How Tenezet assesses salespeople

Why structured assessment is more accurate than an ordinary interview

The free-form “tell me about yourself” interview is a familiar tool. The problem is not the format itself but its accuracy: it is one of the weakest known ways to predict how a person will actually perform.

This is not our opinion — it is one of the most consistent findings in industrial psychology. Schmidt & Hunter's meta-analysis (1998) summarized 85 years of personnel-selection research: structured methods — identical tasks, predefined criteria, behavioral probes — predict job performance substantially better than an unstructured conversation. In practical terms: of ten salespeople hired after an ordinary interview, about five do not last eighteen months; with structured assessment, two or three.

Intuition does not go away — it remains the final filter. It simply works better when it rests on data instead of replacing it.

What we measure: five competencies

Success in sales is not one quality. We assess five competencies, each independently linked to a salesperson's results. The list comes not from consulting practice but from two meta-analyses of sales performance — Verbeke et al. (2011) and Vinchur et al. (1998), covering about 40,000 salespeople combined.

• 1

  Achievement Motivation

  Sustained drive toward results: does the person keep working the goal when the quarter gets hard.

• 2

  Sales Adaptability

  The ability to rebuild the approach mid-conversation when the client's reaction changes.

• 3

  Assertiveness & Proactivity

  Initiative in the contact: does the candidate lead the conversation and bring it to a concrete next step.

• 4

  Stress Resilience

  Whether the quality of work holds under pressure — rejections, objections, hard bargaining.

• 5

  Sales Expertise

  Understanding the logic of a deal: opening contact, uncovering the need, presenting, closing.

Each competency is scored separately, and they do not compensate for one another: a candidate with an even, average profile across all five usually outperforms one with a single strength and four gaps.

A 1–2–3 scale with behavioral criteria

Each competency receives a score from 1 to 3. Every score is tied to a concrete description of behavior — an approach known in psychometrics as behaviorally anchored rating scales (BARS). The score answers not “did we like the candidate” but “what behavior did we observe.”

The scale is deliberately coarser than a five-point school grade: three levels with clear definitions produce a more reproducible score than ten levels without them.

Scores are given in 0.5 steps: the three levels remain the semantic anchors, while intermediate values add sensitivity and working spread for comparing candidates. The final score combines the AI's and the hiring manager's ratings, so its precision reaches a quarter of a point.

AI analysis first, your decision second

After the AI interview and the role-play, the system analyzes the transcript and proposes preliminary scores for each competency — with quotes from the candidate's answers as evidence.

AI does not decide for you. Its role is the first structured analysis: gather the evidence, sort it by competency, show what each score is based on. Only the person who runs the business can weigh a candidate in its context and make the call.

What the AI's first pass gives you:

• Consistency: AI is equally attentive to the first candidate and the tenth, and applies the same criteria to each.

• Protection from the halo effect — the perceptual bias where one strong overall impression inflates every individual score. AI rates each competency separately, from the transcript.

• Speed: an hour of conversation compresses into a structured summary that takes minutes to read.

You can — and often should — disagree with the preliminary scores; your assessment after the role-play carries full weight. The point of the AI is not to be infallible. It is that you start from a structured analysis instead of a blank page.

What you get in the end

Once the candidate has completed all stages and you have entered your scores, the system assembles the final report — a one-to-two-page document:

• Verdict — “Recommended,” “recommended with reservations,” or “not recommended” — with the main reason in a single line.

• Strengths — The competencies where the candidate showed confident behavior — with examples.

• Risks and gaps — Weak spots and areas of uncertainty to weigh in the decision.

• What to verify next — When the decision is not obvious — two to four concrete items: what to clarify, discuss, or test before an offer.

Every conclusion in the report is backed by source data: interview transcripts, per-competency scores with quotes, role-play observations. Any score can be traced to its source.

What Tenezet doesn't do

The method's boundaries are part of the methodology.

• It does not predict outcomes with 100% accuracy. No selection method does. The best structured approaches get it right roughly seven times out of ten — far better than intuition, but not a guarantee.
• It does not remove every bias. The method eliminates the ones tied to fatigue, mood, and the halo effect. But AI is trained on data that can carry biases of its own — one more reason the final decision stays with a human.
• It does not consult on hiring. Tenezet provides the data for a decision. The decision itself, offer terms, and negotiations are yours.
• It does not show candidates their scores. Candidates do not see their scores or comparisons with others. The results are your internal information.
• It does not assess complex B2B sales. Long cycles, enterprise deals, and key-account management are outside the scope. The methodology builds on methods validated for short cycles, where the salesperson closes the deal within a day to two weeks. Long cycles need different predictors that we have not tested.