Exit Survey - Agentic AI 101 Workshop (20 Feb 2026)

This short survey checks your understanding of the concepts from the presentation and your interest in trying the technology. It takes about 3 minutes.

A confirmation email with your responses will be sent to you after you submit.

Your Name *

Please enter your name.

Your Email *

Please enter a valid email.

Part A: Concept Check

Q1. In the case study you just saw, the presenter used an AI agent in an IDE (not a chatbot like ChatGPT). In your own words, what is the key difference between an AI chatbot and an AI agent in an IDE? *

Please answer this question.

Q2. The case study showed a 17-step process for writing a letter to the editor. Several steps involved giving the AI agent specific feedback (e.g., "the draft feels trivial"). Why does specific feedback matter? *

(a) The AI cannot generate text without detailed instructions (b) Specific feedback helps the AI find non-obvious evidence and produce sharper output (c) The AI will refuse to revise without explicit criticism (d) It doesn't really matter — the AI produces good output on the first try

Please select an answer.

Q3. If you wanted to use an AI agent to convert a set of PDF lecture notes into a study guide, which would be the correct input-process-output structure? *

In the KissingNumber case, the workflow had three parts: input (source files), process (analyse, cross-reference, draft), and output (explanation decks, letter drafts).

(a) Input: the study guide you want | Process: ask AI to find PDFs | Output: PDF files (b) Input: PDF lecture notes in a folder | Process: extract key points and reorganise by topic | Output: a Markdown study guide (c) Input: a ChatGPT conversation | Process: copy-paste into a document | Output: a formatted file (d) Input: the AI agent itself | Process: tell it to study | Output: exam answers

Please select an answer.

Q4. Which of the following is an example of critical human judgment from the case study? *

One lesson was that the human's role is not just to give instructions — the human also provided judgment that the AI could not.

(a) The human asked the AI to search arXiv for the paper (b) The human noticed the SCMP article lacked an arXiv link and decided to search for it (c) The human told the AI to convert the video to text (d) The human downloaded the final letter as a PDF

Please select an answer.

Part B: Willingness to Try

Q5. Before today, I thought IDEs (like VS Code, Trae, or Cursor) were only for programmers. *

Please select a response.

Q6. After seeing the case study, I believe I could learn to use an AI agent in an IDE for tasks in my own work or study. *

Please select a response.

Q7. What is the one thing from the presentation that makes you most interested in trying an AI agent in an IDE? (Or, if you are not interested, what is holding you back?) *

Please answer this question.

Q8. Which of the following tasks would you most like to try with an AI agent? *

Converting files (PDF, PPTX) to other formats Summarising or analysing research papers Drafting or revising a piece of writing Analysing data or creating visualisations Processing multimedia (video, audio, images) Something else

Please select at least one option.

Part C: Research Opportunity

Q9. Would you like to join an action research study on using agentic AI for academic purposes?

This is a study exploring how agentic AI can support academic work. Students will have the opportunity to serve as an ambassador for promoting the technology. Staff will gain insights into how to use agentic AI for research. Participation is entirely voluntary.

Yes, I am interested in joining the study

Agentic AI 101 — Exit Survey

Thank you!