The four heights
The same task, four distances: today's deadline, the next reviewer, the stuck moment, the pattern.
Execute — do the immediate task
+I need the forecast stress test and commentary for next week's board pack. Pull the income…
Execute — do the immediate task
+I need the forecast stress test and commentary for next week's board pack. Pull the income statement drivers for North Division for the next eight quarters, recompute the scenario that drops gross margin by three points, and include a short note on which assumptions changed. Send the model to Elena in FP&A first to sanity‑check formulas, then to Tom, the CFO, for sign‑off, with Thursday COB as the deadline — verify Elena's sign‑off before Tom reviews.
Pasted it? When the reply comes back, push once: ask it to sharpen the weakest part. — Did this prompt help?
Improve — make it easier to accept
+Before I send this to the CFO, make the forecast easier to interrogate. Show the key sensitivity…
Improve — make it easier to accept
+Before I send this to the CFO, make the forecast easier to interrogate. Show the key sensitivity (price, volume, margin) at the top, label cells that were manually overridden since last save, and create a one‑line explanation for any driver that moves more than 15% in a scenario so Tom can find the risk without reworking formulas.
Pasted it? When the reply comes back, push once: ask it to sharpen the weakest part. — Did this prompt help?
Decide — diagnose the stuck moment
+I updated the North Division volume drivers and total quarterly units rose 30% compared with last…
Decide — diagnose the stuck moment
+One driver jumped 30% after the last update.
I updated the North Division volume drivers and total quarterly units rose 30% compared with last version. Elena says she did not change inputs. I'm worried someone pasted a year view over a quarter sheet or a link pulled data from the wrong tab. What are the most likely explanations and the exact sequence of checks that will prove which one happened before I tell Elena and Tom?
Pasted it? When the reply comes back, push once: ask it to sharpen the weakest part. — Did this prompt help?
Become — change the pattern
+Over several board cycles we keep shipping forecasts where a director has manually overridden cells…
Become — change the pattern
+We repeatedly deliver forecasts with hidden manual overrides.
Over several board cycles we keep shipping forecasts where a director has manually overridden cells without documenting why, and each time it costs a half‑day to untangle. Where are we losing control — naming, versioning, or review steps — and what single, low‑friction habit would prevent hidden overrides from making it into the next board pack?
Pasted it? When the reply comes back, push once: ask it to sharpen the weakest part. — Did this prompt help?
Where the evidence lives
Who was seen doing this, and what people really ask.
Software tasks in the LLOS Work Atlas come from evidence, never a feature list: careers attested to do the work, real job descriptions, and the questions people actually ask (with their view counts). Facets — feature, workflow, troubleshoot, administer, deploy, scale — are open metadata: the work decides, not a taxonomy.
Copyright © LLOS.ai · 2026 — original pedagogy, voice, and design — all rights reserved.
The rest of the map
Same library, five ways in.