it’s going to go out and search the internet to see what it can find, and it’s connected to my knowledge bases.

It goes through the goals, the summary, and exactly what I’m looking for for information. I tested this, worked like a charm. Can’t wait to see what I get on Monday. There you have it. Those are my top use cases as of today for Agent Studios managed agents built within HighLevel. I never had to leave the system because I put it all in one place, which is the beauty of this system. All right, so I hope you got some great ideas for your business, example use cases, and you will dive in and create a managed agent this week with something that will be actually beneficial to your business.

And if you don’t have these capabilities in your current CRM or with your current system, again, check out HighLevel, 30-day free trial down in the description, or book a consultation with me, and we’ll see if we can’t figure out exactly what you need. Thanks for watching. Make sure you hit like, subscribe, and share with anyone that might benefit.

Want help choosing and configuring an agent around your sales process? Book a consultation with Marketecs to review the opportunities inside your account and explore support through Marketecs Engine.

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.