What we learned building CV screening and AI interviews for Bangladeshi employers — written from our own data, with the sample sizes stated.
A CV score is evidence, not a decision. What an AI genuinely cannot see in a CV, and why a person should press the button.
Our own dataMedian score 38 out of 100, and 35% at 48 or above. The real distribution from 78 CVs across six jobs, with the sample stated plainly.
Cost of hiringJob posting prices, measured screening costs, and the one line item only you can price. A worked breakdown rather than a borrowed average.
High-volume screeningIn three of five real job postings the top-scoring CV arrived in the second half of the pile. What reading in arrival order actually costs.
AI in hiringSix dimensions, what each one reads, which actually separate candidates, and the three things no parser sees. With our own numbers.
PracticalNine parts, the reasoning behind each, and a worked before-and-after of the requirements list — the part that does all the filtering.
Cost of hiringWhy the famous multipliers do not transfer, which two components you can price from your own records, and the one-hour self-audit.
High-volume screeningReading them by hand is 25 hours. Our measured batch rate is 800 CVs an hour. What gets faster, what does not, and what breaks.
AI in hiringRetention, export, deletion, what happens at a limit, what is metered, and who can see your candidates. With our own answers.
PracticalSix dimensions, three real weightings side by side, and why weighting achievement heavily can measure CV writing instead of ability.
Cost of hiringThey used to move together and no longer do. What each contains, the trap that sets, and what to measure instead.
High-volume screeningOur own close-out email has been used zero times: 47 of 56 candidates heard nothing. Why it happens, and what it costs.