In the AI pioneers project we are frequently asked for examples of the use of AI in vocational education and training. Good case studies are not easy to come from. But Bolton College, a further education institution in the UK, has run a series of projects around using chatbots. They are currently beta-testing a platform that empowers teachers, students, and administrators to create and share generative AI chatbots tailored to their academic and operational needs. The project is funded by the UK UFI Voc Tech Trust. This project not only highlights the practical applications of AI in education but also addresses critical challenges such as cost, governance, and pedagogical relevance (see https://aftabhussain.com/).

 Lowering Costs, Raising Standards

The motivation behind Bolton College’s initiative was straightforward: to make GenAI chatbots accessible and affordable for every member of its campus community. While AI tools like ChatGPT have surged in popularity, their costs—particularly for institutions serving thousands of users—remain prohibitively high. By developing a local large language model (LLM), Bolton College sidestepped reliance on expensive third-party platforms, ensuring the project adhered to principles of responsible AI, data security, and transparency. The chatbot platform is intended to be accessible by any college.

This approach aligns with broader trends in the UK education sector, where institutions are increasingly seeking cost-effective, ethical AI solutions. A report by JISC (Joint Information Systems Committee) emphasizes the need for “AI systems that prioritize data sovereignty and explainability” in education, principles that Bolton’s project directly addresses.

Context-Aware Chatbots for Teaching and Learning

The platform allows teachers to design chatbots equipped with custom document libraries and tailored prompts that reflect specific lesson plans, projects, or assessments. For example:

• A biology teacher could deploy a chatbot preloaded with resources on cellular respiration, programmed to reference prior lessons or upcoming lab work.
• An administrator might create a chatbot to guide students through enrollment processes, drawing from institutional policies and FAQs.

Learners will be able to use chatbots to ask questions, clarify concepts, receive feedback, test their understanding, and access additional resources on any topic they are studying. By grounding chatbots in situational context, AI interactions remain pedagogically aligned. Teachers have noted that these bespoke tools help students stay focused on curriculum goals, reducing the risk of “off-topic” or generic AI responses.

Feedback

Feedback from early adopters has been positive:

• Teachers appreciate the ability to “sandbox” AI interactions. As one instructor explained, “The chatbots act as extensions of our teaching—they reinforce what we’ve covered in class and prepare students for what’s next.” The process of crafting prompts and curating document libraries has also become a form of professional development, with educators refining their skills in AI prompt engineering.
• Students report greater confidence in using AI when guided by their teachers. One learner remarked, “Knowing the chatbot is designed by my teacher—and knowing how it’s been programmed—makes the answers feel more reliable.”

This reflects findings from a 2024 Educause study, which found that students prefer institution-specific AI tools over general-purpose models, citing trust and relevance as key factors.

Building AI Literacy

A secondary benefit of the project is the upskilling of staff. Teachers involved in the beta test are gaining hands-on experience with AI design, learning to craft prompts that balance creativity with academic rigor. For instance, prompting a chatbot to “explain the causes of World War I in simple terms, using examples from last week’s lecture” requires both pedagogical insight and technical precision.

Bolton College’s initiative serves as a practical model for improving data lieteracy and UNESCOs Frameowrk for satt and students around AI Literacy.

A Blueprint for Responsible AI in Education

Bolton College’s project demonstrates that GenAI can be both affordable and educationally transformative when developed with intentionality. Key takeaways include:

1. Cost Efficiency: Local LLMs reduce dependency on costly external platforms.
2. Responsible AI: Built-in safeguards ensure transparency, data privacy, and alignment with institutional values.
3. Pedagogical Relevance: Context-aware chatbots enhance—rather than replace—traditional teaching methods.

As the beta test progresses, Bolton College plans to share insights with the wider education community, offering a roadmap for institutions seeking to adopt AI ethically and effectively.