Home/Videos/Workflow API

Video walkthrough · Workflow API

HighLevel Workflow API: Building Automations With AI Tools

What it means when an API can build the workflow itself, not only trigger it.
The video plays here, loaded from YouTube when you press play.Watch on YouTube

In this video Michael Reimer explains the expanded GoHighLevel workflow API. The new endpoints can create workflows, read their full definitions and update triggers, actions and conditions, so AI coding tools and apps can build automations without the visual builder. It is for agencies and developers who build on HighLevel.

Michael is clear about what kind of update this is. It is not a flashy feature that looks great in a demo and never changes your business. He calls it foundational technology, the kind that opens the door to things that could not be built before, and he spends the video on what those things are.

What could the workflow API do before, and what is new?

For years HighLevel has had a powerful workflow builder with one big limitation. If you wanted to create a workflow or change how it was built, you had to go into HighLevel and use the visual builder. There were APIs that could add contacts to a workflow. There were webhooks for sending information between applications. None of that is the same as building the workflow itself.

The new API documentation describes creating workflows, pulling their complete definitions and updating the actions, triggers, conditions, connections and settings that make them work. In Michael's words, that is creating the actual automation, not just triggering one somebody already built.

How would an AI coding tool build a HighLevel workflow?

Picture sitting down with an AI coding tool and explaining what you want. His example is a lead follow up automation for a dental office. When a new lead comes in, send a welcome text. Wait for a reply. Then follow up based on whether they answered.

The AI does not explain how to build that workflow. It generates the workflow and sends it to HighLevel through the API: the triggers, the actions, the steps connected together and the logic that decides what happens next. Then you review it and test it before you publish.

The bigger opportunity, he says, is maintenance. Your AI could read a workflow you already have, look at how it is built, recommend improvements and prepare an updated version. If you want to understand what it is producing, build your first workflow by hand once.

A card engraved Plain English with an arrow pointing to four connected blocks labeled Trigger, Send text, Wait and Reply

What does this mean for client onboarding?

This is one of Michael's favorite examples. An agency signs a new client, and today there is a lot involved in getting them running. You set up the subaccount, install automations, update the messaging, connect calendars and customize the workflows for that business.

Now imagine your own onboarding app. The new client fills out a simple questionnaire: what kind of business they run, their hours, how they want new leads handled and what should happen when someone books an appointment. The app takes those answers and builds an automation package for that business. That means their lead follow up workflow, appointment reminders, missed call follow up, messages after the appointment and review requests, all from what they said during onboarding. For agencies onboarding dozens or hundreds of clients, he says, the time savings are huge. The client onboarding workflow guide shows the manual version.

A clipboard titled Client questionnaire with three arrows pointing to tiles labeled Lead follow up, Reminders and Review requests

Can an app create its own HighLevel automations?

That is the second example. Say you are building a directory, a booking platform or your own software that connects to GoHighLevel. Until now you might have needed a snapshot, or you had to ask users to set up workflows by hand before the app could work.

Now the app itself can create the automations it needs. A business signs up for the booking platform, picks its services, connects a calendar and sets its preferences. Behind the scenes the app builds the right workflows: appointment confirmations, cancellation follow ups, no show recovery, customer reactivation. The customer does not need to know how to build a workflow or how it works underneath. Michael calls that building a real software product on top of HighLevel.

How could agencies update workflows across many subaccounts?

Say you find a workflow that needs to be better. Maybe the follow up strategy changes, or you want to add a qualifying step for a whole group of clients. Normally that is a lot of manual work.

With the workflow API, you can build an app that reads your existing workflows, compares them to your approved standard and prepares the changes. AI can help spot the differences and write the updated versions. With the right permissions and safeguards, those changes can go out through the API. Michael's framing is that you stop treating every subaccount like its own project and start managing automations like a software operation.

How do you start using the workflow API in GoHighLevel?

The video is a concept walkthrough, not a click by click build. These are the moves it points to.

  1. Know the difference. Triggering an existing workflow is one thing. Creating the workflow itself through the API is the new part.
  2. Describe it in plain English. Give an AI coding tool the automation you want and have it generate the workflow.
  3. Review before publishing. Test any AI generated workflow before it goes live.
  4. Generate from onboarding. Use a questionnaire to produce lead follow up, reminders, missed call follow up and review requests for each client.
  5. Let the app build. Have your own app create the workflows it needs when a business signs up.
  6. Standardize across accounts. Compare subaccount workflows to an approved standard and prepare updates through the API with proper permissions.
  7. Check what is enabled. Confirm which API features are turned on for your account or integration before building.

What changes for agencies and developers building on HighLevel?

Michael's point is that this is not about saving a few clicks in the workflow builder. It removes one of the walls between HighLevel and the apps, AI agents and tools people use to build their businesses. He lists what that allows: apps that set themselves up, AI assistants that help build and improve automations, onboarding systems that create workflows for every client, and agencies that manage all of it at scale.

He adds the caution himself. You still need proper authentication, testing and approval before anything goes live, and you should check which API features are turned on for your account or your integration. The page on Claude and MCP with HighLevel covers one way AI tools connect to an account today.

If an API call comes back unauthorized while you test, start with the guide to fix API 401 errors. Not sure which automations a business needs first? The small business automation checklist is a plain starting list. To get an account to build against, start a GoHighLevel 30 day trial through Michael's link. For a wider look at the platform, read Michael's GoHighLevel review.

This article was written from the video transcript. Michael Reimer is a HighLevel Commander and a sponsored affiliate, which means he earns when people sign up through his link.

Questions

Workflow API questions.

What could the earlier APIs do with workflows?

For years, creating a workflow or changing how it was built meant going into HighLevel and using the visual builder. There were APIs that could add contacts to a workflow, and webhooks for sending information between applications. Michael says none of that is the same as building the workflow itself.

What do the new workflow endpoints cover?

The new API documentation describes creating workflows, pulling their complete definitions and updating the actions, triggers, conditions, connections and settings that make them work. In Michael's words, that is creating the actual automation, not just triggering one somebody already built. AI coding tools and apps can use it without the visual builder.

Can changes go live without review?

No, and Michael adds the caution himself. You still need proper authentication, testing and approval before anything goes live. In his example the AI generates the workflow and sends it to HighLevel through the API, and then you review it and test it before you publish. Changes across subaccounts need the right permissions and safeguards.

Can an AI tool build a HighLevel workflow from a plain English description?

That is the direction Michael describes. You explain the automation, such as a lead follow up for a dental office, and the AI tool generates the workflow and sends it to HighLevel through the API: the triggers, the actions, the connected steps and the logic. You review it and test it before you publish.

Does an app still need a snapshot to set up workflows?

Not necessarily. Michael says that until now an app might have needed a snapshot, or users had to set up workflows by hand before the app could work. With these endpoints the app itself can create the automations it needs, such as appointment confirmations, cancellation follow ups, no show recovery and customer reactivation.

What should you check before building on the workflow API?

Check which API features are turned on for your account or your integration. Michael also says you still need proper authentication, testing and approval before anything goes live. The video is a concept walkthrough, not a click by click build, so read the API documentation before you plan a product around it.

More questions

More questions about the workflow API.

Is the HighLevel workflow API available in every account?

Not yet. HighLevel's changelog describes the Workflows API as a private beta that is turned on for selected subaccounts. A subaccount without access gets an error back when a tool calls it, and HighLevel points to a request form for getting added. That is why Michael says to check which API features are enabled for your account or integration before you plan a build around it.

What is new compared with the workflow API that already existed?

The earlier API let a tool work with workflows that were already built, such as adding a contact to one. HighLevel's changelog for the Workflows API lists what is added: create a workflow from a full definition, read one back as an editable definition, validate it, publish or unpublish it, search and delete workflows, browse the available triggers and actions, and read execution logs and statistics.

How does a tool authenticate with the workflow API?

With a private integration token. According to HighLevel's changelog, you create the token under Settings, Private Integrations in the subaccount and send it with each request along with the location ID of that subaccount. The workflow scopes are separate: view, edit, publish and delete workflows, plus view workflow logs. Give a tool only the scopes it needs.

Does a workflow created through the API go live right away?

No. HighLevel's changelog says workflows created through the API start as drafts, and publishing is a separate call with its own scope. There is also a validate call that checks a definition for errors without saving it and points to the step that failed. That matches the order Michael gives in the video: generate the workflow, review it, test it, then publish.

Can the API edit a workflow that is already published?

Not directly. Per HighLevel's changelog, a published workflow has to be set back to draft before it can be edited, and while it is a draft its triggers are off and enrolled contacts are paused. An update replaces the whole workflow, so a tool reads the current version first and sends that version back with its changes. Deleting a workflow is permanent and stops everyone enrolled in it.

Can the workflow API show how a workflow is performing?

Yes. HighLevel's changelog lists execution logs that show enrolled contacts, step outcomes and errors, plus statistics for enrollments, completions and failures per step, grouped by hour, day or week. Logs cover up to 30 days per request and go back 60 days. Test runs from the builder are not included in those logs.

Next step

See it in your own account. Thirty days, every feature.

Start a 30 day trial through Michael's link, 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
⌃

Site preview

Explore the next page

This destination is reserved in the page inventory and is not built yet.