TaizenNew reps ramp on practice, not on live deals.
Every new hire works through scenarios that get harder, and has to pass each one before the next unlocks.
Trusted by leaders at
Practice on buyers built from your real deals.
A new rep doesn't need a generic prospect. Build the buyer from a deal your team already worked, using the CRM, the recorded calls and what's public on the web, and it raises the objections real buyers raised.
Elena Marsh
VP Revenue Operations · Notion
Deal in Salesforce 
4 recorded calls Web signalsShe will bring up
“We tried a tool like this last year and nobody used it.”
Raised on the second discovery call
“Payback has to land inside two quarters.”
Raised twice, both times by the CFO
See who passed what, without sitting in on every call.
One view of the cohort: which stage each rep is on, their best score, and who is stuck. Open any attempt and every rating cites the moment in the transcript it came from.
Tom has failed the hostile room twice, both times on economic buyer. Play the moment at 09:40 before his next attempt.
Sofia Lindqvist · after certification
First 4 real calls · same rubric
On real calls she never asked who signs off. Re-run the hostile room with the same buyer before her next pricing call.
Check that it carried into real calls.
Certification isn't the finish line. Practice and real calls are scored on the same rubric, so you can see which skills a new rep kept once the buyer was real, and which ones need another round.
Products in this use case
How do we get reps to actually use it?
Training plans, rather than an open library nobody opens. A plan is a sequence of scenarios escalating from a friendly buyer to a hostile one, with a passing score required at each stage before the next unlocks, and a certification step at the end. Managers see who has passed what, so it becomes part of ramp rather than an optional tool.
Is this a chat simulation or an actual call?
A live voice call. The buyer interrupts, talks over the rep and changes the subject, and handling that is part of the score. In demo scenarios the rep can share their screen, and the share is recorded in the replay.
Who does the scoring, and can we trust it?
An AI judge scores against a rubric you control, and every rating has to cite the exact moment in the transcript it came from, so you can check its reasoning rather than take a number on faith. It's built to be strict: sessions fail, and a failing score with evidence is more useful to a manager than a generous one. Eight frameworks ship in the box, including MEDDPICC, Challenger, and SPIN, and you can describe your own.
Won't reps just learn to game the score?
That's exactly what comparing practice against real calls is for. A rep whose practice scores climb while their real-call scores stay flat is learning to pass the bot, and you'll see it. It's also why we'd advise against tying compensation to roleplay scores. The moment the score is money, reps optimize for the bot rather than the buyer.
Put your next new hire through it first.
Build a plan from your own deals, set the pass score at each stage, and see who gets to certification.