A CEO told me last month that his team was missing quota because the reps “just need reps.” I asked him how often his VP of Sales was running live roleplay drills. He laughed. Nobody on the leadership team had 30 minutes a week, per rep, to practice. The reps were getting their “reps” on live buyer calls. Paying customers were the practice field. Missed quota was the scoreboard.
This is the most common pattern I see inside SMB and mid-market sales teams. Managers can’t find the time. Reps don’t practice. They wing it on live calls. They miss. They get blamed. Everybody wonders why the number is stuck.
In this article I will name the practice deficit, walk through the two tiers of AI roleplay that have rewritten the math (packaged platforms and build-your-own), give you an honest read on what Sandler and Challenger are doing, tell you why I open every Vistage talk with a live ElevenLabs demo, and finish with a four-step Monday diagnostic you can run on your team this week.
Practice is a leadership output. It is not a tool purchase. By the end of this you will see the difference.
The Practice Deficit: Why Reps Wing It Live
Most sales managers want to coach. They know they should. They have been told their entire career that the best managers are the ones who develop people, not the ones who chase deals.
Then Monday hits. Forecast call at 8. Pipeline review at 10. One-on-ones bleed into deal reviews. Someone’s biggest account is on fire by noon. The afternoon fills with internal meetings, customer escalations, and the VP’s board prep. The manager ends the day having done zero actual coaching. The rep ends the day with three more objections they did not practice handling.
Multiply that by forty weeks and eight reps. Do the math. Most managers I work with are getting maybe five to ten minutes of roleplay in per rep per month. That is a rounding error. That is not coaching. That is the absence of coaching dressed up as “too busy this week.”
The reps notice. The good ones figure out their own practice, usually by recording themselves and replaying calls on the drive home. The rest just show up and wing it. Buyer objections go unanswered. Discovery questions get forgotten. Negotiation moves get improvised. The first time a rep hears themselves fumble an objection is usually the first time a buyer hears it, which is also the first time the deal dies.
Traditional ramp for a new seller is 6 to 12 months to average performance. A big chunk of that ramp is just the rep accumulating live at-bats because there is no structured practice environment. You are paying full salary for 6 to 12 months of on-the-job training that your customers are unknowingly funding.
This is the practice deficit. It is a leadership problem, not a rep problem. Reps do not decide whether the team practices. Leaders do.
What AI Roleplay Actually Does
Picture a Monday morning. A rep sits down at her desk at 7:45, fifteen minutes before her first discovery call. She opens an AI roleplay tool. It has her ICP loaded. It knows the industry. It knows the three objections this type of buyer throws in the first ten minutes.
She runs a five-minute warmup. The AI plays the prospect. It interrupts her. It pushes back on her discovery questions. It throws the “we already have a vendor we like” objection that she flinched on last Friday. She works through it. The AI scores her on messaging coverage, speaking pace, filler words, and objection handling. It flags two things she did well and one thing to fix.
She closes the tool. She takes the call. She nails the objection.
That is AI roleplay in practice. It is not a video library. It is not a learning management module. It is a back-and-forth conversational simulation that adapts in real time, scores performance against a rubric, and gives the rep something to fix before the next live interaction.
The data on what this does to a team is serious. AI-coached teams report 300 to 500 percent ROI in year one. New hires hit full productivity 30 to 40 percent faster, which cuts traditional 6 to 12 month ramps down to 3 to 6 months. AI conversational coaching has been shown to increase call conversion rates by roughly 30 percent. Those are not marketing numbers. Those are the results teams see when practice stops being an optional manager favor and starts being a structural daily habit.
The reason it works is boring and obvious. Reps who practice get better than reps who do not. That has been true since the first sales team existed. The only thing that changed is that the practice field no longer requires a manager’s calendar.
The Two Tiers: Packaged Platforms vs. Build-Your-Own
The AI roleplay market is splitting into two tiers. Any leader who wants to make a smart call needs to understand both, because they solve different problems.
Tier one is packaged platforms. These are purpose-built tools that work out of the box.
Second Nature is the current market leader in voice-based AI roleplay. Its AI conducts real back-and-forth conversations, throws objections, adapts to rep responses, and scores on messaging coverage, speaking rate, filler words, and objection handling. Heavy use in onboarding.
Hyperbound is strong on daily warmups and objection drills. It creates practice scenarios from your actual calls and supports custom AI scorecards aligned to any methodology (MEDDIC, BANT, Challenger, Sandler). A standout feature is industry-specific objections built from real ICP data, not generic ones.
PitchMonster ships with 48 ready-to-use scenarios covering pitch skills, discovery, and objection handling. Good for teams that want plug-and-play without heavy customization.
Kendo AI lets you build AI prospects matching your exact ICP, with the specific objections your reps actually face. It supports both B2B and B2C and lets you target practice objectives (objection handling, discovery mastery, rapport building, demos, urgency creation).
These tools are excellent at what they do. A team of 30 reps can onboard on Hyperbound in two weeks and see measurable ramp acceleration in 30 days. If you want a proven system without engineering work, start here.
Tier two is build-your-own. This is where leaders who want a real edge are starting to spend time.
ElevenLabs has built the most advanced conversational voice AI on the market. 10,000+ expressive voices, emotional intelligence that detects urgency and hesitation in the rep’s voice and adapts in real time, the ability to clone voices so the AI sounds like an actual accent your reps will encounter in the field, and an agent-building API that lets you design custom scenarios around your real buyer personas, your real competitive landscape, and your real sales motion.
The reason this matters is that packaged platforms give you what every other team using that platform has. Build-your-own gives you something your competitors cannot buy. You can build a CFO persona who responds exactly the way your top three lost deals responded. You can load your actual competitive objections, word for word. You can design a simulation where the AI represents your number-one competitor’s sales rep and your team practices head-to-head discovery. None of that exists off the shelf.
The right answer for most teams is both. Use a packaged platform for baseline daily practice and new hire ramp. Build custom agents for the three or four scenarios that are unique to your business and are costing you deals. Leaders who understand both tiers can make that call. Leaders who only know Tier One buy a subscription and call it a strategy.
What Sandler and Challenger Are Actually Doing
I get asked all the time whether the incumbents have caught up. Short answer, no. Honest answer with detail, they are moving, but they are bolting AI onto frameworks that were designed 30 years ago.
Sandler launched the Sandler AI Roleplay Coach, powered by Yoodli. It runs AI-driven scenarios built on Sandler behavioral methodology with real-time feedback. Available standalone or inside Sandler certification. The tool is fine. It works. The problem is what it represents. Sandler is using AI to reinforce the Sandler model, not to rebuild training around AI. Their Summit 2026 agenda features HubSpot’s CEO and marketing experts. No AI-native sales leadership voices on the main stage. AI is a supplement. It is not the method.
Challenger and Richardson launched AccelerateAI in 2025. Scenario-based video challenges simulating buyer conversations, powered by Richardson’s Accelerate Sales Performance System. Challenger also embedded its AI framework natively into Gong and developed AI Smart Trackers that align conversation data to Challenger methodology. This is the most sophisticated AI integration I have seen from a legacy training provider. The Gong embed is genuinely smart. But the core program is still Challenger methodology with AI as an accelerant. If you believe Challenger is the right methodology for your buyer in 2026, great. If you think the buyer has changed since the Challenger book came out, you are paying for a coat of AI paint on a 30-year-old house.
None of this is meant as a knock. Sandler and Challenger built real IP. They know how to train. But their AI investments are defensive, not generative. They are reinforcing the old model. They are not building the new one.
The gap that creates for a practitioner leader is the same gap I write about in Motion Is Not Progress. The tool does not fix the system. If your methodology was designed in 1995 and your AI layer was added in 2024, your team gets a faster, louder version of a thirty-year-old framework. That is not transformation. That is a reskin.
The ElevenLabs Advantage: Why I Demo Live
I open every Vistage talk the same way. I put an ElevenLabs voice agent up on the screen, configured as a B2B CFO with a specific objection profile. I invite a CEO in the room to try to sell her. The CEO pitches. The AI CFO pushes back in a voice that sounds like a real person. The CEO gets visibly uncomfortable. The room leans in.
I do this for a reason. Ninety percent of CEOs have heard about AI roleplay. Maybe ten percent have watched a demo. Almost none have experienced it. Once they hear a voice push back on them in real time, the concept stops being abstract. They feel the practice deficit in their bones, because they themselves just fumbled a CFO objection in front of their peer group.
None of the other training providers can do this. Yoodli is video-based and async. AccelerateAI is scenario-based and prerecorded. The packaged voice platforms work well inside a team’s training workflow, but they are not designed for live stage demos with custom scenarios built in front of the audience. ElevenLabs is.
The reason that matters for a sales leader is not that I get to do a cool demo. It is that the same capability that makes a Vistage room go quiet is the capability your reps need every Monday morning. A custom voice agent that sounds like your actual buyer, pushes back on your actual objections, and adapts in real time. That is the competitive edge. That is why build-your-own matters.
If you are a CEO or VP of Sales thinking about how to give your team a practice environment competitors cannot copy, the ElevenLabs layer on top of a packaged platform is the move. I help companies build exactly that inside CASL, and I demo it live when I speak to Vistage groups.
Building Practice Into the Culture: The Monday Diagnostic
Here is what most leaders get wrong after they see a demo. They buy the tool. They roll it out in a team meeting. They tell the reps to “use it when you have time.” Thirty days later, nobody is using it. Six months later, they renew the subscription out of guilt and declare AI roleplay “not a fit for their team.”
The tool is not the culture. The tool enables the culture. The culture is built by the leader.
Run these four steps on Monday morning.
Step one. Commit to a cadence. Pick three things. A daily five-minute warmup before the first call of the day. A weekly 20-minute objection drill every Wednesday on a specific objection the team is losing on. A monthly pressure simulation where a rep runs a full discovery against a custom AI buyer while the team watches. Write those three cadences on the wall. Tell the team this is the new normal.
Step two. Review the scores. If the AI gives you a scorecard, look at it. Weekly. Pipe it into your one-on-ones. A rep with a consistent 60 percent objection handling score does not need more pep talks. She needs a specific conversation about what she is missing in the first twenty seconds of a “price is too high” response. The data is there. Use it.
Step three. Make practice public. Run a weekly roleplay review where one rep’s recorded AI session plays in the team meeting. Celebrate the wins. Dissect the misses. Peer learning compounds. Private practice is better than no practice. Public practice is better than private practice.
Step four. Measure what changed. Track three metrics 30 days before and 30 days after you install the cadence. New hire ramp time. First-call conversion rate. Objection handling score. If those numbers do not move, something in the cadence is broken and you need to fix it. If they do move, you have the business case to double down.
That is the work. None of it is technical. All of it is leadership. You cannot outsource it to a tool. You cannot delegate it to the VP of Sales and hope it sticks. The cadence has to come from the top, and it has to survive the week where everyone is “too busy.”
If you want peer accountability on making this stick, the Sales Leadership Forum runs a monthly session specifically on building an AI practice culture. If you want to install the full system inside your team with module-by-module guidance, CASL walks you through it. If you want to see the ElevenLabs live demo in your Vistage room or company offsite, book a speaking engagement.
Your reps are not practicing enough. That is on you. The tools exist. The data is real. The only question is whether you are going to build the culture or keep blaming the reps for missing a quota they never got a chance to rehearse for.