Move the Middle: What a 2X Buyer Learned About AI Sales Training

Sales leaders have never had more tools, more data or more pressure. Yet one of the biggest revenue levers sits in plain sight: coaching middle performers. AI can draft emails and score pipeline, but the skills behind a great discovery call still decide who hits quota.

That tension shaped the keynote Yoodli CEO and co-founder Varun Puri gave at CRO Exchange with a guest revenue leader who has rolled out Yoodli at two companies. He brings 20 years of go-to-market leadership and a buyer’s view of what AI sales training actually delivers.

Their core argument is simple. The craft of complex sales is more important than it’s ever been, and AI should help people master that craft, not replace them. Here’s how they break it down.

A small group of sellers carries most of the revenue

Most sales teams run on a handful of stars. The Ebsta and Pavilion 2024 B2B Sales Benchmarks, which analyzed 4.2 million opportunities across 530 companies, found that the top 17% of sellers generate 81% of revenue. The other 83% of the team brings in just 19%.

McKinsey points to the same gap from another angle: the top 30% of reps can outperform the bottom 30% by as much as a factor of four.

Those numbers expose a risk and an opportunity at once. When a few reps carry the number, one resignation can sink a quarter. But the gap also shows how much revenue sits untapped in the rest of the team. So where should a sales leader spend coaching time to close it?

The meat is in the middle: why coaching middle performers pays off

The guest speaker frames sales coaching with a hamburger. Picture your team as a burger: top performers are the top bun, bottom performers are the bottom bun, and your B players are the patty in the middle.

Most leaders pour their energy into the buns. They study what the stars do, and they spend hours trying to rescue the reps who struggle. He allocates his time differently:

• Top performers: Learn what works from them, then let them run. They need less day-to-day attention.
• Middle performers (B players): Focus practice and coaching here to build consistency. This is where coaching pays off fastest.
• Bottom performers: They get lots of attention, but progress stays uncertain.

The middle is where the meat is. Coaching middle performers even part of the way toward your best reps lifts the whole number, and it reduces your dependence on a few stars. That raises the real question Varun and his co-presenter put to the room: how do you move the middle?

AI matters only if it moves your revenue goals

Before any tool earns a place in the stack, Varun’s co-presenter judges it against three outcomes every revenue leader owns:

1. Time to ramp: Get new reps productive faster.
2. Pipeline: Generate more opportunities, and better-qualified ones.
3. Win rates and quota attainment: Get more reps hitting their number.

If an AI tool can’t show impact on at least one of these, it’s a distraction. That lens matters because the problems he saw across several companies all traced back to these same three metrics.

Why traditional sales training can’t close the gap

Across multiple companies, he ran into the same six problems:

• Slow ramp: New reps took longer to produce than the financial model predicted.
• Competitive market: Deals kept getting harder to win.
• Complex subject matter: Reps struggled to master what they were selling.
• Weak qualification: Reps had no way to practice without doing it live, and they lost deals learning on the job.
• No coaching capacity: Enablement and managers lacked the time to fix gaps he already knew about.
• Peers coaching peers: Reps coached each other, with no way to know the quality of that coaching.

Traditional sales training doesn’t solve these problems. Most teams still rely on one-time trainings, seminars, prerecorded videos and occasional human roleplays. They then measure success with participation rates, CSAT scores and qualitative feedback, which correlate only loosely with actual outcomes.

In short, the old model measures completions, not skill improvement. A rep can finish every module and still fumble the first pricing objection. Closing that gap takes a different approach to sales training, one that measures what reps can actually do.

What AI sales training looks like when it works

Yoodli, the AI platform for human skills, replaces one-and-done training with continuous practice and measurement. Reps rehearse discovery, objection handling and pitches in realistic AI roleplays. Then an AI coach meets them in the flow of work, and managers see skill progress tied to business results.

The platform brings five capabilities together:

• Create: Build multimodal learning journeys from a single prompt.
• Roleplays: Run realistic practice for certifications, tool training, partner enablement and everboarding.
• Personas: Let learners work with AI-driven experts in roleplays or on the job.
• Coach: Give every learner a personal coach connected to conversation intelligence, CRM, calendar and messaging apps.
• Measure: Track skill development and tie it to business outcomes.

Enterprise teams already see results across those three revenue metrics:

Customer	Use case	Result
ServiceNow	Ramp time	Cut rep ramp from 3 months to 6 weeks
Google Cloud	Certification	Enabled 15,000 reps org-wide
$5B+ cybersecurity company	Skill diagnosis	29% uplift in asking questions
Snowflake	Manager coaching	Saved 1,200+ manager hours per quarter
USB Payments	Prospecting	19% more meetings within 30 days
RingCentral	Support certification	Cut certification time by 90%
Sandler clients	Pipeline	$6–8M combined pipeline in 10 weeks

Those results show what becomes possible when practice is measurable. Still, Varun and his co-presenter were clear that AI is only part of the answer.

Why sales teams still need humans

AI accelerates skill building, but it doesn’t run a sales org on its own. The speakers flagged three limits every leader should plan for:

• Behavior change is hard. People still default to the ways they’ve always learned, so leaders need to model and reinforce new habits.
• People hold AI to a higher bar. Reps get more frustrated when AI glitches than when a human coach makes a mistake.
• AI shouldn’t make the decisions. AI works well for training, but teams shouldn’t use it alone for screening or evaluating people.

That’s why Yoodli builds technology to help people, not replace them. Managers still set the standard and make the calls. AI gives every rep the reps they need between those moments.

How to choose an AI sales training partner

Drawing on two Yoodli rollouts, the guest speaker offered a short checklist for evaluating vendors:

1. Works with your ecosystem. Ask how the platform connects with your sales stack and your existing trainers.
2. AI-first by design. Look for a product built on AI from the start, not one with AI bolted on.
3. Enterprise-grade security. Confirm it fits your privacy and compliance requirements.
4. A long-term vision. Choose a partner whose roadmap grows with your AI strategy.

The takeaway for revenue leaders is clear. Stop pouring all your coaching into the top and bottom of the team. Start coaching middle performers with consistent, measurable practice, and you’ll see it in ramp time, pipeline and win rates.

Frequently asked questions

How can AI sales training reduce rep ramp time?

AI sales training lets new reps practice real conversations, like discovery calls and objection handling, before they talk to buyers. Reps get instant feedback and repeat scenarios until they meet the standard. ServiceNow cut rep ramp time from three months to six weeks with Yoodli.

Why is coaching middle performers the best use of coaching time?

Top performers need little day-to-day coaching, and bottom performers often show uncertain progress despite heavy attention. Coaching middle performers pays off fastest, because B players respond quickly to focused practice. Moving them closer to your best reps lifts total revenue and reduces dependence on a few stars.

What’s wrong with traditional sales training?

Traditional sales training relies on one-time workshops, seminars and prerecorded video. Teams measure it by participation and satisfaction scores, which correlate only loosely with revenue. It tracks completions instead of measurable skill improvement.

Can AI replace sales managers or coaches?

No. AI scales practice and feedback, but managers still set the standard, reinforce new habits and make decisions about people. Teams shouldn’t use AI alone for screening or evaluation.

What should you look for in an AI sales training platform?

Look for a platform that works with your sales stack and trainers, is built AI-first, meets enterprise security requirements and has a roadmap that grows with your AI strategy. Learn more in our guide to AI sales training or see how Yoodli supports sales enablement teams.

Want to see how it works for your team? Try Yoodli for free or book a demo with our team.