The interest in the utility and applications of large language models (LLMs) such as ChatGPT for business purposes is growing exponentially. It opens up opportunities to explore the ways LLMs can be put to use. One of the key mechanisms to leverage ChatGPT and the potential of LLMs is to develop an understanding of prompting. In the LLM context, prompting is a strategy to provide instructions to the machine. This maximizes the likelihood of eliciting a correct response.

Prompting occurs in many forms. Therefore, one needs to build knowledge of various forms or types of prompts. This enables the use of the most appropriate type given the information one needs from ChatGPT. The idea of prompting is not very different from using the right keywords to find relevant information. An example of which is when using a search engine, with one key difference. LLMs differ in the sense that they have enhanced capabilities in natural language processing, so they offer opportunities to ask complex questions in an interactive, natural, and conversational setting to find answers. Building rapport with machines through effective interactions in the form of prompting is vital to benefiting from LLM.

Entrepreneurs often start and operate their entrepreneurial endeavours with scarce or modest resources at their disposal. So, using an open-source medium such as ChatGPT can be key to their progress as it provides information of all sorts that can be used for marketing, idea generation, financial management, and day-to-day operations, just to mention a few. However, eliciting useful information can be tricky if one does not know the strategies to raise effective prompts. With that in mind, here we propose three types of prompts and snippets of example interactions with ChatGPT that entrepreneurs can use for their business purposes.

Prompts to induce reflection and thinking

This type of prompt induces an LLM to find information that is related to reflection and thinking about something. So prompts like “let’s reflect on this or think about it” could help elicit responses that help in creativity and developing innovative ideas.

As entrepreneurial endeavours are predominantly innovation-driven, this type of prompt can help entrepreneurs refine and build their innovative ideas. Here are some examples of reflective prompts that we used and the answers provided by ChatGPT.

Example of prompt used: Let’s think about innovations needed for mobile phone

ChatGPT: Certainly, there are several areas in which innovations are needed to further enhance the capabilities and user experience of mobile phones. Here are some potential areas for innovation:

1. Battery Life and Efficiency

Despite improvements in battery technology, longer battery life remains a priority. Innovations that extend battery life significantly or make charging even more efficient would greatly benefit users.

(ChatGPT provided a list of 15 ideas for many different features including Contextual Awareness and Predictive Intelligence, Interoperability and Integration, and Ethical Tech Use).

It was also observed that the way a prompt is raised makes a significant difference. So, one needs to have clarity of objectives while raising the prompt. For instance, prompts such as “Let’s think about mobile phone innovation” and “Let’s think about innovations needed for mobile phones” produce different results, though to some they may sound similar. If we want ChatGPT to list ideas for innovations in something, then using the term “innovations needed” elicits a focused and more relevant response where it lists and explains areas where innovations are needed and the types of innovations that can be developed.

Prompts to elicit a response that is based on the integration of knowledge

To build creative and innovative products and services, entrepreneurs often try to connect the dots. It requires knowledge about various domains, aspects, and disciplines. However, that could be challenging as one may not know everything about a certain inquiry of interest. The complexity of business management, always-evolving customer needs, and rapid technological advancements necessitate using cross-disciplinary knowledge to build innovative products and services. The knowledge integration prompts can help entrepreneurs achieve such an objective.

Below, we provide an example of a knowledge integration prompt.

Example of prompt used: Integrate information about neuroscience and use of neuroscience for building new mobile phone technology

ChatGPT: Leveraging insights from neuroscience to build new mobile phone technology holds the potential to create innovative solutions that align more closely with human cognition and behaviour. Here’s how neuroscience can be applied to develop novel mobile phone technology:

1. Cognitive Ergonomics

Understanding how the brain processes information and perceives visual stimuli can lead to the design of mobile interfaces that are more intuitive and user-friendly. This could involve optimizing colour schemes, iconography, and layout to reduce cognitive load and enhance usability.

2. Adaptive User Interfaces

By monitoring user behaviour and brainwave patterns, mobile interfaces could adapt in real-time to provide a tailored experience. For instance, the interface could become simpler when the user is fatigued or adjusts content based on the user’s cognitive preferences.

(ChatGPT provided a list of 15 ideas in total).

Prompts to elicit responses that cluster information

As the term suggests, this type of prompt can be used to cluster information from various sources. Entrepreneurs can use these prompts to cluster information about customers, competitors, markets, products, and challenges that they may face when they start a new endeavour.

Below, we provide an example of cluster prompts.

Example of prompt used: Group the information about challenges faced by entrepreneurs when developing innovations in mobile phone technology

ChatGPT: Certainly, here’s a grouping of the challenges that entrepreneurs might face when developing innovations in mobile phone technology:

(ChatGPT listed and discussed various Technological and Engineering Challenges, Financial and Resource Challenges, Market and Competition Challenges, Regulatory and Legal Challenges, and Technical Expertise and Talent Challenges)

Entrepreneurs can extend this inquiry further by specifying a country or a region to get an understanding of specific challenges.

Concluding thoughts

ChatGPT offers a host of opportunities for entrepreneurs to gather information. It allows that information to progress their entrepreneurial endeavours. The three types of prompts with examples discussed above show that ChatGPT can serve as a one-stop information portal for every type of information that entrepreneurs need. The value produced by GPT in the form of its responses to prompts is well documented.  It goes well beyond traditional means of seeking information. Further, when a certain prompt does not produce the desired information, changing the prompts is always an option. It is pertinent to note that ChatGPT is producing the next most likely word based on probabilities when forming its response to a prompt. Therefore, using the right prompts is vital to working effectively with ChatGPT.

The fast-paced developments in the integration of LLMs (such as ChatGPT) in various AI-based applications are indicative of LLMs potentially becoming a mainstream technology. Hence, preparing for such an eventuality and in particular being able to leverage ChatGPT,  should be part of the futuristic vision of entrepreneurial endeavours.

  • About the Author
  • Latest Posts

Jiwat is a Professor and author. He writes on some of the contemporary issues in business and technology management. His work is focused on building thought leadership and creative discourse to help individuals and organizations build capabilities to solve problems today and tomorrow.

He has a growing portfolio of work on issues related to artificial intelligence, machine learning, large language models (LLMs) and prompt engineering. His work has been published in top scientific journals.