Effect Chat Generative Pre-trained Transformers in Marketing: Possibilities of ChatGPT utilization on iPaaS

4. 11. 2024

Author: Alderd J. Froolik, LIGS University, Honolulu, Hawaï, USA
Supervisor: Dr. Roberto A. Llauro, Adjunct Professor of LIGS University

Author Note
Alderd J. Froolik https://orcid.org/0009-0009-1736-7232
Correspondence concerning this article should be addressed to Alderd J. Froolik, LIGS University, p/a Zwanebloem 47, 2408LT Alphen aan den Rijn, The Netherlands. Email: alderd@me.com
Author Biography
Alderd J. Froolik received a Master of Business Administration (MBA) and a Master Of Science (MSc.) both with a major in marketing from EDU Effective in 2023 and graduated summa cum laude (97%) on both masters. He is currently a Doctor of Business Administration (DBA) candidate at Quest Sales & Marketing Automations, under the supervision of LIGS University. His research areas include marketing automations, AI technologies, and low- and no-code tools.


This study investigates the potential of Chat Generative Pre-trained Transformer (ChatGPT) technology within the marketing domain. As digital marketing landscapes evolve, the integration of advanced artificial intelligence (AI) tools like ChatGPT offers innovative avenues for enhancing customer engagement, personalizing marketing strategies, and streamlining content creation. Through a comprehensive tools review and empirical analysis, this study explores the (non)capabilities of ChatGPT in generating human-like text, its applicability in various marketing functions such as content creation, customer service, and personalized communication, and its implications for marketing efficiency and effectiveness. The methodology comprises qualitative and quantitative analyses, including a case study of a business that has successfully integrated ChatGPT into their marketing operations, and surveys of marketing professionals on their perceptions and experiences with this technology. The findings suggest that ChatGPT significantly enhances the creative process, enables hyper-personalized marketing at scale, and improves customer interaction with brands. However, challenges such as ethical considerations, data privacy, and the need for human oversight are also discussed. The paper concludes with a discussion on the future prospects of ChatGPT in marketing, suggesting that while the technology presents significant opportunities, it also necessitates careful consideration of its limitations and ethical implications. This study contributes to the emerging discourse on the role of AI in marketing, providing insights for academics, practitioners, and policymakers on leveraging ChatGPT for competitive advantage.


“Effect Chat Generative Pre-trained Transformers in Marketing”: Possibilities of ChatGPT utilization on iPaaS

              With the introduction of ChatGPT-4 the usability of AI in marketing has taken another giant step. December 19th 2023 Open AI reported: “the development of GPT-4, a large-scale, multimodal model which can accept image and text inputs and produce text outputs. While less capable than humans in many real-world scenarios, GPT-4 exhibits human-level performance on various professional and academic benchmarks, including passing a simulated bar exam with a score around the top 10% of test takers. GPT-4 is a Transformer-based model pre-trained to predict the next token in a document.” (OpenAI 2023).

For content marketeers, ChatGPT is growing more and more into a respected colleague. The level of interaction on a prompt-based question-and-answer (Q&A) is resulting in a human-level response. Where normally other co-workers had to be asked, ChatGPT is generating responses which reach the skill levels of senior marketeers. The tests performance on the academic benchmark underwrites the capability and useability of ChatGPT in assisting marketeers in their daily tasks.

Despite its capabilities, GPT-4 has similar limitations to earlier GPT models (Brown et al., 2020; Radford et al., 2019, 2018): it is not fully reliable (e.g. can suffer from “hallucinations”), has a limited context window, and does not learn from experience. Care should be taken when using the outputs of GPT-4, particularly in contexts where reliability is important (OpenAI 2023).

Widely used AI tools are based upon or has integrated GPT-4. The paper will go over some popular tools and discuss the pro and cons of that tool in relation to marketing purposes. The tools are divided into writing, video and images. Also an integration Platform as a Service (iPaaS) to visualize data streams, Make.com (formerly known as Integromat) is discussed in this paper.



Just like in regular conversations, users of generative AI have to make clear what they want and/or expect in the response. The description of this question is known as a prompt. The marketing prompts are designed to generate a balanced mix of generic and contextual output. (Tafesse, W., Wood, B. Hey ChatGPT: an examination of ChatGPT prompts in marketing.). Basically, the quality of the input reflects the quality of the outcome. To get high quality out of GPT-4, there is some testing required to make sure the outcome is as expected when used in marketing automation at scale. In GPT-4 a pre-trained assistant (Custom GPT Instruction) can be created to set credentials to act in a pre-defined manner. An example is: “Analyze my writing below in the following categories: syntax, grammar, vocabulary, diction, tone of voice, imagery- and figurative language, rhythm, and pace. Provide an analysis for each category in one sentence. My blog posts: [insert your written blogposts]”. Use the outcome in the custom GPT. Then the user can call the assistant to make use of this writing standard. The assistant will maintain the writing standard to build a range of blogposts in the same style so your writing can be recognized by your audience.



Utilizing the best of all AI applications is by using an iPaaS. With iPaaS data streams are visible through apps connected to each other. Make.com (formerly known as Integromat) or Zapier are the most used iPaaS platforms. There are also other iPaaS platforms which do similar data processing.

Using API access directly or one of the thousands predefined apps within the platform will streamline the data. The data can be pulled from databases like Airtable or Google Sheets. This data can be used for further processing. In the generative solutions AI apps can create content from the given resources. In the same iPaaS this content is distributed to a verity of destinations like social media channels, blogposts, websites or used in live chatbots or saved in databases for future processing. The iPaaS platform Make.com and Zapier are very easy to use. Sources and destinations can be altered within minutes. The benefits for marketeers are numerous; saving time, scalable, concise, audience orientated, etc. Content allows businesses to connect with customers, raise brand awareness, influence consumer attitudes, and get feedback on products and services, thereby helping to increase gross revenue (Algharabat et al., Citation 2018). In Figure 1 the scenario of an automation is showing. This automation is searching for data in an Airtable table. If this is found the text is re-written into a compelling Tweet which is then posted on X (formerly Twitter). Then the Airtable record is updated so no duplicated content will show. This scenario runs every 60 minutes during business hours and saves about 6 minutes per tweet. On yearly basis about 200 working days times 8 times per day times 6 minutes approximately 20 8-hour working days are saved.

Afbeelding met tekst, schermopname, Lettertype, logo

Automatisch gegenereerde beschrijving

Figure 1 iPaaS Make.com scenario X posts automations



GPT-4 is very capable to write content over about any subject. That makes GPT-4 the right choice for writing blogposts. When the GPT-4 assistant is used, compelling blogs with the tone of voice of the author can be created in large numbers. When the blogposts are written to a defined audience the effect of will show with the implementation of social media and SEO techniques to boost sales in retail business (Mohammad & Hoque et.al 2020). The source for subjects to write about can be connected through an iPaaS. Connecting an XML/RSS feed, a database containing keywords, a website, or just a text file together with the prompt is a guaranteed way of producing blogpost at scale. In Figure 2 the scenario on Make.com is retrieving a feed from a new item in a Shopify webshop. Then the image of that item is pulled out the database of Shopify. Also the image is pulled from the database using a http-request. As the text is retuned as HTML the text parser will set the text into plain text. Then ChatGPT will analyze the image, writes a blogpost and a title which is then posted on Shopify blogpost. On Facebook, X (formerly Twitter) and LinkedIn a notification of the new blogpost is posted.

This scenario runs once a day and saves 1 hour per day. Based upon 365 days per year this automation is saving at least 45 working days annually.

Afbeelding met schermopname, diagram, Lettertype

Automatisch gegenereerde beschrijving

Figure 2 iPaaS Make.com scenario Blogpost generator



Dall-E is a popular AI app for generating images. Similar like ChatGPT there is prompt required to let the app know what kind of image to design. Examples of a cowboy riding on the moon or ironing on the back of taxi in New York are great images which will attract people. DALL·E 2 is a remarkable achievement in the field of AI, and its ability to generate high- quality images from textual inputs has broad implications for many fields. While there are still limitations and risks associated with the technology, continued research and development will undoubtedly improve its capabilities and expand its potential applications. (Sudershan Manasvi Malhar 2023). However, Rassin et al. 2022) pointed out that DALL-E 2 does not enforce a constraint where each word has a single meaning.

While DALL-E 3 is a significant step forward for prompt following, it still struggles with object placement and spatial awareness (OpenAI 2023). For generating images for blog posts Dall-E 2 and Dall-E3 will certainly contribute to the overall message to be delivered to the reader. On large scale the Dall-E imaging embedded in a iPaaS scenario will reduce the time spend on searching and implementing the right image into the blog posts. Another advantage is the use of Dall-E generated images are free of copyright (OpenAI article 6425277). This reduces the costs of copyrighted images to nihil.

Figure 3 DALL-E 3 AI-generated image.

A DALL-E 3 AI-generated image of "A 3D render of a coffee mug placed on a window sill during a stormy day. The storm outside the window is reflected in the coffee, with miniature lightning bolts and turbulent waves seen inside the mug. The room is dimly lit, adding to the dramatic atmosphere." A DALL-E 3 AI-generated image of "A 3D render of a coffee mug placed on a windowsill during a stormy day. The storm outside the window is reflected in the coffee, with miniature lightning bolts and turbulent waves seen inside the mug. The room is dimly lit, adding to the dramatic atmosphere." OpenAI 2023




The latest in the ChatGPT family is Sora. Sora is a text-to-video generator. It uses the recaptioning technique from DALL·E 3, which involves generating highly descriptive captions for the visual training data. As a result, Sora can follow the user’s prompt in the generated video.


Social Media

Distribution of the content on social media with an iPaaS is easy. For each major social media channel an app is available for distribution. The data will flow through the scenarios and will deliver the content to the desired channel. For the best results the content has to be created for each specific social media channel in order to address the right audience as they differ for each channel.


ChatGPT as Colleague

In a survey conducted on LinkedIn the public was asked: “Do you use ChatGPT as some sort of colleague to assist you on various questions you'd normally ask a colleague?”. Here 43% of the respondents said yes. 57% said no. Another survey on LinkedIn with the question: “Do you see ChatGPT fill in an active rol as marketeer in producing content like blogs or ads in writing, imaging or video?”, indicating ChatGPT as an assistant resulted in 91% positive responses and 9% negative responses.


Benefits of using ChatGPT in marketing

As the examples show, a great deal of working hours can be saved by implementing ChatGPT into the field of marketing. In the world of today, where skilled employees are hard to find and have strong rates, ChatGPT can bring relief on HR. Also the consistency of daily blog posting and consistency in writing style will contribute to the branding of the company. Effective branding can increase brand awareness among customers and when a brand is well known, it can make it easier for new customers to find the product and for the company to expand and grow into new market segments (Padubidri, Swathi. (2023). CONSUMER PERSPECTIVES ABOUT THE EFFECT OF BRANDING ON PRODUCT MARKETING.).



Chat Generative Pre-Trained Transformers like Open AI’s ChatGPT is disrupting the current way of working. Great benefits like saving on labor costs, increasing consistency, building a brand, are just some examples. Implementing ChatGPT into the daily operation will unlock these benefits. When implemented by using an iPaaS solution like Make.com will unlock the benefits at scale. Only a few examples are mentioned but when implemented into all of the business, only lack of imagination will stop integration of ChatGPT. At Make.com ChatGPT can be connected to many more solutions. Think about email responses, DMs (direct messages), responses on comments underneath (blog) posts, automated offers to customers, etc. ChatGPT is also capable to see trends in your data which means you can analyze your data streams and act upon it.

              The biggest thing today is the integration with video. Together with other apps it is possible to clone yourself and appear in podcasts or on tv-shows published on YouTube. With ChatGPT a script can be created based upon whatever source you want, and your avatar can say the generated lines like it is real.

BEWARE: this might also result in unwanted results like abuse and fake news. So, use ChatGPT wisely!


Algharabat, R., Rana, N. P., Dwivedi, Y. K., Alalwan, A. A., & Qasem, Z. (2018). The effect of telepresence, social presence and involvement on consumer brand engagement: An empirical study of non-profit organizations. Journal of Retailing and Consumer Services, 40, 139–149. https://doi.org/10.1016/j.jretconser.2017.09.011 [Crossref], [Web of Science ®], [Google Scholar]

Brown et al. (2020) Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D. Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Language models are few-shot learners. Advances in Neural Information Processing Systems, 33:1877–1901, 2020.

Mohammad & Hoque et.al (2020) E-commerce Optimization with the Implementation of Social Media and SEO Techniques to Boost Sales in Retail Business DOI: https://doi.org/10.31580/jmis.v3i1.1193

OpenAI https://arxiv.org/html/2303.08774v4 December 19th 2023.

OpenAI https://cdn.openai.com/papers/dall-e-3.pdf

Padubidri, Swathi. (2023). CONSUMER PERSPECTIVES ABOUT THE EFFECT OF BRANDING ON PRODUCT MARKETING. International Journal of Humanities & social Science studies (IJHSSS). 12. 131-136.

Radford et al. (2018) Alec Radford, Karthik Narasimhan, Tim Salimans, and Ilya Sutskever. Improving language understanding by generative pre-training. 2018.

Radford et al. (2019) Alec Radford, Jeff Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. Language models are unsupervised multitask learners. 2019.

Rassin, R., Ravfogel, S., and Goldberg, Y. (2022). Dalle-2 is seeing double: Flaws in word-to-concept mapping in text2image models.

Sudershan Manasvi Malhar, "DALL. E 2", International Journal of Scientific Research in Computer Science, Engineering and Information Technology (IJSRCSEIT), ISSN : 2456-3307, Volume 9, Issue 5, pp.48-56, September-October-2023. Available at doi : https://doi.org/10.32628/CSEIT239052 Journal URL : https://ijsrcseit.com/CSEIT239052

Tafesse, W., Wood, B. Hey ChatGPT: an examination of ChatGPT prompts in marketing. J Market Anal (2024). https://doi.org/10.1057/s41270-023-00284-w

Application for study

Interactive online: