top of page

2023 Oct-Nov Cumulative Development Report

LEAP Development Report

Dec 3, 2023

Alpha V1.0 of CONCIS-E for UBC Sauder School of Business

Dear Followers of CONCIS-E:

It's with immense pride and excitement that we present the latest milestones and advancements in our CONCIS-E project at UBC Sauder. Your steadfast support and confidence in our vision have been instrumental in propelling us to this pivotal point.

Significant Milestone: Alpha Phase Initiation

Today marks a remarkable chapter in our journey. Juggling demanding academic schedules and project responsibilities, our team has triumphantly launched the Alpha V1.0 trial of the CONCIS-E tool. This feat stands as a testament to our team's unwavering commitment and perseverance, bolstered by your invaluable support and mentorship.

Access CONCIS-E Trial Here

As we progress beyond the Alpha phase, our sights are set on the Beta stage, where we'll expand our trials to a broader audience at UBC Sauder. This phase will feature advanced machine learning techniques and more comprehensive data analysis.

Please note: The link provided ( is temporary as we explore further options to align with evolving privacy assessments, student needs, and organizational development strategies.

Tech Overview – CONCIS-E for UBC Sauder

Language Model Insights

CONCIS-E is built on PALM 2, distinct from ChatGPT. Although smaller than the GPT-4 model, it offers a comparable level of accuracy with greater energy efficiency.

Purpose-Aligned Advising

We've dedicated considerable effort to refining the model, ensuring it aligns with academic advising roles. The chatbot responds in a structured manner, informed by practical and pedagogical advice from UGO, LT Hub, and New Venture Design instructors. This design is focused on educational and human-centric interactions.

Areas of Improvement

Please note that while the LLM is capable of generating, it's accuracy is still at risk. Accuracy will continue to be refined as we move towards the final live release. A great advantage of our current design is that we are able to switch to more accurate language models as technology evolves. This ensures that both students and office staff would be more beneficial in the long term use of the product.

UI Design

Our design philosophy emphasizes simplicity and clarity. We aim for a user interface that is both minimalistic and intuitive, enhancing the user experience for students and administrators. The UI features:

  • A clean, uncluttered layout with our signature color scheme.

  • An easy-to-navigate text box for user queries, mimicking real human conversation.

  • Adaptive design for seamless use across various devices and screen sizes, with adjustments in design and color for different devices.

Area of improvement

The current UI is designed for simplicity and easy access. There's currently no authentication feature built into it, as we've encountered some issues in adapting industry methods to this service. This is expected to be fixed in early January next year.

Admin UI design is still being implemented. We ask you to kindly wait for the next cumulative development report.

Prioritizing Privacy

Privacy is paramount in our development process, in line with UBC's Privacy Impact Assessment standards. We ensure:

  • Strong encryption protocols for data handling.

  • Rigorous access controls and authentication to restrict data access.

  • Transparency through clear privacy policies.

  • Proactive engagement with regulatory bodies.

  • A comprehensive response plan for security incidents.

Next Steps

The team is currently working on the redesign of both the frontend (UI, etc.) and backend (language model, datasets, etc.), and will soon come up with a method to automatically scrape new policies and data from both MyBCom and the UBC academic calendar website. This will ensure that students always have the most up-to-date information when receiving academic advising from the entity.

From there, we are also beginning to design a data panel for UGO staff. The new admin console will include insights on how the unit is handling student requests, from aspects of number of requests, common questions asked, complexity of requests handled, etc. Please note that we are still exploring more features to help users get insights from the tool using our current technologies, but things may vary a lot. We will keep in transparent communication with all subscribers on our progress of development.

MiniCONCIS: Idea in the incubator

We plan to introduce the MiniCONCIS, which is a variant of the current CONCIS-E that we are designing. The MiniCONCIS is designed for each course to answer specific syllabus and assignment questions related to that course. The initial user survey has shown that many students and instructors are eager to have this tool to better help them succeed in this course. The technology we propose to use will not be able to generate content directly with the assignments, only with the advice.

bottom of page