The debrief for a director-level role runs forty minutes and ends where it started. Four interviewers, four scorecards, each mapped to the competency framework the people team rolled out last year. The scores cluster around "meets expectations." Every competency has a number next to it. And the room still cannot agree on whether to make the offer. The framework produced a grid. It did not produce a decision.
If you own hiring outcomes, you have sat in that room. The competency framework was supposed to end exactly this conversation. It cost a quarter of someone's year to build, possibly a consulting engagement on top of that, and it now lives in a slide deck that interviewers open twice a year. The framework is not wrong. It is just not being used the way it was designed, and the gap between the document and the interview is where hiring decisions actually get made.
The framework was a reasonable answer to a real problem
Competency frameworks exist because unstructured interviewing is genuinely unreliable. When every interviewer asks different questions and weighs the answers by instinct, you cannot compare candidates fairly, and you cannot defend the decision later. Decades of selection research back this up: structured interviews are significantly more predictive of performance than unstructured ones, while unstructured conversations stay vulnerable to bias and show weak relationships with how someone later performs. A framework is an attempt to impose that structure. The instinct behind it is correct.
The problem is not the goal. It is the form the goal takes. A competency framework is a taxonomy: a set of abstract qualities, often a dozen or more, each with behavioral definitions and proficiency levels. It reads as rigorous because it is comprehensive. Comprehensiveness is exactly what makes it hard to use.
The framework decays faster than anyone updates it
A competency framework is a snapshot of what a role needed on the day it was written. Roles do not hold still. LinkedIn's 2025 Work Change Report estimates that 70% of the skills used in most jobs will change by 2030, with AI emerging as a catalyst. A framework built for a software engineering role two years ago was written before the current expectations around AI-assisted development existed.
Most organizations review competency frameworks on a two or three year cycle, if they review them at all. That cadence loses to the pace of the work. The result is a framework that describes a version of the job that no longer exists, scored by interviewers who can see the mismatch and quietly route around it. Nobody is doing anything wrong. The artifact is simply older than the role.
A framework names the destination, not the route
One gap does the most damage. A competency framework tells an interviewer what to assess. It does not tell them how. An interviewer opens the scorecard, sees "drives results" or "strategic thinking," and has to invent, in the moment, a question that will surface evidence of it. Two interviewers assessing the same competency ask two different questions and hear two different things. The framework promised consistency and delivered a label.
This is not a training failure on the interviewer's part. Translating an abstract competency into a sharp, job-relevant question is a real skill, and it is being asked of people on top of their actual jobs, minutes before a candidate joins the call. The framework offloaded the hardest part of structured interviewing onto whoever happens to be in the room. The scores that come back look like data. They are closer to four separate unstructured interviews wearing a shared vocabulary.
What holds up: questions tied to the actual job
The durable unit of a hiring process is not the competency. It is the question, tied to a concrete task the role actually involves. SHRM's recent guidance on building a hiring process makes the same point from the other direction: effective selection starts from core competencies tied to actual job demands, not a sprawling model that tries to capture everything.
In practice, the lighter alternative is short and concrete. Pick the four or five things the role genuinely turns on. For each one, write the actual question, and write the evidence that separates a strong answer from a weak one. That is the whole instrument. It is faster to build than a full framework and faster to update when the role shifts. It also removes the on-the-spot invention that makes panel scores incomparable. An interviewer is not asked to design a structured interview. They are asked to run one.
This is closer to how good recruiters already work. The strongest interviewers do not consult the competency taxonomy mid-conversation. They carry a tight set of questions they trust, calibrated against real candidates over time. The lighter model makes that discipline the standard instead of leaving it to the few people who built it on their own.
Where structured screening becomes consistent by default
Question-level structure rarely sticks because it depends on every interviewer running the same instrument the same way, every time, under time pressure. That is hard to enforce across a busy panel and a full req load. It is the part of the process that should not have required a human to hold it together.
Eximius is built for that layer. Sia, the Eximius screening agent, runs structured screening conversations against the job-specific criteria the recruiter sets, asking every candidate the same job-anchored questions and collecting responses against the same evidence standard. The recruiter defines what the role turns on. Sia keeps the screen consistent across a slate of two hundred candidates without drifting. The recruiter then reviews the structured signal and decides who moves forward. The judgment stays with the hiring team. The repetitive work of asking the same questions the same way does not need to.
If your competency framework is not changing how interviews actually run, the fix is not a better framework. It is a smaller one, expressed as questions instead of abstractions, and a way to make sure those questions get asked the same way every time. Pull up your current framework and try to name the four questions an interviewer should ask from it. If you cannot do that quickly, neither can they.
Want to see what structured screening looks like when every candidate answers the same job-anchored questions? Book a pilot and we'll run your next role through the Eximius workflow.