An interview competency score is a rating of one specific evidenced behavior — how you solve problems, how you communicate, how you own an outcome — scored on its own, separately from every other behavior and from whether anyone liked you.
The separation is the entire mechanism. Score a candidate once, overall, and you mostly measure the first five minutes: warm and quick in the opening pulls every later rating up with it, and one fumbled answer drags them all down. Scoring dimension by dimension, each against its own evidence, is a structural defense against that. It is also why a weak score is worth having. “Six out of ten” tells you nothing you can act on. “Strong on technical depth, weak on impact” tells you which thirty minutes of preparation will change the outcome.
What is a competency in an interview?
A competency is a repeatable, evidenced behavior that can be observed in how you describe your past work. It is not a personality trait. “Collaborative” is a trait — a claim about who you are. “Resolved a disagreement between two teams by mapping both sets of constraints and proposing a shared deadline” is a competency demonstration — a thing you did, with a mechanism and an outcome someone else could check.
The distinction matters because traits can’t be evidenced and competencies can. An interviewer scoring you on communication is not asking whether you seem articulate; they are asking whether, when you described a complex project, a listener who wasn’t there could follow it. That is a testable property of the answer you gave, not an impression of you as a person.
This is why competency-based interviewing asks for specific past incidents rather than hypotheticals or self-descriptions. The question “tell me about a time when…” exists to force an answer that contains evidence rather than assertion.
Why do interviewers score competencies separately?
Interviewers score competencies separately because scoring them together produces one number that mostly reflects the first impression. This is the halo effect — a bias described by the psychologist Edward Thorndike in his 1920 paper “A Constant Error in Psychological Ratings”, in which a rater’s judgment on one attribute contaminates their judgment on every other. An interviewer who finds a candidate articulate will tend to rate them higher on technical depth too, without ever hearing evidence for it. The reverse happens too: one fumbled opening answer drags down scores on dimensions the answer never touched.
Splitting the evaluation into separate competencies, each with its own evidence, is a structural defense against this. It forces the rater to answer “what did they actually say about problem solving?” rather than “how do I feel about this candidate?”
Structured interviewing — same questions, same dimensions, scored independently — is the standard practice built on this idea, and it is the idea scorecards are built on.
What competencies are assessed in an interview?
Competency frameworks vary by company and by role. The table below covers the dimensions I see recur across engineering, product, commercial and operations interviews, with what a weak and a strong answer sound like in a transcript.
These are general industry dimensions, not SiviGen’s — SiviGen derives its areas per role, which is covered further down.
| Dimension | What it measures | Weak answer sounds like | Strong answer sounds like |
|---|---|---|---|
| Technical depth | Whether you understand the tools and domain you claim, below the level of the job title | Names technologies without describing use; “I worked with Kubernetes” and nothing more; deflects follow-ups to “the team handled that” | Explains a specific decision and its tradeoff: why this datastore over that one, what broke first at scale, what you’d do differently |
| Problem solving and judgment | How you decompose an ambiguous problem and choose between imperfect options | Jumps straight to the solution; describes what happened but not why that path; presents the chosen option as the only option | States the constraints, names the options considered, explains why one won, and admits what the choice cost |
| Communication and structure | Whether a listener with no context can follow your account start to finish | Starts mid-story; no situation set up; long, unsegmented answer that circles back; leaves the outcome implicit | Sets context in a sentence or two, moves through action to result in order, lands a clear ending; the listener never has to ask “wait, who are you in this?” |
| Ownership and impact | Whether the outcome traces back to something you did | Uses “we” throughout; describes the team’s result with no individual action; no outcome, or an outcome with no connection to the action | Uses “I” for the actions and “we” for the context; states a concrete result and the causal link between your work and it |
| Collaboration and influence | How you move work forward through people who don’t report to you | Describes conflict as something that resolved itself; frames disagreements as other people being wrong; no mention of how agreement was reached | Describes the other party’s position accurately, names what you changed in your own approach, and says how the disagreement actually ended |
| Motivation and role fit | Whether your interest in this role is specific or generic | Praise that would apply to any company; “I’m looking for a new challenge”; no reference to what the role actually involves | Connects a specific part of the job description to something you’ve done or deliberately want to move toward, with a reason |
Two patterns run through the whole table. A weak answer is almost always assertion without mechanism — the claim is there, the how is missing. A strong answer contains something falsifiable: a number, a decision, a tradeoff, a named consequence.
How SiviGen scores competencies
SiviGen’s AI mock interview does not use a fixed competency list. The dimensions are derived per interview, at the moment the session is created, from two real inputs: the requirements parsed out of that specific job description, and the fact graph built from your resume. The result is an agenda of typically four to six areas, each tied to an actual requirement of the role or an actual claim on your CV.
That means a competency dimension in a SiviGen assessment reads like “distributed systems debugging” or “stakeholder management across regions” rather than a generic label — because those are the things this job asks about and your background can speak to. Each area is scored 0–100, given a tone of strong, adequate or weak, and attached to one line of evidence drawn from what you actually said. The interview readiness score is then computed in TypeScript as the mean of those competency scores, not guessed by the model.
The tradeoff is honest: per-role dimensions are more useful for preparing for one job and less comparable across different jobs. Two sessions for two different roles score different things.
How to improve a weak competency score
Improvement is dimension-specific, which is the point of splitting the score in the first place.
- Technical depth — pick the three technologies most central to the job description and, for each, prepare one decision you made and one thing that went wrong. Depth reads as specificity about failure, not fluency about features.
- Problem solving — before describing what you did, say what you were choosing between. One sentence naming the discarded option does more for this score than any amount of detail about the chosen one.
- Communication and structure — use the STAR method as a container, not a script. Most weak structure scores come from skipping the situation and starting at the action.
- Ownership and impact — audit your own pronouns. If an answer is all “we”, rewrite it until the sentence describing the decisive action says “I”. The same failure shows up in writing, where “responsible for” hides the thing you actually did — trade the passive phrasing for verbs that name what you did, without inflating the claim into one you would have to defend.
- Collaboration — prepare one story where you changed your mind. Interviewers scoring influence are listening for evidence that persuasion went both ways.
- Motivation — read the job description again and find the one responsibility you’d genuinely want. Specificity is the whole score here.
If you have an assessment in front of you already, the practical order of operations is in the guide on how to use your AI interview feedback.
What a competency score cannot tell you
A competency score derived from a transcript measures what you said and how you structured it, not whether you would be good at the job — and not what a company will decide, which the explainer on reading a readiness score covers in full.
What the score is genuinely good for is the gap between what you know and what you managed to convey. That gap is fixable in an afternoon. The underlying competence is not, and no score should pretend to measure it.
Practice against the real requirements of a real job you’re applying to, get the per-competency breakdown, and fix the one area that came back weakest. That’s the whole loop.
FAQ
What is an interview competency score?
An interview competency score is a rating of one specific evidenced behavior — problem solving, communication, ownership — assessed independently of the other dimensions and scored on its own scale, which varies by employer. Each score is attached to evidence from what the candidate actually said, so a weak score points at a specific gap rather than a general impression.
What is the difference between a competency and a skill?
A skill is something you can do; a competency is a behavior observable in how you describe doing it. “Python” is a skill. “Chooses an appropriate solution under ambiguous constraints and explains the tradeoff” is a competency. Interviews score competencies because they can be evidenced from an answer, whereas a claimed skill cannot.
Why do interviewers score competencies separately instead of giving one overall rating?
To reduce the halo effect, where a rater’s impression of one attribute contaminates their judgment of every other. Scoring dimensions independently forces each rating to rest on its own evidence, and makes weak results diagnosable — you learn which behavior to work on instead of receiving a single unactionable number.
What competencies does SiviGen assess?
SiviGen does not use a fixed list. It builds typically four to six competency areas per interview from that specific job’s parsed requirements and the fact graph built from your resume, so the dimensions match the role you are actually applying for. Each is scored 0–100 with a tone and one line of transcript evidence.
What should I do with a weak competency score?
Take the weakest area and work only on that one. Read the evidence line under it first — the quoted sentence usually shows the failure faster than the number does — then rehearse that single answer out loud rather than rewriting all of them. Improvement is dimension-specific, which is the reason the score is split up in the first place.