Standard SRC issue this week. No founding-pricing pitch in the first paragraph. If you skipped Tuesday's email, that's fine. This stands on its own.
Here's the argument:
Post-call coaching is structurally broken, and it's not the vendors' fault.
I want to walk through why, because once you see it you can't unsee it, and it changes how you think about every dollar your enablement team spends.
Start with the question nobody asks
Why does sales coaching happen after the call?
Pull up the org chart of literally any other coaching domain.
Athletic coaching
Music coaching
Surgical training
Military tactics
Pilot training
In every one of those domains, the coach is present at the moment of execution. The coach is on the sideline during the game. The coach is in the room during the rehearsal. The coach is in the OR during the operation. The coach is in the cockpit during the simulation.
Sales is the only high-stakes execution domain that inverted this. We coach the play after the game is over.
That's strange. And it's strange in a way most people in our industry have stopped noticing because it's been like this for 25 years.
So why is it like this?
The historical answer is technical
For the entire history of the sales coaching industry, call it 1998 to about 2023, we did not have the technology to coach a call live. Real-time speech-to-text was bad. Real-time language understanding was nonexistent. The only viable artifact was the recording.
So the industry built around the constraint. Gong, Chorus, ExecVision, Clari Copilot. All of them are post-call platforms because in 2008 (when the early ones were built) you couldn't do anything else.
That's not a criticism. That's good engineering. Build what's possible.
But the constraint is gone now.
Real-time speech-to-text now runs at human latency on consumer hardware. LLMs can analyze conversation phase, detect objections, surface relevant context, all in under two seconds. The exact technology that makes ChatGPT possible also makes live call coaching possible. The architectural shift happened in the last 18 months.
And here's the part the post-call platforms cannot say out loud: their entire $584M+ raised market cap is based on an architecture that is now optional, not mandatory.
The structural problem with post-call coaching, in three layers
Layer one: timing
Coaching has a half-life. The closer to the moment of pain, the more it sticks. Coach a tennis player on her serve at the moment she misses, and she internalizes the correction. Coach her on Friday from a recording, and she'll have hit 200 more serves in between, half of them reinforcing the bad pattern.
The post-call review window is structurally too late. By Friday, the rep has had 3 to 5 more calls. The deal is either advanced or dead. The lesson is decoupled from the moment of pain. The rep can intellectually understand "I should have said X" and still freeze on the same objection next Tuesday because the correction never reached the muscle memory layer.
Layer two: scarcity of attention
A sales manager has 12 reps and 80 hours a week. Listening to every recorded call is impossible. So the post-call platforms invented "highlight reels." Algorithmic flagging of the moments the manager should review.
Even with highlights, the math doesn't work. If each rep takes 4 calls a day, that's 48 calls/day or 240/week for a team of 12. If the manager spends 5 minutes reviewing each flagged moment, and the algorithm flags 2 moments per call, that's 40 hours/week. Half the manager's time, just on review.
So managers triage. They review the deals that closed (to validate the playbook) or the deals that died (to autopsy what went wrong). They almost never review the deals in motion, which is exactly when coaching could change the outcome.
The post-call platform forces the manager to choose between coverage and timeliness. The manager will always choose coverage. The reps lose.
Layer three: the methodology gap
Every enterprise sales org buys a methodology. MEDDICC, Challenger, Sandler, Force Management. The methodology lives in the playbook PDF and the onboarding deck. Adoption is enforced via Salesforce stage gates: rep can't move the deal forward until they've filled in "Economic Buyer" and "Pain."
So the rep fills in the field to clear the gate. The methodology becomes a CRM hygiene exercise instead of a conversational practice.
Pull any 10 random discovery calls from your team this month and check whether the rep actually pressure-tested the Economic Buyer or just put a name in a field. The gap between methodology compliance in Salesforce and methodology compliance on the actual call is usually 5 to 7x.
Post-call platforms can score the gap. They cannot close it. Closing it requires intervention at the moment the rep is in the call.
The architectural shift
Three things happened simultaneously in 2023 to 2025 that made live coaching technically viable:
Whisper-class real-time transcription dropped latency to sub-second
LLMs got fast and cheap enough to run inference on a live conversation without breaking the bank
On-device fine-tuning made it possible to ship a private model per customer without the data ever leaving their infrastructure
Any one of these three would have been insufficient. The combination unlocks a category that was technically impossible 24 months ago and is now obvious.
The post-call platforms know this. Most of them have small "real-time" features bolted on. None of them have rebuilt their core architecture around live coaching, because doing that means abandoning the post-call moat they spent a decade building. Their boards won't let them do it.
That creates a 12 to 18 month window where the architectural shift to real-time happens, and whoever owns the customer-specific layer owns the moat.
That's why I'm building Parallax now and not in 2028. By 2028 the category is closed.
What this means for you
If you're an enablement leader: stop benchmarking your team against post-call adherence scores. Those scores measure whether the recording matched the playbook, not whether the rep executed it on the call. The number is structurally inflated.
If you're a sales manager: stop spending half your week on call review. The deals you can affect are the ones in motion, and post-call review can't reach them. Move your time to live ride-alongs and real-time enablement.
If you're a rep: the smartest thing you can do this quarter is find a way to get coaching to fire during your calls, not after. If your company has a tool that does this, use it. If they don't, push for one.
That's the structural argument. The product I'm building is the consequence.
If you want a 60-second walkthrough of what live coaching actually looks like in production:
-Coen
