1. Home
  2. Reference Manual
  3. Getting Started With ChatGPT Integration

Getting Started With ChatGPT Integration

This document will help you get the most out of the ChatGPT MarketMuse integration. Even if you’ve used ChatGPT before, you’ll still find this helpful.

What is ChatGPT?

ChatGPT is a large language model (LLM) that has been trained on vast amounts of data to learn which words typically follow other words. It uses this statistical knowledge to mimic human language. And it does it very convincingly, to such a degree that some mistakenly believe it’s sentient. 

Limitations of ChatGPT

There are many limitations to LLMs in general and ChatGPT specifically. Here are four important ones.

Lack of Common Sense

Although it has mastered the rules of human language it doesn’t understand what the words mean to a human being. As an article in The Conversation notes, “Humans are biological entities that evolved with bodies that need to operate in the physical and social worlds to get things done. Language is a tool that helps people do that. GPT-3 [ChatGPT] is an artificial software system that predicts the next word. It does not need to get anything done with those predictions in the real world.”

Inability to Understand Context

LLMs typically have difficulty understanding the nuances of human language and the context. Although you can help by providing context and background information in the prompt, ChatGPT can operate without it, leading to potential misunderstandings.

Lack of Explainability

The decision-making processes of LLMs like ChatGPT are complex and opaque. This is something to consider if you plan on using ChatGPT in situations where transparency and explainability are critical.


The data that ChatGPT is trained on may be biased, which could lead to unintended consequences in the generated output, including insensitive language and the promotion of stereotypes.

ChatGPT is Multilingual but Monocultural

An excellent article by Jill Walker Rettberg, Co-Director of the Center for Digital Narrative at the University of Bergen, and Professor of Digital Culture goes into details on why this exists and the data used to train ChatGPT.

Always Edit The Output

As a result of the above-mentioned limitations, it’s imperative that you always edit the output of ChatGPT.

Just like a lawyer verifies every line in a legal contract, you need to verify each line that ChatGPT generates. Be especially wary when generated output includes proper nouns, links, numbers, citations, or quotes. 

A well-crafted prompt does not remove the need for human oversight — you still need to edit.

Who Can Use The ChatGPT Integration

This integration is available on all paid plans.

What Can I Do With This Integration

Let’s take a look at using the pre-defined ChatGPT prompts. Keep in mind you can customize these prompts as you see fit and we’ve written about that here — How to Create a Blog Post Outline Using ChatGPT (and MarketMuse) and Turning an Outline Into an Article Using ChatGPT (and MarketMuse).

There are four prompts from which you can choose:

  • Create Outline
  • Ask a Question
  • Expand a Concept
  • Article Draft

All of these prompts, with the exception of “Ask a Question” incorporate MarketMuse topic model data.

Here’s a suggested workflow for using MarketMuse ChatGPT integration.

Create Outline

Use this prompt to get an initial outline. It won’t be prefect by any means but it’s something to get you started. As the outline forms the foundation of your article, you want to take the time to get this right.

ChatGPT response standard prompt outline
Why you have to always verify AI generated output.

Once your outline is completed, you’ll have a bunch of subheadings each with a number of topics which need to be covered.

Ask a Question | Expand a Concept

Now it’s time to build upon the foundation you’ve created through the construction of an outline. You can do that through either of the following prompts — Ask a Question or Expand a Concept.

Although both of these prompts generate content, their output is slightly different, as the names of the prompts imply. In the first instance you’ll get a direct answer. In the second case, it will be more detailed and expansive.

Experiment with both prompts, using them as you see fit, to build out the body of content.


There are two basic workflows.

  1. You can use the prompts to create the entire piece and then edit it.
  2. Develop one major point at a time — use a prompt to generate the content and then work with that to develop it further. Once you’re happy with the result, move on to the next major point.

Both ways are equally valid, it’s just a matter of personal preference. Either way, you’ll still want to another round of editing once the article is complete.

Article Draft

As this prompt’s name implies, it’s used for creating an initial draft. Since it includes MarketMuse data, it puts you miles ahead of other article generation prompts. Try it out to determine if the output suits your needs.

Keep in mind that it doesn’t offer you the same control as using the other three ChatGPT prompts in concert. You may find that the amount of editing required isn’t worth the investment.

Mechanics of the MarketMuse ChatGPT Integration

ChatGPT Integration and FAQ provides details on creating the connection between MarketMuse and ChatGPT.Once connected you can automatically pass MarketMuse data to ChatGPT using the pre-defined prompts. 

The integration allows you to interact with ChatGPT exactly as you would on OpenAI’s platform. The only exception is that these interactions can not be retrieved at a later date. You can copy and paste a response into MarketMuse, but once the ChatGPT window is closed, that conversation is gone. Opening up the ChatGPT integration starts a new conversation.

Updated on April 13, 2023

Was this article helpful?

Related Articles