Monday's team training, ready in ten minutes

Reading time: about 10 minutes

The situation. You want to spend fifteen minutes on Monday getting the team up to speed on something — how to dispense the new thing, what to ask on a question that keeps coming up — and there is never time to prepare it, so it never happens.

What stops this session is not knowledge: it is preparation. And preparing — ordering it into steps, writing three questions, thinking of an example — is exactly what can be delegated. What cannot is the clinical content, which is why the prompt separates the two explicitly.

What you need to hand: One narrow topic — the narrower the better — and fifteen minutes on the clock as a real limit. "Training on diabetes" is not a topic; "what we ask when someone comes for test strips and does not know how to use them" is.

Step by step

Pick a topic that fits in fifteen minutes

The test: can you say in one sentence what you want them to be able to do at the end? If not, the topic is too broad and the session becomes a talk nobody remembers on Tuesday.

The best topics come from what is already happening: the question that keeps repeating this week, the mistake somebody made (no names), the new product nobody knows how to place.

Ask for the outline with the verification separated

The key to this prompt is point 5: have it tell you which of what it wrote must be verified before you say it. An outline that does not distinguish "this is organisation" from "this is a clinical claim" makes you repeat something unverified in front of your team.

Act as a pharmacist preparing a short in-house training session.

TOPIC: <<what we ask and what we explain when somebody comes in with a
new inhaler>>
TEAM: two technicians and one pharmacist
REAL DURATION: 15 minutes, standing, before opening.

Give me:
1. The objective in one sentence starting "by the end we will know…".
2. A 4-block outline with minutes for each.
3. The 3 most typical mistakes we want to avoid.
4. A 3-line practical case to run on the spot, with the expected answer.
5. IMPORTANT: at the end, a list titled "CHECK BEFORE SAYING IT" with
   EVERY clinical or product claim you included that I should verify in
   the product information. If something is our own organisation and not
   a clinical fact, do not put it there.

No brand names. Do not invent figures.

Verify that list. All of it.

Product information in hand, one by one. This is not a formality: it is what makes this session yours. What you say on Monday your team will repeat at the counter for months, and an unverified claim multiplies by three people and two hundred patients.

And if something on that list turns out to be false, do not hide it: tell them in the session. "The AI gave me this and when I checked it was wrong" is by far the most useful five minutes you can give your team about how to use these tools.
Remove one block

It will have given you four blocks adding up to fifteen minutes on paper. In real life that is twenty-five: people ask, somebody comes in, the phone rings.

Cut the block that adds least — usually "context" or "introduction" — and keep three. A fifteen-minute session that lasts fifteen minutes gets repeated next week; one that runs to half an hour never does.
Turn half of it into questions for the team

A session where one person talks for ten minutes is forgotten before Thursday. One where the team answers two questions sticks — and you also find out what they genuinely do not know, which is almost never what you thought.

Ask for it like this:

"Turn it into 4 questions to put to the team out loud, with the right answer underneath for me. Make them about deciding what you would do, not recalling a fact. And make one a case where the right answer is to REFER."

That last condition is what makes the exercise useful in a pharmacy: if no question is answered with "this is not for us", the whole session is teaching people to solve and not to recognise the limit, which is half the job.

And it works better if a different person asks the questions each week. The rotation means whoever prepares has to understand it, which is when learning actually happens.
Finish with one written thing

At the end, ask: "summarise it in three lines to stick up in the back room". That is what survives the week.

Everything else — the outline, the case, the timings — is for you. For the team, three lines on a piece of paper at eye level beat any presentation.

A full example, and what is wrong with it

The topic asked for: "how to explain the use of a pressurised metered-dose inhaler properly". This is the script it returned, with its verification list separated as the prompt required.

What it replies

Script (8 min): Shake the canister. Breathe out fully. Seal the lips on the mouthpiece. Start inhaling slowly and actuate at the same time. Keep inhaling for 3-4 seconds. Hold your breath for 10 seconds. Wait 30 seconds between doses. Rinse the mouth if it contains a corticosteroid.

To verify before teaching: "The recommended breath-hold (10 s) and the interval between doses (30 s) are worth checking against the specific device’s data sheet, as they vary between manufacturers."

Proposed exercise: "A patient tells you the inhaler is not working. What do you ask?" — Expected answer: check the technique before assuming the drug is not working.

And now, what is wrong with that reply:

The verification flags the right things — and misses the important one. The 10 and 30 seconds do vary, and flagging them is right. But what genuinely needs verifying is not on the list: the "inhale slowly and actuate simultaneously" order is the part most often done wrong and the one that differs between pressurised and dry-powder devices. The model flagged what sounds like data — the numbers — and took as settled what is technique.
That is the pattern: it flags figures, not judgements. And that is the opposite of what suits you, because a wrong figure gets corrected in two seconds when somebody checks it, whereas a wrong judgement stays in the team for years. The verification list it gives you is not the verification list you need: use it as a starting point and add the technique part yourself.
The exercise is too easy. The answer "check the technique" is got by everybody the moment they hear the question, so it teaches nothing and reveals nothing. A useful case has two defensible answers: "two weeks in with no improvement — do you send them back to the doctor or check the technique first?" gets discussed, and the discussion is where what needs correcting comes out.
And there is no case resolved by referring. All eight steps are about solving. None is about recognising the limit — a patient who has also had three days of worsening breathlessness is not an inhaler-technique question. If that never appears, the session is teaching how to solve and not how to distinguish, which is half the job.

The script is correct and the order is right. What needs changing is everything around it: what gets verified and what gets asked.

And there is a check that costs nothing and changes everything: do the exercise yourself before the meeting. If you get it without thinking, it is no good — and if it makes you hesitate, that is the one.

And if your pharmacy is not like that

If the team is two people. Change the format entirely: do not hold a meeting, take the exercise and do it at the counter in a quiet moment. With two people, ten minutes standing up discussing a case beats any script — and the written script becomes just your prompt sheet.

If you are not the one giving it. Then the document needs the answers separated and marked, because whoever gives it will read them first. And rotate the person weekly: whoever prepares has to understand the topic, and that is where the real learning happens — more than in the half hour of listening.

If the topic is one supplier’s specific product. Careful asking for the script plainly: it will write in the register of promotional material, because that is what it has read about that product. Ask explicitly for "no commercial language, and separate what is data sheet from what is sales argument" — the separation is the useful part.

If nobody has ten minutes. Cut it to one: a single question, out loud, with nothing else prepared. One question a week for a year is fifty situations discussed, and that is more training than two away-days that never get scheduled.

When it does not work first time

I check the list and everything is right. So what is the point?

Then you have already won: you taught something knowing it is correct, rather than believing it is. What must not happen is the reverse — giving the session unchecked and finding out when somebody applies it at the counter. Verification is not a formality: it is what separates this from forwarding a PDF.

It gives me a twenty-minute script even when I ask for ten.

Models estimate speaking time badly. Give the limit in words, which it counts approximately well: "800 words maximum" is about eight minutes read calmly. And cut a block yourself afterwards, as step 4 says — it is nearly always the context block at the start.

Nobody joins in when I put the case out.

Nearly always the question has an obvious right answer and nobody wants to state the obvious. Try giving a deliberately bad answer yourself and asking what is wrong with it — it breaks the ice far better than waiting for volunteers, and it teaches the exercise this whole school is built on.

I prepare it and then forget what we decided.

That is why step 5 is ending with one written thing. Not minutes: one sentence, in the same place as that task’s protocol. If it is not written down, in three weeks everybody will be doing it as before and the session will have changed nothing.

Before you use what it gave you

You verified every claim on the final list against the product information.
The objective fits in one sentence and starts with "by the end we will know".
There are no brand names and no unsourced figures.
You kept three blocks, not four.
Half the time is the team talking, not you.
At least one case is resolved by referring.
The questions are about deciding, not recalling.

It only flags what it considers doubtful. And what it treats as certain is exactly what it does not question. If the outline contains a clinical claim that is not on that list, add it yourself: the list is a starting point, not a certificate.

The practice case comes out too easy. It tends to pose the textbook case, and nobody gets the textbook case wrong. Explicitly ask for "a case where a piece of data is missing" or "one where the right answer is to refer": that is where the learning is.

The questions test recall, not judgement. "What is the maximum dose of…?" is either answered or not and teaches nothing. "Somebody comes in with this and says that, what do you do?" gets discussed, and the discussion is where what needs correcting comes out. If the questions can be answered with a number, ask again.

It mixes recommendations from different countries. Regimens, vaccination ages, device names and even referral criteria differ between countries, and the model has read them all. In an internal session that propagates to the whole team unchallenged. Which is why you verify against the local data sheet or your regional guideline, not against what sounds reasonable.

Now, with your own. Look at which question reached you three times this week. That is Monday’s topic, and you already have the prompt. The first time costs ten minutes; from the third, five.

← All practical cases Back to the school →

Monday's team training, ready in ten minutes

Step by step

Pick a topic that fits in fifteen minutes

Ask for the outline with the verification separated

Verify that list. All of it.

Remove one block

Turn half of it into questions for the team

Finish with one written thing

A full example, and what is wrong with it

And if your pharmacy is not like that

When it does not work first time

Before you use what it gave you