Using AI on the Course

Learning Objectives

• Identify a wide range of tasks on the course that AI can help with
• Distinguish between acceptable and unacceptable uses of AI during Software Fundamentals
• Use AI to support learning and day-to-day work

Workshop

(Coach Note): this is a short, informal 30-minute session. Trainees have already been given the AI rules for Software Fundamentals; this session makes those rules concrete and then pushes them to use AI more widely within them. The goal is for trainees to leave feeling confident and encouraged, not warned off.

This workshop should be delivered towards the end of the week, remotely. Holding it here gives trainees a few days of hands-on coding first, ensuring they build initial confidence in their core problem-solving skills before introducing AI assistance, and giving them real moments of friction to reflect on. Running it remotely mirrors how they will actually use AI day-to-day at their desks.

Think (2 mins)

Task

Look back over your first few days on the course. Jot down:

• How you've used AI in the past
• Every time you used AI for something on the course
• At least three times you could have used AI to help you learn or get feedback, but didn't

Don't share these yet.

Where are we now? (4 mins)

Draw a line (on a whiteboard, or as an emoji scale in a Slack thread) from "never used AI" to "use it constantly".

Ask trainees to place themselves on it.

Ask a few people at each end what they've used it for, or what has held them back.

Probe their past experiences with AI. Many trainees carry baggage from university or school where AI was strictly banned or treated as academic misconduct.

Note down common blockers.

• "It feels like cheating"
• "I'm worried about accidentally plagiarising"
• "At uni/school we were told never to touch it"
• "I'm worried I won't learn properly or will become reliant on it"
• "Real engineers should write everything from scratch"
• "I didn't know what I was allowed to use it for"
• "I forget it's an option"

The tools site (4 mins)

Introduce the tools site. Acknowledge that the full list of tools can look intimidating at first glance, but in Week 1 we only focus on two.

Demonstrate how to use Learning Mode and the Code reviewer specifically.

Give trainees 2 minutes to navigate to the site, bookmark it, open both tools, and ask any immediate questions.

The rules (5 mins)

State clearly: we want trainees to use AI. Our clients expect people who are comfortable working with it, and that only comes from daily practice. The rules are there to protect learning, not to discourage it.

Recap the rules for Software Fundamentals:

Good uses

• Rephrasing complex concepts
• Suggesting possible options
• Getting feedback
• Generating dummy data

Bad uses

• Generating code for you
• Completing curriculum tasks
• Writing emails/assessments
• Identifying issues/bugs in your code

Explain the reasoning in one line: right now trainees are building the skills you'll need to judge whether AI output is any good. Using AI to do the work (which we treat as plagiarism) skips that step; using it to help think speeds it up.

Run a quickfire round. Read each scenario out and get trainees to vote ✅ (good use) or ❌ (bad use). Briefly discuss any that split the room.

Scenario	Answer
Asking Learning Mode to explain list comprehensions using a cooking analogy	✅ Rephrasing a concept
Pasting in your failing function and asking what's wrong with it	❌ Identifying bugs
Asking what a KeyError means in general, without sharing your code	✅ Rephrasing a concept
Asking for three different ways you could represent a deck of cards, then picking one and coding it yourself	✅ Suggesting options
Asking AI to write calculate_score() "just to compare with mine"	❌ Generating code
Generating 20 realistic Blackjack hands to test your game with	✅ Dummy data
Asking AI to write your weekly update message to a coach	❌ Writing emails
Writing that message yourself, then asking AI for feedback on tone and clarity	✅ Getting feedback
Running your finished, working code through the Code Reviewer for feedback on readability and naming	✅ Getting feedback
Using AI during the Friday quiz	❌ Assessments

Highlight the pattern: if you did the thinking and the work, and AI helped you understand it or improve it, you're fine.

Offer a simple check: "Would I be happy showing this conversation to a coach?" If yes, go for it.

Revisit the blockers from the start. "It feels like cheating" and "I'm worried about plagiarising" should now have clear answers.

Reassure trainees that if they're unsure about something, they should just ask a coach. Nobody will be in trouble for asking.

Possible uses (10 mins)

Split trainees into pairs.

Challenge each pair to list as many good uses for AI on the course as they can in 5 minutes, using their pre-work notes. Encourage them to think well beyond code.

Gather ideas onto a shared Excalidraw board.

Fill in any gaps so that the following are covered:

Learning (Rephrasing concepts, Learning Mode)

• Explaining a concept a different way, at a different level, or with an analogy
• Quizzing you on a topic before an assessment
• Turning workshop notes into flashcards or a summary
• Explaining confusing documentation, or what a type of error means
• Explaining what a git or terminal command does before you run it

Thinking (Suggesting options)

• Brainstorming different ways to approach a problem, before you choose and code one yourself
• Suggesting the kinds of situations a function should handle, so you can write your own tests
• Breaking a big challenge or your week into smaller steps

Feedback (Getting feedback)

• Code Reviewer on finished, working code: readability, naming, structure
• Feedback on a Slack message or email you've written
• Feedback on your presentation plan or slides
• Feedback on your interview answers
• Practising stakeholder conversations with the Stakeholder Communication tool (used later in SF)

Data (Generating dummy data)

• Realistic test data: player names, card hands, CSVs, JSON

Demonstrate one or two of the less obvious ones live (2 mins max). Good options:

• Learning Mode explaining a concept from this week in three different ways
• Presenting two ways you're thinking of modeling a Blackjack hand (e.g. list of tuples vs list of dicts) and asking the model to critique the trade-offs of each

Point out that most of these have nothing to do with writing code, and that's where most of the untapped value is right now.

Commit (5 mins)

Challenge each trainee to post a short message in their career channel saying what two new uses of AI they'll experiment with over the next week.

Tell them we'll come back to this at the end of the next week: who tried what, and what worked?

Stress that being savvy with AI is essential preparation for joining the industry. Many of our clients will expect engineers to use AI best practices and cutting-edge tools to save time effectively, and we want to help trainees build that capability from day one. The habits built now (checking output, asking good questions, using it daily) are what make that possible.

Close by reiterating that AI use is expected and encouraged, as long as it's helping you learn rather than doing the work for you.

(Coach Note): at the end of the week (e.g. in retro), ask for one "AI win" per person. Repeat this regularly through SF to keep the habit going.