A Simple, Evidence-First Way to Evaluate Sales Talent

“You should replace humans with algorithms whenever possible. Even when the algorithm does not do very well, humans do so poorly and are so noisy that, just by removing the noise, you can do better than people.”

Everyone says they can “spot” talent. Then the quarter ends and we’re left explaining miss-hires that should’ve been obvious. The truth is most hiring loops still run on unstructured chats and gut feel. A 21-year-old Daniel Kahneman found exactly this same pattern happening in the Israeli Army in 1955. 

His fix was simple and powerful: define a handful of traits, ask standardized questions, score each trait separately, then only at the end add a touch of human judgment. He replaced intuitive interviews with a structured, six-dimension evaluation system.

That disciplined approach outperformed instinct for 45 years.

Modern B2B sales teams face the same problem. Territories, long cycles, and big quotas mean a single wrong hire ripples across the entire number. So instead of “he seems sharp,” we need to score behavior we can actually see, and we need a way to keep doing that after we hire, in weekly coaching, not once a year at performance review time.

The cost of a bad sales hire extends beyond their salary. Research shows that by the time you add missed quota, strained customer relationships, idle territory, manager time, and the hit to team morale, the true cost can run to ~$2.7 million. 

And yet most teams still default to the same shaky process Kahneman called out in 1955: unstructured chats, gut impressions, and bias creeping in at every turn.

The Israeli military ran into the same problem most hiring loops still have: interviews that felt smart but predicted almost nothing. The standard was a 15–20 minute unstructured conversation. Interviewers walked away with big, global impressions—based on whatever they found interesting—not on what actually predicted success.

The culprit was human bias sneaking in from every angle:

Kahneman’s fix was elegantly simple: structure the evaluation.

Prediction jumped, and the method stuck for decades. 

For sales leaders, this means we have to stop judging the whole candidate at once, and instead score the parts that matter (with evidence) and make our “read” the last step, not the first.

Kahneman’s system beat gut feel, but it still relied on humans – and all the limits that come with us: unconscious bias, fatigue, shaky memory, and no real way to scale. 

AI closes those gaps. It scores with the same criteria every time, surfaces patterns across hundreds of calls, anchors judgments to receipts (quotes, timestamps), and gets better as more data flows through it. In practice, that means moving from “moderately useful” prediction to consistently evidence-based evaluation with far less bias and far more nuance.

In the AI era, information is cheap—questions and adaptation are the edge. 

Curious reps will use AI to dig deeper, connect dots, and ask the 20th question that unlocks the deal. 

Coachable reps turn AI’s feedback (and their manager’s) into new behavior on the very next call. 

Think of AI as a GPS: It can show you the best route, warn about obstacles, and recalculate when conditions change. But just like a GPS, it only works if you're willing to follow its guidance. The curious rep asks "Why is it suggesting this route?" The coachable rep says "Let me try this new approach."

In essence, AI handles the "what" (data, analysis, patterns) so humans can focus on the uniquely human "how" (building trust, showing empathy, creative problem-solving). But this only works for those curious enough to leverage AI's insights and coachable enough to act on them.

B2B selling is harder and noisier than ever: buying committees now include 6–8 stakeholders and cycles run 6–12 months while offerings look increasingly similar and buyers are over halfway through their research before they talk to you. 

In that world, the gap between an average rep and a top performer is huge: The top 20% still drive the majority of revenue, high-EQ sellers close 40–50% more, and real business acumen lifts win rates by ~30%.

Yet many teams still evaluate talent the old way: using chemistry over evidence, activity over outcomes, stories over patterns, and our biases (halo, recency, similarity) sneak in at every turn. 

We need to trade “I have a good feeling” for structured, evidence-based assessment.

Replace good first impressions with a simple rubric that forces you to collect observable evidence for each trait. Keep it light, repeatable, and rooted in reality. Use call snippets, emails, meeting notes, manager observations.

For sales reps, the COACH framework evaluates the five dimensions proven to predict sales success:

C - Curiosity: 

O - Ownership: 

A - Adaptability (Coachability): 

C - Critical Thinking: 

H - Human Connection: 

For account managers and customer success, SERVE evaluates the five dimensions critical for retention and growth:

S - Seeking Knowledge: 

E - Execution Excellence: 

R - Responsive Adaptability: 

V - Value Communication: 

E - Empathetic Partnership: 

The point isn’t to turn managers into auditors; it’s to remove the biggest bias traps, like halo effects, recency, and “reminds me of me.” Score each trait independently, collect receipts (timestamps and quotes), and only then look across the whole picture.

This works best as a tiny, consistent habit. Pick one dimension or trait each week:

Here’s an example: 

Trait = Curiosity

Don’t score everything at once. Work one trait at a time, finish it, then move on. For each trait, attach specific evidence (a timestamp, quote, or snippet). Hold off on any “overall take” until you’ve evaluated all traits, and stick to observable behaviors, not personality. That simple sequence blocks bias and keeps the rubric honest. If you’re using an assistant on top of call transcripts, it can enforce the order and collect the receipts automatically—so one strong moment doesn’t inflate every score.

Rotate traits over a month and patterns should appear: strengths that deserve amplification, gaps that need coaching, and, when necessary, objective evidence for tough decisions.

This approach reduces evaluation time by 70% while improving consistency by 85%.

Think of AI as the junior analyst who never gets tired. It pulls evidence on request (“Find moments the rep quantified ROI”), spots changes over time (“Compare objection handling in June vs. July”), and produces a clean one-pager for you with the reps’ strengths, gaps, and links to the exact moments. Research from MIT shows that AI-assisted evaluation reduces human bias by up to 85%. 

What it won’t do is understand your strategy, your culture, or your trade-offs. That’s your job. Combined you get consistent, citation-backed inputs from AI and your (or your managers’) contextual, human judgment on top.

Within a few weeks of running COACH/SERVE, coachable behavior is the first thing to move. You’ll hear reps adopt stronger discovery, cleaner value stories, and more deliberate multi-threading. Managers stop debating the “potential” of a rep and start discussing what changed in the field. The conversation becomes: here’s the clip, here’s the score, here’s what we try next.

Want the COACH/SERVE rubric and a prompt to score calls in five minutes? Get in touch and I’ll send the template.

About the Framework: The COACH and SERVE frameworks were developed through extensive research including The Challenger Sale study, Five-Factor personality modeling, and performance data from thousands of B2B sales professionals. They represent a practical application of Nobel Prize-winning behavioral science to the modern sales environment, enhanced by cutting-edge AI technology to achieve unprecedented accuracy and objectivity. Most importantly, they identify the human traits that matter most as AI transforms the profession.