Assessment #4 eliminates. Accenture: "mandatory assessment. You are required to meet our
sectional cut-off in order to proceed." The cut-off value is not published.
It has a written email section — officially scored on Etiquette, Content, Grammar. Every
prep site online says this section does not exist; they are describing a different test.
Bad audio can mean no score at all, not just a low one. Wired headset, no Bluetooth, not
the laptop mic, empty quiet room. This is the highest-value hour of preparation.
Accent is not scored. Intelligibility, pace, pause placement and word stress are. Do not fake
an accent.
The interview is one panelist, ~30 minutes, the final round, decided on "overall
performance". Four of its questions are known verbatim.
Read first. Confirmed facts, the official scored parameters, how the machine grades speech, the hardware protocol as a checklist, and ranked failure modes
The conflict table · why every aggregator is wrong · the accent question resolved
research/04-qa-report.md
What QA found and fixed, including my own errors
data/claims.csv
40 claims with source, date and confidence
How to read the confidence tags
Tag
Means
[CONFIRMED]
3+ independent sources, or stated by Accenture officially
[LIKELY]
2 independent sources
[REPORTED]
1 source, uncorroborated
[INFERRED]
Reasoned from other evidence
[UNVERIFIED — model inference]
Stated with no evidence
Two-thirds of the candidate evidence for these rounds is one person's recollection. The official
Accenture material is solid; the candidate detail is thin, single-cycle, and comes entirely from
people who passed. Where something is unknown, this library says so rather than filling the gap.