How Do I Know Which Producers Run the Process on the Phone?

You see the numbers, not the calls

Your dashboard shows quotes, binds, premium and retention by producer. One producer binds at twice the rate of another on the same leads. You do not know why. You suspect the second one is reading quotes and hanging up, but you have not heard a full call in months, because you also run the agency. When you do sit in, the producer runs the call the way you taught them, because you are sitting there.

So you coach from the numbers. "Your bind rate is low, sell the value." The producer nods. The number does not move, because the number is not the problem. The problem is a specific thing the producer does or does not say on a specific kind of call, and nobody in the agency can see it.

73% of sales managers spend under 30 minutes a week coaching a rep (ATD 2025, Gong 2025). In an agency, the manager is you, and most of that 30 minutes goes to gathering information rather than coaching. Fix the gathering and the coaching gets its time back.

Write the process before you try to see it

You cannot see whether producers run the process if the process lives in your head. Write it down as steps a call either does or does not contain. Something like: qualify before quoting, present the quote to the household not the request, handle the price objection with the comparison, ask about the second policy, ask for the bind, set the next step. Add the required language your agency or carriers expect, because scoring against it is the only way to know it was said.

Keep it short enough to hold in one hand. Six to eight steps. Every step should be observable in a transcript. "Build rapport" is not observable. "Ask what prompted the shopping before opening the rater" is.

This document does three jobs. It is what new producers learn. It is the script the AI caller practice is scored on. And it is the standard every real call is measured against, which is what makes the score a fact rather than an opinion.

• Qualify before quoting
• Quote to the household, not the request
• Handle the price objection with a dollar comparison
• Ask about the second policy
• Ask for the bind and set the next step

Score every call against your process

Recording is not the answer by itself. A library of calls nobody listens to is a recording trap. The answer is a score per call, produced without anyone listening to the whole thing. In Revenue OS, the Intel Suite records with consent, transcribes and scores every real conversation, whether from the producer's phone, an uploaded office recording or a Zoom review, on six skills: value stacking, objection handling, close timing, price presentation, follow-up setup and talk-to-listen ratio. Then it scores against your written process at 60% of the weight, because your steps matter more than the generic method.

Every call comes back with a number, a timestamped best moment and a timestamped worst moment. You do not open the call at minute zero. You open it at minute four, where the shopper said they were comparing and the producer read the quote anyway.

A call under 80 does not sit in a list. It becomes a redo due 6pm the next business day, a role play against an AI customer seeded with the real caller's words. The producer practices the exact call they lost. You see whether the redo was done, and that feeds the manager enforcement piece of their score.

Read the exceptions, not the recordings

Once calls are scored, your morning is a queue. Managers get a coaching queue sorted by urgency: the calls that scored lowest, the producers with the most misses on a single step, the redos not yet completed. You open two or three, jump to the timestamp, hear the moment and have a five-minute conversation with a specific fix. Ten minutes total. No ride-along, no CRM archaeology.

The pattern view is where the real information is. If four producers all miss the second-policy question, that is a training gap, and next week's lesson writes itself. If one producer misses the qualify step on every shopper call, that is a habit, and the fix is a persona who says "I just want a number" run ten times before the phones open tomorrow.

One number that tells you who does the work

Real-call scores tell you how the producer performs. A behavior score tells you whether they are doing the work that improves performance. In Revenue OS that is the Compliance Score, one number per producer, 0 to 100, recalculated nightly: training completion 30%, daily tasks 25%, role play quality 25%, manager enforcement 20%. Producers see their own number and their rank. You see the team sorted.

The two together answer the question you actually have. A producer with a strong real-call score and a strong Compliance Score is doing the work and it shows; leave them alone. A producer with weak calls and a strong Compliance Score is trying and needs coaching on a specific step. A producer with weak calls and a weak Compliance Score is not doing the work, and now that is a fact on a screen rather than a suspicion.

State Farm agents run this system. It does not change your lead sources or your agency management system. It gives you the thing you have never had: the call itself, scored against your own process, with the moment marked. Coach that moment, every day, for a quarter, and the bind rate on the leads you already buy is how the agency breaks its revenue record.