Skip to main content

How Iris Helps You Learn

Iris is not a search engine and it is not a solution generator. It is a virtual tutor designed around educational research on how students learn most effectively. This page explains the principles behind how Iris responds to your questions.

The 4-Tier Scaffolding System​

Iris uses a calibrated scaffolding approach with four tiers of support. The idea is to give you the minimum help needed to make progress, so that you do as much of the thinking as possible — which is where the real learning happens.

Tier 1: Subtle Hints​

Iris draws your attention to a specific part of your code, a particular concept, or a blind spot you may have overlooked. These hints are intentionally minimal — they focus your attention without telling you what to do.

"Take a closer look at the condition in your while loop on line 14."

Tier 2: Guiding Questions​

If a subtle hint is not enough, Iris asks questions that encourage you to reflect and reason through the problem. The goal is self-discovery: when you figure something out by answering a well-placed question, you are more likely to remember it.

"What value does i have when the loop exits? Is that what you expect?"

Tier 3: High-Level Conceptual Feedback​

When you need more direction, Iris provides strategic guidance about your overall approach — without revealing the implementation. This tier helps you course-correct at the level of strategy, not syntax.

"Your sorting algorithm works for most cases, but it doesn't handle duplicate values correctly. Think about what should happen when two elements are equal."

Tier 4: Generalized Examples​

As a last resort, Iris illustrates the underlying pattern using an analogous example from a different domain. This shows you the shape of the solution while keeping the actual answer for your exercise opaque.

"Consider how a phone book lookup works: you open to the middle, check if the name comes before or after, and repeat with the relevant half. How could you apply a similar strategy here?"

tip

If you feel like Iris is being too vague, tell it what you have already tried. The more context Iris has about your current understanding, the better it can calibrate its help.

Citations​

When Iris references course materials — lecture slides, transcripts, or FAQs — it includes numbered citation markers in the response (e.g., [1], [2]).

  • Hover over a citation to see the source (which slide, video timestamp, or FAQ entry).
  • Use citations to verify what Iris tells you and to find additional context in the original materials.

Citations make Iris's responses transparent and traceable, so you always know where the information comes from.

Follow-Up Suggestions​

After a response, Iris may display clickable buttons below the message with suggested follow-up questions or topics. These are designed to help you:

  • Explore a concept in more depth
  • See related topics you might not have thought to ask about
  • Continue the conversation productively

You are never limited to these suggestions — you can always type your own questions instead.

Proactive Hints​

In some situations, Iris may reach out to you before you ask. For example, after repeated build failures on a programming exercise, Iris might proactively offer a hint to help you get unstuck.

Proactive hints follow the same scaffolding approach — they start subtle and only become more specific if you engage with them.

What to Expect​

Guide, Not Solve​

Iris will not give you complete solutions to exercises. This is by design: research shows that students who receive scaffolded guidance develop stronger problem-solving skills and higher intrinsic motivation compared to those who receive direct answers.

Stay on Topic​

Iris is focused on your course content and learning. It will not help with topics outside the scope of your course, and it will redirect you if a conversation drifts off-topic.

Own Your Work​

Iris is a learning aid. Your submissions, code, and answers should always be your own. Use Iris to understand concepts and get unstuck — not to generate work you submit as yours.

Quality May Vary​

If you are using the on-premise AI option, responses may sometimes be less detailed or accurate than the cloud option. See Getting Started for more on the AI experience options.

Next Steps​