There is a moment in sales training that managers recognize almost immediately.
The rep is doing fine. The opening is clean. The product explanation is accurate. The discovery questions are in the right order.
Then the prospect says something nobody prepared for.
The rep pauses.
Maybe they start talking faster. Maybe they return to the product pitch. Maybe they ask a question that has little to do with what the buyer just said. Sometimes they simply say, “That’s a good question,” and reach for the next line they remember.
The conversation has changed.
The training has not.
That is the part worth examining.
I have become skeptical of the usual explanation that these reps simply “lack confidence.” Confidence matters, but it is often the visible symptom. The deeper problem is that the rep has practiced retrieving a correct answer while the buyer is asking them to make a judgment.
Those are different skills.
A script works until the buyer stops following it
Scripts are useful.
A new rep needs some structure. Without it, the first few weeks can become a string of avoidable mistakes. The problem begins when the script becomes the training rather than the starting point for training.
Imagine a new SDR has practiced this sequence:
- Introduce yourself.
- Explain why you’re calling.
- Ask about the prospect’s current process.
- Identify a pain point.
- Schedule the next meeting.
It sounds reasonable.
Then the prospect says:
“We’re actually pretty happy with what we’re using. Why would I change?”
That sentence isn’t in the sequence.
Now the rep has to decide what the prospect actually means.
Are they genuinely satisfied?
Are they protecting an incumbent vendor?
Are they trying to end the call?
Do they have a problem but don’t consider it urgent?
Is another stakeholder driving the decision?
Those possibilities require different responses.
A memorized answer cannot reliably solve that problem because the first task is interpretation.
That distinction shows up in current sales-training guidance too. Salesforce’s objection-handling material recommends clarifying what sits behind an objection rather than assuming the first statement tells you everything.
That is a judgment skill.
It has to be practiced as one.
The uncomfortable version of role-play
Most sales organizations already know role-play is useful.
The problem is how they use it.
A manager announces, “Let’s practice objections.”
One rep becomes the seller. Another becomes the prospect. The manager gives the prospect an objection. The rep answers.
Everyone nods.
Then they move on.
The exercise looks productive.
Sometimes it isn’t.
A role-play becomes weak when the buyer is too cooperative, the objection is announced beforehand, and the rep knows the conversation is heading toward a predetermined answer.
The rep can become good at the exercise without becoming better at selling.
Current AI role-play guidance makes the same distinction. Useful simulations introduce buyer resistance, changing circumstances, and multiple stakeholders rather than simply walking a rep through a fixed script.
A useful practice conversation should occasionally make the rep uncomfortable.
Not humiliated. Uncomfortable.
The buyer should interrupt.
The buyer should misunderstand something.
The buyer should change the subject.
The buyer should say the competitor is cheaper.
The buyer should ask for a discount before explaining the problem.
And sometimes the buyer should be right.
That last one gets missed.
What actually happens when a rep freezes?
There are several different failure modes, and treating all of them as “confidence” makes coaching less useful.
1. The rep goes back to the script
This is probably the easiest one to spot.
The buyer asks an unexpected question and the rep returns to a memorized product explanation.
You can almost hear the training module playing in the background.
The rep isn’t necessarily panicking.
They may simply be searching their memory for the next approved response.
2. The rep answers too quickly
This one is harder to catch because it can sound confident.
The buyer says:
Your pricing is higher than what we’re paying today.”
The rep immediately explains why the product is worth the price.
But they haven’t established what “higher” means.
Higher by how much?
Compared with what?
Is price actually the concern?
What would make the investment worthwhile?
The rep has responded to the sentence instead of the situation.
That is a very different mistake.
3. The rep becomes defensive
This is particularly common with competitor objections.
The buyer says:
We’re already using another platform.”
The rep hears:
Your product isn’t good enough.”
So they start explaining why their company is better.
That can turn a useful conversation into a contest.
A better response may be a question:
“What is working well with your current platform?”
Now the buyer has to tell you something.
Maybe the incumbent is genuinely good. Maybe there is no opportunity. Or maybe the answer exposes the exact weakness that created the buying conversation.
The objection was information.
The rep treated it as an attack.
4. The rep starts talking
A lot.
This is the one managers sometimes mistake for effort.
The buyer raises a concern and suddenly the rep gives a three-minute explanation.
More features.
More proof points.
More customer stories.
More product detail.
The rep feels productive because they are saying useful things.
The buyer may be thinking, “I still haven’t been heard.”
That is a training problem.
More product knowledge won’t necessarily fix it
This is where many onboarding programs take a wrong turn.
A rep struggles with unexpected questions.
The organization responds with another product-training session.
The rep struggles with objections.
The organization adds another objection-handling document.
The rep struggles with discovery.
The organization gives them a better discovery checklist.
Sometimes those interventions help.
But if the failure happens because the rep cannot adapt during an unpredictable conversation, more information can actually give them more material to search through mentally.
The skill is closer to:
Listen → interpret → decide → respond → check
That sequence is harder to teach because there isn’t always one correct sentence.
There is a correct process.
Recent AI sales role-play research and guidance increasingly focuses on this distinction. The value of simulation is not simply having an AI recite buyer objections. It is creating a conversation in which the buyer reacts to what the rep actually says, allowing the rep to practice discovery, objection handling, negotiation, and recovery.
That is a better definition of practice.
Practice needs uncertainty
This is the first thing I would change in a sales training program.
Don’t only practice the objection.
Practice the uncertainty around the objection.
Suppose the scenario is:
Buyer: “Your price is too high.”
A weak exercise asks the rep to respond to that sentence.
A stronger exercise changes what happens next.
Version A
Buyer: “Your price is too high.”
Rep asks a question.
Buyer explains that the company has a strict budget.
Now the rep has to explore budget and value.
Version B
Buyer: “Your price is too high.”
Rep asks a question.
Buyer reveals that the real concern is implementation risk.
Now the conversation moves somewhere else.
Version C
Buyer: “Your price is too high.”
Rep immediately starts defending price.
Buyer says:
We’re not even comparing you on price. We’re worried about whether our team will use it.”
Now the rep has to recover.
That recovery is the useful part.
The buyer should have a reason for pushing back
This is a subtle difference between a demonstration and a real drill.
Many role-play examples are written from the seller’s perspective.
They tell the rep what objection is coming and then give them a response.
I would reverse that.
Brief the buyer.
Give the buyer a reason to resist.
Give them information they will reveal only if the rep asks the right question.
Give them a rule for what makes them more receptive and what makes them more skeptical.
Then let the rep find the path.
That approach is appearing more frequently in current sales-roleplay design. Recent scenario guidance argues that a useful objection exercise should define the buyer’s hidden motive and conditions under which the buyer becomes more or less receptive, rather than giving the seller a predetermined conversation.
That matters because real prospects have motives.
They don’t wake up wanting to deliver objection number seven from the sales-training worksheet.
AI role-play can solve one very practical problem
There is a reasonable argument for human role-play.
A good manager knows the business, understands the rep’s weaknesses, and can introduce nuance that a simulation may miss.
That remains valuable.
But there is another problem:
Availability.
A manager cannot realistically role-play every objection with every rep every day.
Current sales-training research and vendor reports consistently identify this as one of the practical reasons teams are turning to AI simulations: reps can practice on demand without waiting for a manager or peer to become available.
That creates an important distinction.
AI doesn’t have to replace the manager to be useful.
It can increase the number of times a rep gets to practice before the manager needs to intervene.
That is a much more defensible claim.
But AI role-play can recreate the same problem
This is where the category needs some skepticism.
If the AI simply says:
“Prospect: It’s too expensive.”
and waits for the rep to deliver the approved response, we’ve recreated the original training problem.
The rep is still practicing a predetermined conversation.
A better simulation might work like this:
Buyer: “I’m not sure this is worth the money.”
The rep asks:
“What makes you feel that way?”
The buyer responds:
“Honestly, our team tried something similar last year and nobody used it.”
Now the rep has a new problem.
The price objection was actually an adoption objection.
That is the kind of turn that matters.
The rep should have to discover it.
Recent AI role-play platforms increasingly describe this adaptive behavior as a core feature, with simulations changing their responses based on what the rep says rather than following a fixed transcript.
That is the difference between a chatbot and a training exercise.
Where Practis.ai fits
This is one of the more natural use cases for Practis.ai.
The value isn’t simply that an AI system can play the buyer.
The more interesting question is whether a rep can repeatedly practice the moments that tend to break down in real conversations.
For example:
- A pricing objection that turns into an implementation concern
- A competitor objection that reveals incumbent loyalty
- A discovery call where the buyer refuses to answer questions
- A prospect who repeatedly interrupts
- A skeptical executive who wants the answer immediately
- A buyer who asks for a proposal before explaining the problem
- A conversation where the rep gives a poor answer and needs to recover
The point is not to produce a perfect script.
It is to make the rep less dependent on one.
Practis can sit in that practice layer, giving reps another environment to rehearse realistic conversations and work through difficult moments before those moments occur with a real buyer.
That is where I think the product category makes the most sense.
Not as another content library.
As rehearsal.
The better test for readiness
I would stop asking:
“Can this rep handle the standard objection?”
That test is too easy to game.
Ask instead:
“What happens when the objection isn’t what the rep thinks it is?”
That’s a much better test.
Give the rep the same opening three times.
Change the buyer each time.
On the first attempt, the prospect says price is the problem.
On the second, price is a proxy for risk.
On the third, the buyer simply doesn’t believe the claimed outcome.
If the rep gives the same answer three times, they have learned the objection.
If they investigate what is underneath it, they are learning how to sell.
There is a difference.
A practical test managers can run this week
You don’t need a new training program to test this.
Take the last ten calls where a deal stalled because of an objection.
Write down the exact buyer language.
Don’t clean it up.
Keep the awkward wording.
Then choose three of those conversations and role-play them with a rep.
But don’t give the rep the objection in advance.
Start the conversation normally.
Introduce the objection naturally.
Then add one unexpected turn.
If the rep handles the first objection well but collapses after the second turn, you’ve found something useful.
That is a coaching signal.
And you have something better than a generic note saying the rep “needs more confidence.”
You know where the practice needs to happen.
Recent sales-training guidance similarly recommends building role-play scenarios from real calls and repeating the specific moments where reps struggle rather than relying only on broad training exercises.
That is a much better use of training time.
One thing I would not claim yet
I would not claim that AI role-play automatically produces better sales performance than manager-led coaching.
The evidence is not strong enough for that statement.
What I am more comfortable saying is narrower: AI can make realistic repetition easier to provide at scale, and repetition is difficult to provide when every practice session depends on a manager’s availability.
That distinction matters.
I would also be careful with claims about direct revenue impact. Some vendors report improvements in readiness, ramp time, or performance, but those results are not automatically transferable from one organization to another.
The more interesting question is whether practice performance predicts what happens on actual calls.
That is the measurement I would want to watch over the next two quarters.
The goal isn’t to eliminate the script
A good rep still needs structure.
They need messaging.
They need product knowledge.
They need a sales methodology.
They need to understand the buyer.
But those things should give the rep something to work from, not something to hide behind.
When the buyer says something unexpected, the rep needs to leave the script without losing the thread of the conversation.
That is the skill.
And if a sales team keeps seeing reps freeze when buyers go off script, I would look at the practice environment before blaming the reps.
The uncomfortable possibility is that the team may have trained them exactly as instructed.
The buyer simply didn’t cooperate.
And that is where the next training experiment should begin.