Analyzing text in 2025

Should you analyze text with ChatGPT? There are so many different ways to analyze text in 2025 - whether that's categorizing survey responses manually or using a text analytics platform (like Displayr) to fast-track this process. However, recent advances in AI now mean that a free tool like ChatGPT can be used quickly generate insights.

Here, we'll examine some of the reasons why you might want to use ChatGPT for text analysis, provide a how-to guide, and suggest some prompt ideas to help uncover valuable information.

ChatGPT for Text Analysis: The Pros and Cons

There are pros and cons to using ChatGPT in any context, and it's no different when it comes to text analysis. These are some pros and cons to consider when you're deciding if you want to analyze text with ChatGPT.

Pros

  1. Speed: ChatGPT can analyze customer reviews, social media posts, reports, or whatever else you need to analyze in seconds. There is no extensive implementation or training to complete—it is as simple as writing your prompt, uploading your text, and then pressing enter.
  2. Easy: As mentioned above, ChatGPT has a familiar interface that makes it easy to add your text and get results. There is a very minimal learning curve when it comes to actually using ChatGPT. Most skill development comes in the form of prompt engineering.
  3. Cost-effective: ChatGPT's free offering provides great value, while the paid ChatGPT Plus subscription is significantly cheaper than many other text analysis solutions. It is a great starting point for a business that wants to automatically analyze their text without the price tag or for a small business with a limited budget.

Cons

  1. Character limit: OpenAI enforces a character limit per prompt, limiting the amount of text you can enter at any one time. The free version of ChatGPT (ChatGPT-3.5) has 4096 token limit between the input and the output (a token is how OpenAI measures text and is approximately four characters). The paid version (ChatGPT-4) has a limit 32,768 tokens. There are ways to try and bypass this limit (such as uploading text as a PDF), but ultimately, this limit will slow you down when analyzing large amounts of text.
  2. Limitations for complex analysis: As we'll see below, ChatGPT is a fantastic tool for quick, high-level insights, such as sentiment or entity extraction. However, when it comes to more complex analysis—such as principal component analysis—it can be challenging to come up with the right prompts to get there.
  3. Prompt skills: Prompting is the secret to effective results from text analysis with ChatGPT, and like any skill, it takes time to master. Although we'll show you some good prompt ideas to start your text analysis, ultimately, it is something you will have to take the time to learn and master.
Screen capture of ChatGPT showing that message too long

Avoid the dreaded character limit for best results.

How-To Guide for Text Analysis with ChatGPT

Step 1 - Set your goals

Defining your goals should always be the first step in any successful ChatGPT text analysis project. As mentioned above, ChatGPT has clear limitations compared to all-in-one text analytics platforms. As such, the goal for any project should be realistic and achievable. Think about use cases where the text you need to analyze falls within the maximum token limit and analysis that won't require multiple steps.

Step 2 - Choose between ChatGPT-3.5 and ChatGPT-4

ChatGPT offers paid and unpaid options. The free version of ChatGPT (GPT-3.5) handles basic text analysis tasks, but may struggle with complex or nuanced text, and has a smaller token limit (around 4,096 tokens). It is still highly serviceable, making it a great option for those with tight budget limitations.

The paid version (GPT-4) offers enhanced capabilities, including better accuracy in sentiment analysis, topic understanding, and handling longer texts (up to 32,768 tokens). Additionally, GPT-4 performs faster, with more reliable access and advanced customizability, making it ideal for in-depth or large-scale text analysis.

Step 3 - Decide what analysis you want to perform

There are endless possibilities for analyzing text in ChatGPT. However, choosing a method that's right for your data and appropriate for ChatGPT's capabilities will determine the success of your project. Something like sentiment analysis, thematic coding, or topic modeling is going to work well, as they leverage ChatGPT's ability to understand context, identify patterns, and generate insights from large volumes of text.

Step 4 - Enter your text

Actually inputting your text into ChatGPT is one of the main steps when conducting text analysis. This can be as simple as just copying and pasting the text from your source. However, given the character limit, there might be smarter ways to do this in order to analyze more text.

Traditional data cleaning techniques, such as stemming, lemmatization, removing stop words, and tokenizing the text, will not only improve the quality of your analysis but also reduce the overall character count of the text and allow you to input more data. If the text is still over the character limit, you will need to break it up and upload each section separately until you have included all of the data.

Step 5 - Start your analysis

Once your text is in ChatGPT, it's time to begin your analysis! You must specify the type of analysis you want performed by the AI as part of your prompt. Remember always to be as specific and prescriptive as possible. Make it clear what you want to get out of the analysis and carefully think about how you structure your prompts for best results.

Step 6 - Optimize, interpret and implement

It may take some trial and error with different prompts before you are happy with your text analysis output. AI is not perfect, so it's essential to check for any irregularities or mistakes in your data with an eyeball test. Giving feedback when the text analysis is incorrect will help refine ChatGPT's ability to perform these tasks moving forward.

Once you're happy with your data, you can start to actually analyze and interpret the results. This is where you see the answers in your data. If you've asked ChatGPT to perform a sentiment analysis, you might be able to see which section of your customers is unhappy. This is where you implement the findings. This means sharing the data with your organization, developing strategies to improve, and monitoring progress moving forward.

Constructing the Perfect Prompt 

The secret to effectively analyzing text with ChatGPT - and to using any generative AI software - comes in the form of prompt engineering. AI is very good at following instructions and copying examples, so the more specific you can be with your prompts, the better your outcome is going to be.

Providing as much context as possible and asking ChatGPT to explain why it has made certain decisions will also help with accuracy.

Prompt Ideas for Text Analysis with ChatGPT

The prompts you use for your text analysis project will depend entirely on your data and what you are trying to achieve. The list below serves as a starting point for you to come up with relevant ideas.

  • "Analyze the sentiment of the following text. Is it positive, negative, or neutral? Please explain your reasoning: [insert text]."
  • "Read the following text and identify the main themes. Summarize the key themes and provide specific examples from the text to support your analysis: [insert text]."
  • "Analyze the following set of customer reviews. What common themes can you identify regarding the product or service? Provide examples from the reviews: [insert set of reviews]."
  • "Given this passage, list the potential topics that are being explored and describe how they are related: [insert text]."
  • "List the most frequently used words in this passage and their frequency. Ignore common stop words: [insert text]."
Example of a text analysis prompt in ChatGPT.

Remember to always customize your prompts so they suit your text data.

Advanced AI Text Analysis

Analyzing text with ChatGPT is fast, effective, and affordable. However, it is highly dependent on prompt engineering, has limitations around how much text it can process at once, and can be one-dimensional in its output.

If you have a larger dataset or want to perform more advanced analysis, you will need a more robust tool. Displayr's AI-powered text analytics tool allows you to do everything you can do in ChatGPT and more. By combining advanced AI with decades of market research experience, the text analytics tool not only categorizes text into reliable themes but makes it easy to analyze, visualize, and share insights.

Try it free today.