CecilIA: A Chatbot That Transformed Mechanical Engineering

During my 10 years of teaching chemical and mechanical engineering, I noticed that many students in my classes bring questions and poorly learned concepts from prior semesters. For example, in class, my students are required to design thermal machines, but they often struggle to understand energy balances, the distinction between stress and pressure, and their relationship with viscosity, which is necessary to comprehend how forces behave in a moving fluid. The reality is that it is impossible to review basic concepts every time we begin a new training unit. Although I would love to sit down with each student to clarify the fundamental questions they have, my other work as a researcher greatly restricts my availability for consults, so I have to rule that option out. This situation really began to frustrate me and my students, so, although it sounds like a joke, I decided to clone myself so that I could accompany all of my students at all times without the restrictions of schedules or other limitations.

I needed to find a 24/7 solution that doesn’t get tired or need caffeine to function. Our institution was already encouraging us to use Artificial Intelligence (AI). I confess that, as a Millennial, I am always leery of changes, but the situation was critical. So, as a good scientist, I began researching the subject and training myself. I took courses at Tecnologico de Monterrey and elsewhere. When I finally felt prepared, I took the step to create CecilIA. I felt like Dr. Frankenstein. I must admit that my ego led me to name her after myself, but I think I had no choice; she was really cloning me. This is how CecilIA was born – a virtual assistant designed to support engineering students when they need it most.

AI in education is a clear marker of our direction

The relevance of this project goes beyond my classroom. Several studies have demonstrated that many engineering students worldwide drop out because they do not feel adequately prepared to meet the academic challenges in STEM fields (Rosenzweig et al., 2024). Those of us who work in higher education know that an issue left unattended will quickly snowball.

For years, I have accompanied my students in face-to-face, remedial, or group consultations. Those are helpful, but very limited given the time I have to dedicate to each activity. Another determining factor in learning is that many students are timid about asking for help or seeking counseling, thereby retaining their poorly structured concepts that, in the long run, cause significant problems.

The world of education is changing, and most Top 10 or aspirational universities are integrating AI into their educational models (Moorhouse et al., 2023). AI in education has demonstrated its ability to support learning processes in a flexible and personalized manner (Dogan et al., 2023). During the research I conducted for this purpose, I realized that the use of AI in education is not a fad, but a permanence that we need to accept and integrate into our teaching.

I cloned myself into a chatbot that I named CecilIA

With all these ideas in mind, I felt encouraged to give it a try. I designed CecilIA, my virtual assistant in the classroom, as a kind of “second me,” one that had infinite time, inexhaustible patience, and precise, kind communication. I gave it a relatable personality so that my students wouldn’t feel like they were talking to something stiff or intimidating. I spent hours training her to communicate as closely as possible to my way of expressing myself, just a little less “norteña” (northern). I asked her to explain everything step-by-step, to follow up with as many questions as necessary to ensure the student really comprehended, and to switch between Spanish and technical English to help on that front as well.

What happened next exceeded my expectations. The students began using it at any time, even when I was present. It was strange to witness that they utilized this tool as if it were another person, a kind of teaching assistant. My time for one-on-one counselling improved and became more efficient; my students arrived prepared with specific questions, rather than conceptual ones. I could increase the complexity of my classes because my students no longer had deficiencies in basic concepts. Thus, I could optimize my class lessons. I received comments from students saying they felt they studied with me by their side, or that I unblocked them when they couldn’t absorb any more. When I get this kind of feedback, I’m not going to lie – I feel very proud. Behind every student who keeps trying is a future engineer full of confidence.

Of course, there are still areas for improvement. I wish CecilIA could solve more complex problems, generate simulations, and model systems more effectively. Additionally, I would like to integrate the tool with the institution’s platforms to facilitate easier integration into formal courses, and, of course, continue expanding the content library to support other subjects.

This experience reminded me of something fundamental: teachers do not have to be everywhere to make an impact. We can rely on tools that extend our reach without losing the essence of what we do. AI is not here to replace anyone; it is here to help us reach places we could not have reached before. It can accompany our students when we are unable to be present.

AI tools for education

To create CecilIA, I relied on common artificial intelligence tools. I utilized advanced language models, such as those underpinning conversational platforms like ChatGPT and Gemini, to enable the bot to explain concepts clearly and adapt to each student’s level, always in a friendly and accessible tone. I also adjusted her personality, manner of speaking, and approach to guiding students, so that she would seem like a patient tutor who accompanies each student step by step. The idea was not that I would “answer questions” but that I would have an educational conversation like the ones I usually have with my students during a one-on-one assessment.

Moreover, several didactic strategies are integrated into the bot’s configuration, leveraging the flexibility of AI tools. For example, I taught it to devise explanations in stages, detecting when a student is confused, and to reformulate a simple explanation. I also incorporated practical examples, the option to switch between Spanish and English, and the ability to adapt the answers according to the type of problem that the student wants to solve. In other words, CecilIA is not just a chatbot; it is the product of integrating artificial intelligence technology with pedagogical intention, a deep love for teaching, and a teacher who had to evolve.

Reflection

I decided to share my story because I would love to see more teachers encouraged to experiment with similar tools. We do not need to be programming experts to create solutions. Sometimes it is enough to be clear about the purpose: to help those who need it most, at the time they need it. The use of AI tools in education is a highly topical issue. I hope that my experience will help other teachers to seek and embark on similar adventures.

If anyone reading this is interested in giving it a try, has questions, or would like to discuss educational AI, I’d love to hear from you. After all, education doesn’t change on its own; we change it together, trying, failing, tweaking, and trying again.

Along the way, I have learned that innovating in education doesn’t require superpowers, but having a chatbot that’s available around the clock definitely helps.

About the Author

Cecilia Daniela Treviño Quintanilla (cdtrevino@tec.mx) is a research professor at the Institute of Advanced Materials for Sustainable Manufacturing, specializing in polymer, materials, and thermosciences. She develops educational innovation projects and AI applications to improve engineering learning.

References

Dogan, M. E., Goru Dogan, T., & Bozkurt, A. (2023). The use of artificial intelligence (AI) in online learning and distance education processes: A systematic review of empirical studies.Applied sciences, 13(5), 3056.

Moorhouse, B. L., Yeo, M. A., & Wan, Y. (2023). Generative AI tools and assessment: Guidelines of the world’s top-ranking universities.Computers and Education Open, 5, 100151.

Rosenzweig, E. Q., Chen, X. Y., Song, Y., Baldwin, A., Barger, M. M., Cotterell, M. E., … & Lemons, P. P. (2024). Beyond STEM attrition: Changing career plans within STEM fields in college is associated with lower motivation, certainty, and satisfaction about one’s career. International Journal of STEM Education, 11(1), 15.

Editing

Edited by Rubí Román (rubi.roman@tec.mx) – Editor of the Edu bits articles and producer of The Observatory webinars- “Learning that inspires” – Observatory of the Institute for the Future of Education at Tec de Monterrey.