Home/HighLevel/Features/Workflows/How to add a GPT-4 Turbo step in HighLevel workflows

Workflows and Automation

How to add a GPT-4 Turbo step in HighLevel workflows

In short

The GPT-4 Turbo action is an AI step that can be placed inside a HighLevel workflow. It lets you choose a model, write a prompt, optionally add a system prompt, and receive generated text that can be stored or sent in later steps.

AreaWorkflows · Workflow Builder
Used forDigital marketing agency, Home repair service, Online education platform
Setup steps6 steps in 1 task
Last updatedNovember 16, 2026

01Who this is for

  • Marketing agencies can generate custom copy for emails and ads without manual writing.
  • Service providers can create quick follow up messages that reference client data.
  • Sales teams can produce concise conversation summaries to keep records up to date.

02GPT 4 Turbo, explained

The GPT-4 Turbo step appears in the workflow editor as an action you can add after a trigger or filter. When you add the step you select the AI model, enter a prompt, and optionally provide a system prompt that tells the model what tone to use. The step processes the prompt, consumes tokens, and returns text that can be saved in a custom field or passed to an email, SMS, or webhook action.

The step lives inside the workflow canvas under the list of actions. You can mix it with other AI steps that use GPT-3.5 or GPT-4, each step choosing its own model. After the step runs you can test the workflow with a sample contact to see the generated output before publishing.

03How businesses use GPT 4 Turbo

  • Digital marketing agency. The agency adds a GPT-4 Turbo step to a client lead capture workflow. They write a prompt that asks for a friendly follow up message using the lead's name and service interest. The generated text is then sent automatically via email.
  • Home repair service. The service creates a workflow that captures a phone call transcript. A GPT step with a summarise prompt creates bullet points that are stored in a custom field, which the team reviews before scheduling work.
  • Online education platform. The platform adds a GPT step that writes a brief email inviting a student to a new course, inserting the student’s first name and course title via dynamic variables. The email step uses the AI output directly.

04The numbers and rules that matter

  • GPT 4 Turbo is a premium Content AI feature.
  • Token pricing: GPT 4o input $2.5 per million, output $10 per million; GPT 4o mini input $0.15 per million, output $0.60 per million (per HighLevel's help center, checked November 2026).
  • With $10 you can process roughly 3 million input words or generate 750 000 output words on GPT 4o (per HighLevel's help center, checked November 2026).
  • Model Switcher UI lets you choose GPT 3.5, GPT 4 or GPT 4 Turbo per step.

05Before you start

  • Your account must have access to the premium Content AI features that include the GPT-4 Turbo action.

06How to set up and use GPT 4 Turbo in HighLevel

Add a GPT step

  1. Open the workflow editor and click to add a new action.
  2. Select the action type labeled GPT powered by OpenAI.
  3. Choose the model you want from the Model Switcher UI.
  4. If you use a custom prompt, type the prompt text in the provided box.
  5. Optionally insert dynamic variables such as {{contact.first_name}} into the prompt.
  6. Save the step, then test the workflow with a contact to see the AI output.

07What it looks like in HighLevel

08Settings that matter

  • Model Switcher UI. Allows you to pick GPT-3.5, GPT-4, or GPT-4 Turbo for each AI step.
  • System Prompt. Provides optional instructions that shape the AI’s tone and behavior.
  • AI Billing Settings. Shows token usage and cost details for each AI action in your account.

09Things to know before you rely on it

  • Pricing is token based with zero markup; you pay the exact OpenAI rates for input and output tokens.
  • GPT-4 Turbo is a premium action and may not be available on all HighLevel plans.
  • Tokens are counted per execution, so cost scales with how many words you send and receive.

10How it connects to the rest of HighLevel

Michael's take

Michael Reimer
Michael ReimerFounder of CRM Pros, HighLevel Commander and Certified Admin

In my experience, the token based pricing of GPT 4 Turbo can be advantageous for high volume tasks, but for low volume or low complexity needs, a simpler model may be more economical.

Get More InformationOpens on HighLevelPublished November 2, 2026 · Updated November 16, 2026

Common questions

Common questions about GPT 4 Turbo.

What is GPT-4 Turbo and how does it differ from GPT-4 or GPT-3.5?

GPT-4 Turbo is a faster and more cost efficient version of GPT-4. It offers better contextual understanding than GPT-3.5 and lower token costs than the standard GPT 4 model.

Can I use GPT-4 Turbo in any HighLevel plan?

No. GPT-4 Turbo is part of the premium Content AI feature set and may require a plan upgrade.

How is usage of the GPT step billed?

Usage is billed per token. You pay the exact OpenAI price for input and output tokens with no additional markup.

Is it possible to switch models between steps in the same workflow?

Yes. Each AI step lets you select its model independently, so you can mix GPT 3.5, GPT 4, and GPT 4 Turbo as needed.

What can I do with the text generated by the GPT step?

The output can be stored in a custom field, used directly in an email or SMS step, or sent to a team channel via Slack or a webhook.

How many words are roughly in one million tokens?

Around 750,000 words, based on an estimate of 0.75 words per token.

Next step

Pick the features that matter. Then get the plan that unlocks them.

Explore the features you want, start with the five at the top, and upgrade through my link when you are ready. The bonuses and the team come with it.

The features directory explains every HighLevel feature in plain English. Browse it to see what the platform can do.

Back to the FeaturesSee the full details
⌃

Site preview

Explore the next page

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