AI has at least for most people been this ominous idea of a technology prophesized to overthrow the world and destroy humanity. Although it may still be this terrifying idea, our first interactions with AI were all based on assistance and user aid (Siri, ChatBots, ChaptGPT).
Our group took this notion and conceptualized how the idea of human assistance has slightly been encapsulated by GenAI. Yet, it does not adhere to the fundamental merits of what it means to be mentored. To find these merits a literature review was conducted. The identified gap provided an opportunity to create value for users by having the GenAI mimic the personalities, communication styles, and knowledge of real-life figures from both the past and present.
The model is trained using publicly accessible information, obviously drawing reference to the mentor and using their unique style of delivery to create a tailored experience. The delivery will also reflect literature outlining what exactly successful mentoring is all about (Balse et al, 2023) (Tepper, 1996).
The proposed solution to attack the market that we argued as being vulnerable creates value by modifying the standards associated with the concept. Opposed to traditional mentoring, we were able to centralize the fragmented market (You typically cannot ask a finance mentor for marketing advice), mentors are also often costly (Try finding an expert for 19,99 EURO per month), and best of all this thing has no location constraints and its delivery is near instantaneous!
The value created is then captured via advertising and a subscription to our service. Users are still able to take free advantage of the model but there will be advertisements for them. To ensure that our value is continually monitored for beneficial adjustments to be made, success metrics were identified. The metrics chosen encompass the diffusion of our technology, profitability, and user-centered value creation.
The generation of a prototype using the custom GPT function was deemed important for several reasons. A prototype allows us to first get a grasp of how the idea would look in practice, so in essence, we were able to understand our offering better (In our report we referred to this as the validation of the concept), obviously, we were also able to test the functionality of idea and test how feasible it really is (Arnowitz et al, 2007). All of these insight were then used to create a rough estimation of risk.
Please try it out!
https://chatgpt.com/g/g-rYpcrrwZF-aidvise
References:
Arnowitz, J., Arent, M., & Berger, N. (2010). Effective prototyping for software makers. Elsevier.
Balse, R., Prasad, P., & Warriem, J. M. (2023, December). Exploring the Potential of GPT-4
in Automated Mentoring for Programming Courses. In Proceedings of the ACM
Conference on Global Computing Education Vol 2 (pp. 191-191).
Tepper, K., Shaffer, B. C., & Tepper, B. J. (1996). Latent structure of mentoring function scales. Educational and psychological measurement, 56(5), 848-857.
