Why the pipeline walk always wins
The default 1:1 is a deal-by-deal walk through the CRM, and it survives for good reasons. It needs no preparation, because the CRM prepares it. It produces visible output, because deals get updated. It makes the manager feel current, which is a real need in a job where being surprised is expensive.
It is also necessary work. Forecasting is not busywork. The problem is that it consumes the only recurring block of time a manager and a rep have alone together, and it consumes it every single week without ever touching how the rep actually talks to buyers.
A number is an address, not an instruction
Reviewing numbers is genuinely useful for one thing: deciding where to look. If a rep's conversations stop turning into appointments, that tells the manager which door to open. It says nothing at all about which words fail behind it. Between the metric and the behavior there is always a conversation nobody has examined.
There is a second, quieter problem. A rep describing a call and a rep making a call are two different people using two different registers. The reporting voice is composed, retrospective, and edited. It has had days to decide what the story was. The selling voice is improvised at speed against someone who interrupts. Spend the whole meeting in the first register and you will get a rep who is excellent at explaining calls.
One rule holds the whole structure together. The rep should be in their selling voice for more of the meeting than in their reporting voice. Everything below is scheduling detail around that.
Thirty minutes, arranged around a roleplay
| Minutes | What happens | Why it is there |
|---|---|---|
| 0 to 5 | One number the rep owns, and one moment from a real conversation where they got stuck. The rep brings both. | Anchors the session in reality without letting it become a status update. |
| 5 to 10 | Reconstruct the moment out loud. The manager plays the buyer and says the line the buyer actually said, in the buyer's wording. | Turns a story into a situation. Stories can be narrated. Situations must be handled. |
| 10 to 20 | Run the roleplay. The rep responds live, gets exactly one correction, then runs the same moment again from the same opening line. | The second attempt is where a correction turns into a behavior. |
| 20 to 25 | The rep states the change in one sentence. The rep, not the manager. | If they cannot say it in their own words, it did not land. |
| 25 to 30 | Commit to repetitions, not intentions. How many times, on what scenario, before the next session. | "I will work on it" is not a commitment. A count is. |
Notice what is missing. There is no deal-by-deal walk, no forecast, no update on the three accounts in legal. That work still has to happen. It happens in writing, or in a separate short call, or in the team meeting. It does not get to eat this half hour, because it will if you let it.
What the manager gives up
- The update. The hardest part for most managers is walking into a 1:1 without getting caught up on the pipeline. Move it somewhere else and defend the boundary for a month, which is roughly how long it takes to stop feeling wrong.
- Being the one with the answer. The urge to demonstrate how you would have handled it is the single most common way a coaching session collapses back into a monologue. Model it once, briefly, then hand the microphone back.
- Fixing everything visible. One behavior per session. Write the rest down in a file the rep can see, and take it in order over the quarter.
- Comfort. A meeting where the rep performs is more uncomfortable than a meeting where they report, for both people. That discomfort is not a design flaw. It is the signal that the session is doing something a status meeting cannot.
Where it goes wrong
- It quietly becomes an evaluation. If anything a rep fumbles in a coaching session resurfaces in a performance conversation, they will bring their safest moment forever after. Separate the two explicitly, say so out loud, and then never break it once.
- Coaching the deal instead of the rep. Working the live opportunity is tempting because it produces a result this week. It moves one deal. Coaching the behavior moves the next fifty conversations. Do both, but know which one you are in.
- No second attempt. A correction with no immediate rerun is advice, and advice decays before Friday. If the clock only allows one thing, keep the rerun and cut the discussion.
- Five improvements at once. Nobody rebuilds five habits simultaneously. A session that fixes everything fixes nothing.
- Cancelling it. A cancelled coaching 1:1 tells the rep exactly where their development ranks. Reschedule, shorten, do it standing up, but do not skip it.
Small teams have a different failure mode
A broker-owner with six agents, or an acquisitions manager sitting ten feet from four cold callers, has an advantage a VP of sales does not: they hear the calls as they happen. They rarely lack information. What they lack is repetition.
The natural habit on a small team is the correction in passing. You hear a rep fold on price from across the room, you tell them afterwards what to say next time, and it takes eleven seconds. That is efficient at telling, and telling is the cheap half. The rep has still never said the better sentence out loud, and the next tired landlord will get the old one. Thirty scheduled minutes is what converts an overheard problem into a roleplay the rep has actually run.
There is an honest constraint to state alongside this. Thirty minutes per rep per week does not survive contact with a team of twenty. At that size something has to give: fewer coaching sessions run properly rather than many run badly, or practice the rep can run repeatedly without the manager sitting in the room for every repetition.
How to tell it is working
Not by pipeline in month one. Watch the meeting itself first. The rep should be talking more than the manager. Reps should start arriving with a moment already chosen, which is the clearest sign they have stopped experiencing the session as an inspection. The same correction should appear less often across months, and that is the one worth tracking, since a correction that keeps returning means the second attempt never happened.
After that, look at whether the specific behavior you coached is moving in whatever you use to score conversations. A session should be aimed at one row of the sales call scorecard at a time and never at the total, and that row is the only number you should expect to move this month.
The part of this that scales badly is the middle twenty minutes, because it needs a second person who can play the buyer convincingly and be available on demand. That is the constraint pichi.ai is designed around. The rep runs the roleplay by voice against an AI buyer as many times as they need, before and after the session, and the manager arrives already knowing which moment to work on. The half hour stays a coaching session instead of turning back into a report.