A half-day workshop that gives first-time supervisors a repeatable structure for the conversations they most want to avoid.
Open resource
Newly promoted supervisors were avoiding performance conversations until the problem was large enough to require HR. Exit interviews named "my manager never told me" as a leaving reason four times in one quarter.
Twenty-two supervisors promoted from individual contributor roles within the last eighteen months. Most had no management training of any kind. Roughly a third work primarily in Spanish.
Six interviews and a review of one quarter of exit data. The gap was not willingness — it was that nobody had given them a structure, so every conversation had to be invented from scratch under pressure.
A half-day workshop built around one four-step structure, practised three times on escalating scenarios. Participants leave with a pocket card and a filled-in template from their own real caseload.
AI generated the first draft of the twelve practice scenarios, which were then rewritten against real (anonymised) cases — the generated versions were plausible but conflict-free, and the whole point is the conflict. No AI in the assessment rubric.
1 of 2
The room projector would not take HDMI, so the deck ran from a borrowed laptop. Twenty minutes lost at the start; the Frame segment was cut to 15 minutes to recover.
| Priority | Issue | Recommendation |
|---|---|---|
| Major | Scenario cards 7–12 were never printed. | Ran round three from the screen; added the print run to the checklist. |
| Minor | Pocket card text was too small to read at arm's length. | Reset at 13pt for cohort 2. |
| Critical | Two participants had no Spanish materials despite requesting them. | Language preference is now collected at registration, not on the day. |
| Normal | Session ran nine minutes long. | No change — the overrun was the closing round, which is worth the time. |
Print everything the day before, not the morning of. The Spanish materials failure was a registration problem wearing a logistics costume.
Breakout rooms dropped twice. Rounds two and three were run in the main room with cameras off.
| Priority | Issue | Recommendation |
|---|---|---|
| Minor | Remote pairs finished rounds at very different speeds. | Added a written extension prompt to each scenario card. |
Remote breakouts need a written prompt on the card. Anything only said aloud is gone the moment the room splits.
"I used it the next morning. It was still awkward, but it happened."
"The pocket card is in my badge holder. That is the whole reason I remember it."
"I finally had words for the thing I'd been avoiding for six months."
The structure mattered more than the content.
Practice with a real case beat practice with a written one.
Spanish-first supervisors used the scripts verbatim and said so without embarrassment.
Facilitator observations
Round three is where the learning happens; rounds one and two are warm-up.
Nobody used the rubric on themselves without being asked to.
Partially
80% name all four steps at 90 days.
Documented conversations up by half.
No participant reports being unable to work in their stronger language.
Two of three met. Recall landed at 82% and documented conversations rose. The language criterion failed in cohort 1 and passed in cohort 2, so it is counted as not met for the round as a whole.
Collect language preference at registration and print from that list.
Cut rounds one and two to one round and give the time to round three.
Send the pocket card as a phone wallpaper as well as a card.
Pocket card reset at 13pt.
Extension prompt added to all twelve scenario cards.
Registration form now asks for materials language.
Run cohort 3 with the single-round structure and compare recall at 90 days.
Build the Spanish observation rubric, which is still English-only.