For staff and administration. Fifteen exercises, at your own pace.
Faculty have their ladder. Students have theirs. This one is yours: fifteen short exercises across five rungs, each run on the work you actually do — the emails, the processes, the documents, the policies. No coding, no jargon, nothing on the line.
Every exercise is drawn from how administrative offices everywhere are actually using AI — drafting communications, turning tribal knowledge into written procedures, taming long documents and messy meeting notes, and, on the last rung, catching the confident-but-wrong answer before it goes out under anyone’s name.
Why this order? Maybe you’ve never tried AI, or you tried it once and it was wrong, or nobody ever showed you where it fits in an office job. That’s why the climb starts with play — and ends with the habit that matters most in administration: nothing goes out under your name unchecked.
Any AI chatbot works — ChatGPT, Claude, Gemini, Copilot. Free versions are fine. Switching between them is fine too — Rung 4 requires it.
One hard rule before you start: no protected data. Never paste student records, grades, personnel matters, health information, or anything confidential into a public chatbot. Every exercise here works on public or made-up material — when one calls for a document, use something already public or swap in invented names and numbers.
Notes are optional. If you want to remember what happened, a line after each exercise is plenty: what surprised you, and what it got wrong. Nobody collects them; they are only there for the write-up at the end.
The starters are a floor, not a ceiling. Every exercise gives you something to paste so you never face a blank box. Swap in your own material — a starter you changed is worth more than one you pasted clean.
Each rung ends with a self-check. If the answer is no, repeat the rung’s last exercise with new material before moving on. Nobody sees your answer.
| # | Exercise | The question to ask |
|---|---|---|
| Rung 1 — PlayNothing on the line. Build fluency. | ||
| 01 | Turn a procedure into song lyrics Paste to start: “Turn this into song lyrics in the style of any artist I name: [paste any procedure, form instructions, or policy you work with].” | How good was it — and what did it get wrong, drop, or make up? |
| 02 | Make a campus concept explain your own job Paste to start: “Explain [something you keep hearing about — enrollment funnels, shared governance, AI itself] using my job: [say what you do].” | Does the concept actually fit your work, or only sound like it does? |
| 03 | Play a confused newcomer to your own process Paste to start: “Play a brand-new employee who has to use my process: [describe it]. Ask me the questions a real newcomer would — including the awkward ones.” | Are the confusing parts of your process the newcomer’s problem — or the process’s? |
| Before you climb on: could you start a conversation with an AI about your own work, without a script? | ||
| Rung 2 — DirectYou steer. The answer depends on you. | ||
| 04 | Make it argue against itself Paste to start: “Answer this work question: [e.g. should our office move this form online?]. Then write the strongest case against your own answer — actually try to beat it.” Then: “Which side is right? Commit to one.” | Which side actually won? |
| 05 | Change the tone, watch the message change Paste to start: “Draft a short announcement about [something routine — a deadline, an office move].” Then: “Same announcement, but [warmer / firmer / for students instead of faculty / half the length].” | What did your edit cause — and which version would you actually send? |
| 06 | Have it quiz you on a process you own Paste to start: “Quiz me on [a process you run — travel reimbursement, event approvals, onboarding], one question at a time. Go harder where I hesitate. Stop when you find a step I do by habit but can’t explain.” | Where did your own command of your own process actually run out — and what did you do about it? |
| Before you climb on: could you say what your instruction changed and what it didn’t? | ||
| Rung 3 — ApplyReal work of your own. This is where the hours come back. | ||
| 07 | Draft a real email — then mark what you changed Paste to start: “Draft an email: [describe the situation and what you need — no names, no confidential details]. Professional but human.” Then: edit it your way before sending. | What did you have to change — and were your edits style, accuracy, or judgment the AI couldn’t make? |
| 08 | Turn tribal knowledge into written procedure Paste to start: “I’m going to describe, out loud and messy, how I actually do [a task only you know how to do]. Turn it into numbered steps a substitute could follow: [describe it].” | Could a colleague really run your process from it? The test is what it missed, and only you know that. |
| 09 | Turn messy meeting notes into minutes — then check what it invented Paste to start: “Turn these rough notes into clean minutes with decisions and action items, each with an owner: [paste messy notes from a non-confidential meeting, or invent some].” Then: “Mark anything you added that wasn’t in my notes.” | What did it invent — a filled-in owner, a decision nobody made — and would you have caught it without asking? |
| Before you climb on: could you point to the spot where the AI stopped helping and you took over? | ||
| Rung 4 — InterrogateExamine the tool itself. The decision stays yours. | ||
| 10 | Compare two AIs on the same work question Paste to start: ask the same question in two different chatbots. Then: “Another AI answered this way: [paste]. Where do you disagree, and which disagreement matters?” | Where do they disagree — and which disagreement matters for your office? |
| 11 | Have it argue the side you reject Paste to start: “I believe [a work position you hold — e.g. this form still needs a signature]. Make the strongest case against it — the version that would worry someone who actually holds my view.” | Was the case against you stronger than you expected — and did it move you? |
| 12 | Mark the line: its recommendation, your judgment Paste to start: “Given [a real decision in your office — scheduling, a vendor, a process change], give me your recommendation labeled RECOMMENDATION. Then list every call only I can make — people, politics, history, budget — labeled MINE.” | Exactly where did the AI stop and you start? This is the point of the rung. |
| Before you climb on: could you name a disagreement between two AIs that matters, and make the call yourself? | ||
| Rung 5 — VerifyIndependence. Nothing goes out under your name unchecked. | ||
| 13 | Fact-check before it goes out Paste to start: “List every fact, number, date, and deadline in this draft: [paste something AI helped write]. For each, say how I’d verify it.” Then: verify the three most important ones yourself. | Would anything have gone out wrong — and which kind of claim was it? |
| 14 | Trace a policy claim to the actual policy Paste to start: “You said [a rule or policy the AI stated]. Find the actual CSUN or CSU policy that says this — the document and section, not a summary. If you can’t, say so plainly.” | Does the policy exist, and does it say what the AI said? AI gives generic-university answers with total confidence; your campus’s rules are the ones that count. |
| 15 | Assess the source before you trust it Paste to start: “Here’s a source you cited or I found: [paste it or its link]. Who published it, what’s their incentive, and what would they leave out?” | Does the source’s credibility match how the AI treated it — or was the AI too trusting? |
| Before you climb on: could you tell a colleague when to trust this and when not to? That is the top of the ladder. | ||
No schedule, no office tie-in. Nothing here depends on a particular department or system — where an exercise needs material, use whatever your desk happens to hold this week, minus anything confidential. The order is a suggestion; it exists only because each rung assumes the one below it.
From writing song lyrics about a procedure to tracing a policy claim to the actual policy. That distance is the lesson. Start wherever you are — and welcome to the Ladder.
One thing is worth writing, once, at the end: a short account of your climb. One to two pages. What you tried, which exercises surprised you and how, and what you would now hand to AI at your desk without a second thought and what you would not let it near. Name the exercises. Bring it to the college’s AI conversation — a person reads it and replies.
If your write-up shows the climb, guide a colleague up. The ladder spreads one climber at a time — and the college is building its AI cadre from faculty and staff who have climbed it.