Catherine Hill (VP of Marketing) and Terrence Fox (VP of Sales) from CoderPad opened with a familiar mishire: a demand gen marketer who aced every behavioral interview, then showed up on day 1 unable to do the math.
Every role is technical now, and few teams assess that way. In a study CoderPad commissioned, 93% of recruiters said all roles are now technical, but only about half run any skills assessment outside engineering. When they do, it’s usually a take-home case study scored with a thumbs up or down, and they can’t see how the candidate got there.
Mishires are expensive. Terrence cited about $220K for a bad sales hire. He added that sellers are excellent at sounding polished. A customer described as a giant search engine called its sales hiring gap a $40M problem.
Define AI fluency for your company first. Every company is at a different stage of AI maturity, and some can’t use AI at all for legal reasons. Start by studying the person inside your company who best shows what good looks like, and build your benchmark from how they work.
Grade the process, not just the artifact. With an LLM built into the assessment, you can see which model a candidate picks, how they prompt, whether they push back and how they iterate. Written take-homes are increasingly AI-generated filler, so the reasoning is now the signal.
Make it real. The closer the exercise is to the actual job, the better the signal. Terrence runs live role-plays where sales candidates use an LLM to build a prospecting plan, and he scores video interviews against CoderPad’s 5 core values.
