Move your organization from class to work

What changes, and what keeps running

Your organization's records stop being marked as course practice and start being marked as everyday work. Everything else keeps running: Evy's daily check, the Charter Steward's daily check, your teams' rule sheets, Kansa's watch lists, your shelves, and the standing rule for anything new.

On the Hattie B's practice inbox

The student said "Switch my organization from course mode to everyday mode." The Charter Steward filed the ruling with the student's words and the date. The Chief listed what changed (new records are marked as everyday work) and what keeps running, all six items. The next morning's brief looked the same as the day before.

Do this now: switch

In the same Chief thread, say, in your own words:

What you should see

Your ruling, word for word, with today's date, and a short list: what changed, and everything that keeps running.

Pick a time for Evy's daily check

Right now Evy's daily check runs at your first brief of the day. You can give it a time instead. Say Hattie B's wants yesterday's catering replies graded before the 7:30 AM prep call. You pick 7:30 AM once. Every morning at 7:30 your Chief starts Evy's check, Evy grades yesterday's work, and your brief is already waiting, with "How your teams did yesterday" filled in.

Do this now, if you want a set time: pick it

In the same Chief thread, say:

Your Chief reads the time back in words and saves it. Then it walks you through adding the morning digest in your app, one step at a time, and nothing is created until you say yes. At the end you click "Run now" once and watch it work. You can also set the time yourself in your dashboard's Settings tab, under "When Evy grades"; your Chief then helps you set the morning digest to the same time.

What you should see

Your Chief reads your time back, like "Once a day, at 7:30 AM". After the test run, it shows you your morning digest. If you ran the test before your time, the digest says Evy's check is "not yet", and that's right: it waits for your time. At your next scheduled time, Evy's check runs and the digest shows how your teams did.

What we've tested, and what we haven't:

• Tested: the morning digest's instructions, started with nobody in the chat, run Evy's check and save the digest.
• Tested: the timing. At your time the check is due. Before it, it waits. It never runs twice in a day.
• We haven't tested this yet: your app's own alarm clock, the Routines page in Claude Code and the Scheduled list in Codex. So your first morning is the test. At your next brief, your Chief checks that Evy's check was started by the schedule, and tells you plainly if it wasn't.

If a scheduled run hits a step that needs your OK, it doesn't guess. It stops, saves a short note saying what it was waiting for, and your next brief tells you. That's why you click "Run now" once while you watch and approve what it asks: then the morning run has what it needs. We saw this ourselves: our first test run stopped on a step nobody had approved, and said so.

Your first brief of the day still runs it, if it hasn't run yet. So if the scheduled check didn't run, saying "brief me" covers it. And if you ask for a brief before 7:30, Evy's check runs then instead. If your computer was asleep at 7:30, Claude Code runs one catch-up when it wakes. Either way, Evy's check runs once a day.

While your computer is off

Your agents run on your computer. They work only while it's awake and the app is open. When you come back, your first brief catches up on what was missed. For most people, that's enough.

Ways to keep them running, simplest first

1. Cloud routines in Claude Code, for jobs that only read, like a nightly summary. They work from your copy of your workbench on GitHub, so they need your evening save point. They use your connected tools without asking, so connect read-only tools.
2. ChatGPT's cloud tasks, for jobs that need only your connected tools, not your workbench folder.
3. An always-on computer, like a small server that never sleeps, with everything wired in. That's the harder path.

None of these can read your chats with your Chief, so Evy's daily check stays on your own computer. The course doesn't teach these: they're there when you want them.

Your "what's next" page is the last page of this week, "What's next, after the course", at course/workforce/week4/lessons/anw-w4-l13.md in your workbench. It lays out each option, what it can't do and where your keys live, and how to keep Langfuse and Evy's daily check running.

Do this now: ask for your page

In the same Chief thread, say:

Read it once. If anything on it isn't clear, ask The Professor, in its thread.

Bounce an idea off the Thinking Partner (Riff)

Riff talks a choice or a rough idea through with you, in two separate voices:

• The Optimist finds what's possible, and what you're already good at.
• The Challenger finds the risks, the trade-offs and the question you've been avoiding.

Then Riff hands it back with one reflection prompt and up to three follow-up questions. Riff never makes the decision, never ranks your ideas as a verdict, and never tells you what to do. It doesn't give medical, legal or financial advice either. When you talk to Riff directly, it replies in the conversation and saves nothing.

We show you: the Hattie B's practice inbox

Sunday is the busiest day, and the inbox fills up while everyone's cooking. The example student had three rough ideas for the Sunday rush email and took them to Riff: a short holding reply that says when someone will answer, a note in the reply listing the quietest times to come in, or sending every Sunday email straight to the owner.

Riff's Optimist liked the holding reply: the team already writes in the practice voice, and the reply promises nothing only the owner can give. Riff's Challenger asked what a holding reply does for someone with a question about today, and whether the owner really wants every Sunday email on a phone during service. The reflection prompt asked which of the three the student would still be glad of on the worst Sunday of the year. The decision stayed with the student.

Do this now: ask Riff about one idea

Pick one improvement you're weighing for one of your teams. In Claude Code: start a new thread, type @ and pick aibl-riff. In Codex: start a new thread and begin with "Ask the Thinking Partner (Riff) to do this, and show me its reply word for word:". Say:

Plan the research with the Research Planner (CLEAR)

Some questions need research before anyone can decide. CLEAR makes a fuzzy question clear before anyone searches:

• It asks you at most three questions, through your Chief of Staff, and only ones whose answer would change the plan.
• It writes a research brief: the question, what's in and out, the questions ranked most important first, where to look, and what "done" means. Anything you didn't answer is marked as a gap, not guessed.
• The External Researcher (Scratch) runs the research, when you say go. CLEAR never searches.
• Then CLEAR checks the findings cover every question. It says COVERED, or RERUN with the questions still missing. Whether the answers are good enough is your call.

Your Chief carries CLEAR's questions to you, runs the check on the brief, and sends Scratch the brief on your go.

We show you: the Hattie B's practice inbox

The example student asked: "Should we answer catering questions by email?" CLEAR asked two questions through the Chief: "Is this about first questions, or about taking orders by email too?" and "What would make email a bad idea for you: speed, mistakes, or something else?" The student answered both. CLEAR's brief ranked three questions, starting with what customers expect to happen after they email a catering question, and named Scratch as the researcher. The student said go. When Scratch's findings came back, the check said COVERED: every question had an answer or a stated gap. The student read the findings and made the call.

Do this now: have CLEAR plan one research question

Pick one question from your own work that you'd want researched before you decide. In your Chief of Staff thread, say:

What you should see

From Riff: the Optimist, the Challenger, one reflection prompt and up to three follow-up questions, ending with "Both voices are here to help you think. The decision is yours." From CLEAR: no more than three questions, then a brief with its ranked questions and any gaps marked. Scratch starts only after your go. When the findings come back, CLEAR says COVERED or RERUN.