← All Episodes

Neha Gupta joins Warren Kucker to discuss A New Approach to Sales Roleplay. Key takeaways include Off-the-shelf chatbots are built to be agreeable, so they stop pushing back after a few exchanges; realistic practice needs personas that hold their ground. and Traditional role play fails because practicing in front of a room is embarrassing and colleagues avoid pushing hard on each other..

EP. 28 32 min July 29, 2026

A New Approach to Sales Roleplay

with Neha Gupta

"What's nice about AI now is that it can be pushy. It doesn't care about your feelings. It can say no."

About This Episode

Neha Gupta, founder and CEO of Breakthrough AI, joins Warren Kucker to discuss how AI is changing sales role play. She explains why traditional role play and off-the-shelf chatbots fall short, how fine-tuned buyer personas and an objective scoring rubric make practice realistic and consistent, and how to roll out a practice program that reps actually use. The conversation also walks through rehearsing a 10 percent discount request and where role play is heading next.

About the Guest

Neha Gupta

Neha Gupta is founder and CEO of Breakthrough AI, a practice-first coaching platform where sellers rehearse hard sales conversations with AI buyer personas. She spent 25 years in consulting and consumer brands, including several years at Unilever and Clorox, and later worked with mission-driven brands.

Key Takeaways

  • ✓ Off-the-shelf chatbots are built to be agreeable, so they stop pushing back after a few exchanges; realistic practice needs personas that hold their ground.
  • ✓ Traditional role play fails because practicing in front of a room is embarrassing and colleagues avoid pushing hard on each other.
  • ✓ Sales is one of the few disciplines where people walk into the most important conversations without rehearsing them.
  • ✓ An objective AI rubric scores every rep on the same scale and removes manager bias and calibration sessions.
  • ✓ Rehearsing a discount request builds the muscle memory to defend value instead of defaulting to a concession.
  • ✓ Make practice an event with onboarding, leaderboards and competitions, and run it for a fixed window of about six weeks.
  • ✓ The next frontier for AI role play is detecting tone, not just the words a rep uses.

Chapters

00:00 Introduction to Breakthrough AI
01:18 Neha Gupta's Career Journey
03:41 The Evolution of Role-Playing
05:17 Role-Playing in Sales
29:01 The Future of Role-Playing in Sales

Show Notes

Selling is a skill most people never rehearse

Neha Gupta, founder and CEO of Breakthrough AI, started her career in consulting, where she was put in front of skeptical clients early and had to learn quickly how to present to them. She did not grow up around corporate speak, so she learned the unwritten rules by watching others and picking up tips from peers. She then spent several years at Unilever and Clorox, where she found that selling is not only external. Inside a large organization, pushing an idea or a new product direction is its own sales job.

The through line, Gupta says, is how much communication skills shape a career. She repeatedly saw technically capable people who could not get their ideas through, then wondered why they were not being promoted. School does not teach these skills, and new hires are expected to take years to ramp up. Kucker adds that most of what people do at work is sell their ideas, even if they do not see it as selling.

Asked why reps should role-play at all, Gupta describes sales as a form of conflict. You are asking someone to do something for you, you are exposed, and they might simply say no. A few charismatic people handle that naturally, but most need practice, and she believes it is a skill anyone can build. Nobody puts a soccer player on the field without practice, yet people spend 40 to 80 hours a week at work and rehearse almost none of it.

Kucker sharpens the point. Engineers do not write code in production, a CFO does not assemble the financials during a board meeting, and marketing does not one-shot its materials. Salespeople walk into a $500,000 deal with general methodology and experience, but no rehearsal of the conversation that matters most.

Off-the-shelf chatbots make poor sparring partners

Many people already role-play with ChatGPT or Claude. The problem, Gupta says, is that general-purpose models are tuned to keep you engaged.

they're built to be agreeable with you. They're trying to make you like them

A real buyer will push back, so Breakthrough AI has spent the last 12 months fine-tuning 11 personas, including a nice one for easy starts, a demanding one, an analytical one, a bargainer and a novice who lacks context. Gupta says that if you tell ChatGPT to act as a demanding customer, it holds the role for one or two exchanges and then turns agreeable again. Role plays are built around a customer's own company, product and culture, and users often say they got surprisingly emotionally invested.

She compares this to athletes and musicians, who do not just practice a lot but practice the hard things they are not good at yet.

Why traditional role play rarely happens

Role play is not new. The classic format puts two people in front of a workshop while everyone watches, with an instructor stopping them to give feedback before they retry. Gupta says that is the best version of it, and it is still awkward and sometimes humiliating. One early customer tried assigning reps to practice in pairs and report back. The reps ended up talking about what they would say instead of actually doing the role play, because playing a demanding client against a colleague strains a relationship you are trying to keep.

What's nice about AI now is that it can be pushy. It doesn't care about your feelings. It can say no.

Kucker adds the moderator problem. The person playing the buyer has to be inventive on the spot, and in almost every session they say something no real VP of customer success would say, which breaks the flow. He argues AI is better at this because it is far more inventive in conversation, and nobody has to play moderator.

An objective rubric removes the calibration problem

The second advantage Gupta describes is scoring. Breakthrough AI built its rubric from academic research on what makes a good communicator and salesperson. A manager coaching a rep brings their own biases and their own idea of what works, and large companies spend hours in rooms calibrating managers who grade easily against those who grade hard.

With AI, everyone is measured on the same scale. A feedback agent reads the transcript and looks for micro behaviors: whether the rep acknowledged the buyer's concerns, brought back specific data or value points, ended with momentum through timelines and next steps, explained the product clearly, rambled, or talked more than they listened. Gupta says that the more precisely those behaviors are defined, the more objective the feedback becomes. Managers can see results across every session and spot where a rep is struggling, which would be a manual chore if each rep ran their own prompts in ChatGPT and nobody knew what they had actually practiced.

Rehearsing the 10 percent discount request

In Breakthrough AI's pricing scenario, the rep first gets a short framework so a new seller does not practice bad habits. Then the buyer says everything looks great but asks for a 10 percent discount, with no problem in the proposal behind it. The rep's job is to hold off on giving in and talk about the value, the expertise and the reasons behind the price. If the rep does the right things, the agent eventually agrees to move forward, so the rep experiences what success feels like.

Kucker puts a number on it: each discount conversation has a dollar value, so a team can estimate what an untrained 10 percent concession costs across every deal. Most sales leaders cover pricing once during onboarding. Without practice, he says, about 80 percent of reps fall back on the same line about not being authorized and needing to check with their boss. His framework is to ask whether the buyer is working to a set budget, does not see the ROI discussed on the last call, or is simply being a good steward of company money, which he finds disarming.

That doesn't become a learned behavior. The default behavior becomes the behavior.

Gupta calls this muscle memory. Reps who have rehearsed come in calm and confident instead of desperate to close, and in Breakthrough AI's last pilot, participants reported feeling 40 percent less anxious after the role plays.

Make practice an event, then end it

Rolling out a tool and saying little about it does not get usage. Gupta recommends treating it like an event, with an onboarding session people are invited to, leaderboards that show how much peers are practicing, and competitions. The most successful pilot turned it into a contest with streaks and levels, where the buyer gets more demanding as the levels rise. She also argues for a fixed training window of around six weeks because one-time workshops fade and self-paced video training rarely gets done.

People don't want to train forever.

Kucker's goal is for role play to become part of how a team warms up before calls rather than something extra. Gupta practices on her own product too. The feedback told her she talks too much, so she is working on shorter answers, more discovery questions, and playing a buyer's own words back to them. She also practices different styles, such as hunter, farmer and gatherer, depending on whether a lead is warm, a mutual connection or completely cold.

Looking three years out, Gupta expects more realism, starting with video, which she currently finds too clunky. She is most interested in AI that detects tone, such as whether a rep sounded sarcastic or unenthusiastic. Kucker adds a caution: the value comes from repeated exposure to a consistent tone and learning a response that lands, not from throwing random tones at a rep within a single call.

Full Transcript Show

00:16 Warren Kucker: Hey there, and welcome to another episode of Selling AI. Today we're diving into something every sales leader claims to love and every rep secretly dreads. This is sales role-playing. I've got Neha Gupta on the show, founder and CEO of Breakthrough AI, a practice-first coaching platform where you rehearse your hardest sales conversations with an AI that actually kind of pushes back with the right personality traits. Naha spent 25 years in consumer brands and consulting, learning the unspoken rules of professional communication the hard way, and she turned that into a tool that treats Soft skills as measurable business capabilities. In today's episode, Naha and I get into why most AI roleplay becomes outdated quickly, and actually whether empathy is something you could practice and score well on. alright, let's get to it. Hey Neha, thank you so much for being here on the show today to talk about the future of role-playing in sales. Welcome to the show.

01:07 Neha Gupta: Thank you, my pleasure. I'm excited for our conversation.

01:10 Warren Kucker: same here. I'm excited to dive deep into this. This is on it's close to my heart. Cool. So listen, give us a little bit about your background. Tell us about your career arc, twenty-five years in consumer brands, consulting, mission driven work, before starting Breakthrough AI, tell us about your career and what was the through line to why you started started Breakthrough.

01:28 Neha Gupta: Yeah, so I started in consulting where I got thrown in front of clients pretty quickly and had to really very quickly learn how do you actually, you know, present something to a client that may be skeptical of what you're presenting. And that was a really hard learning curve, honestly, very steep. I mentioned in my bio that I'm a first generation, you know, corporate employee. So I didn't really grow up with a lot of like, here's how you speak corporate speak, or even, you know, how do you be at a client dinner or something like that. So I had to do a lot of like observing and seeing what others were doing and kind of get tips from my fellow peers and all of that in order to really figure it out. I then went into consumer brands. I worked at Unilever and Clorox for several years. And again, kind of learned internally, these are huge organizations. You're learning internally how do you actually sell something like an idea you have or a new direction for a product. And so sales to me is not just external, but also internally, there's so much sales that needs to happen, especially if you really want to make an impact. and then last, I try to take all of that sort of brand experience and marketing experience and bring them sort of mission-driven brands where they could then they might have a great product or a service and they want to really be able to reach more people. Everybody wants growth at the end of the day. And so, how can we bring those same ideas to small organizations that might not be able to afford, you know, a full time brand manager? And I would say the through line through all of this was that because I've been in front of customers and clients for so long, but then also internally needing to learn how to sell something is. That I really became really aware of how communication skills can make a huge difference in the trajectory of your career. I mean, I would see time and time again about people being technically really capable, smart, you know, all of those great things. But then when it came to hacking the sales ability to the social proof type of ability to really get their ideas through, that didn't really happen. And then they would find that they weren't getting promoted and wondering why they're not getting promoted because they're doing all of the great work. And so that frustration and seeing all of that is what really made me think there's gotta be a better way to do this because in school you don't learn it and then you get thrown into a job and so you're supposed to just take years to ramp up and there's gotta be a faster way to go.

03:51 Warren Kucker: That's a great great through line. I a similar experience where I didn't necessarily grow up around a bunch of corporate speak. and I went to business school and I started my job and I didn't know I was calling everyone Mr. and Mrs. at work. in the mid two thousands, which apparently had fallen out of favor or never been in favor. But like, I don't know, a vice president at UPS would come down and be like, Well, thank you, Mr. Tringali and he's like, Listen, man, it's Joe. Like there's no Mr. Tringali. I didn't know. Who would you know? I've been calling everyone Mr. and Mrs. my whole life. You this is this like like really pedantic kind of small thing that when people job

04:23 Neha Gupta: Yeah.

04:24 Warren Kucker: I'll throw that out there just in case they didn't know, which cracks me up. along the way. So it's a really good call out. And Yeah. Yeah, right.

04:28 Neha Gupta: Yeah, or even like how do you ask for a day off? Or how do you say, Hey, I'm not feeling well and I need, you know, to go home early? Like even things as simple as that when you're first starting out can be nerve wracking. So

04:40 Warren Kucker: No, agreed. And again, like more complex, but like fundamental to what you're doing at work all the time. I think people forget that you're really all you're trying to do at work all the time is sell your ideas all the time, right? Like that's what you're I have this I thing I'm working on. Here's what why I'm doing it and here's why you should care and I need your help on it, right? So like there's a lot more of that in the workplace that people don't see that as like selling the way that they should. And learning how to engage in that is a really interesting space. So I'm I'm excited to dive in. Cool.

05:06 Neha Gupta: Great. Yeah.

05:07 Warren Kucker: Great. Tell us a little bit more specifically about Breakthrough. what you're trying to figure out is like sorry, let me let me start that over. That was my fault, right? great. tell me a little bit more about breakthrough. What are you what problem are you sp trying to specifically solve and how are you solving it?

05:27 Neha Gupta: Yeah, so what the nice thing is about AI is that like now you have this tool where you can really go back and forth with it. And I know a lot of people do role play through Chat GPT or Claude, whatever it is, their prefer preferred LLM is. And that is a great experience already in its own. But what we found when trying to do role play through those, you know, off-the-shelf LLMs is that they're built to be agreeable with you. They're trying to make you like them because obviously if you like them, you use them more and now you know you're using more tokens, et cetera. And so they're never going to push back on you. The way a real human is going to be. So, what we've been doing at Breakthrough AI for the last 12 months is fine-tuning 11 different personas. We do have a nice one if you want to start off nice and easy and have someone agree with you. But we also have ones that are demanding. We have one that's analytical. We have one that's a bargainer. We have one that's like a novice, a brand new learner. If you're trying to actually explain something to someone, how do you do that with someone who just really doesn't understand like context and stuff like that? So by having all these different personas, now we can actually build you a role play that's custom, not only to your company, your product, your culture, et cetera, but also you can be. Sure, that whoever you're, you know, whoever is actually going through that role play is really in a high pressure situation, right? Like they're really. emotionally feeling very connected to the role play because we found that if you can get them to connect emotionally, they really feel very real to them. And I hear that over and over again where many different people who've tried the product have said, wow, I can't believe how emotionally invested I got into like wanting to do well in this actual role play. So that's what we're really trying to do is like just like an athlete or a musician needs to not just practice a lot before like they perform, but they need to practice hard stuff, the things that they're not actually good At and people don't necessarily get that from just like a standard LLM experience, and that's what we're really trying to elevate here.

07:25 Warren Kucker: Yeah. So talk walk us through the evolution of roleplay, right? Think about like what what was roleplay? People that were doing role playing well five years ago. What were they doing and what were the like positives and negatives about that? And then what are we being able to solve for with AI that just makes that experience so much better?

07:42 Neha Gupta: So role play has been around forever, right? I mean, it's not like a new concept or anything like that. Everyone gets that you need to practice. But usually the way it's been done, it's like two people will come in front of an entire workshop or a classroom of people and everyone's watching their role play. So one is like it's nerve-wracking because you're in front of a bunch of people. And usually the way it was done well is like you might do a little bit of the role play and then the instructor would stop you and then give you a little bit of feedback and maybe you'd retry something again. And so that would be the best way. But it's also kind of Awkward and maybe even a little humiliating to be stopped so much and definitely embarrassing, if nothing else, right? So it feels like very awkward. And one of our early customers actually said that they found that they would then assign people role play and they would say, Hey, okay, you know, we're not gonna make you do it in front of everyone. We realize that that's really awkward, but go find a partner, another colleague, and go and do some role play and then report back. And what they would find out is that people would actually talk about the role play. Like what they might say or what they would say, but they wouldn't actually do the roleplay because being

08:46 Warren Kucker: Okay.

08:46 Neha Gupta: like a demanding client with your colleague is not fun. Like you're still trying to maintain

08:50 Warren Kucker: Mm-hmm.

08:51 Neha Gupta: that relationship. You don't really want to push back on them so much. So that's been, I think, the problem with role play all along is that if you're doing it with someone that you have a relationship with. colleague, coworker, I don't know, friend, et cetera, they're only gonna be so pushy. What's nice about AI now is that it can be pushy. It doesn't care about your feelings. It can say no. It can like have all the sales objections, you know, it wants and stuff like that. And at the same time it's staying very objective and just continuing the role play and not worrying about like the cringe factor, I would say. So that's what's exciting.

09:23 Warren Kucker: Yeah.

09:24 Neha Gupta: Yeah.

09:25 Warren Kucker: That's awesome. Yeah, I mean, even as the moderator in the roleplay, right? Like the person that's helping the other person, you have to be weirdly inventive in a way that's not a natural skill. You in every role play you invariably make like a you say something that's kind of like out of step, right? Like it didn't flow with the conversation. You're trying to move along and that's not how anyone would talk and it completely breaks the flow of the roleplay, right? Like I'm trying to

09:46 Neha Gupta: Yes, that's right.

09:47 Warren Kucker: be VP of customer success that like my rep might want to sell to, right? Well, the pain part four is weird, right? Like it's an unnatural conversation. I think AI really helps smooth smooth that problem. It's much better at being inventive because it's guessing every syllable based on everything that's ever been said. and you don't have to worry about the moderation piece. So that's that's really interesting. well you know fundamentally, I'll take it back 'cause I haven't asked this question in a while. W like why should we roleplay in sales? What's the what what are we getting we're trying to get better at? What's the end result? How fast can we get better at? Give me give me the background as to why this is an important skill that practice to have.

10:23 Neha Gupta: Yeah, I just don't think humans like really actually are that great at sales in general. It's it's it's a form of conflict, if you think about it, right? Like you're negotiating, you're wanting them to do something for you potentially or support you in some way, maybe buy something, obviously. That's like sort of the traditional thing we think about with sales. And that's a little bit of a form of conflict because like now you're coming forward, you're being vulnerable in a sense, you're like really putting something out there and then looking for their response back. And they might just say, No, or I don't like this, and you have to kind of be ready to. be able to respond to that. And I just do not think humans are naturally that good at that. Obviously there's some star people who are just charismatic and maybe able to kind of pull that off just because their personality is that way. But I think that's like very few people. And we know those people, right? Like we all know like, this person's very charismatic. They're sort of a star seller. But most of us actually need to practice it. And I think it's a skill that you can have. I don't think it's something that you have to be born with. I think it's something you can practice and get better at. Just like we don't throw, you know, a soccer player onto the field and expect them to be just be amazing at shooting goals. It's like, why would we do this with like the most important part of our life, which is work where we spend 40 to 60 to 80 hours, but then like we do no practice for that, right? So to me, it seems like kind of a no-brainer to do some sort of practice. I think like this is why people practice a presentation in front of a mirror or videotape themselves, or like I said, do this sort of cringy role play with somebody that they know. But with AI, you don't have to do any of that. I mean, that's the great thing. Is like, and then the best thing is that we've built in this rubric after looking at all of these different academic journals of like what is a good communicator? What actually like makes someone a good salesperson, et cetera? And that rubric is very objective. So even if you're a manager trying to coach someone, you've got your own biases coming in. You've got your own idea of like this is what's good and this is what's bad because of like what's worked for you. So that's honestly like the lens that you're doing on everyone. With AI, we can be objective. Everybody can be on the exact same scale, the exact same rubric. You do not have to worry about some managers maybe being easy graders and some being harder graders, like that calibration piece. doesn't really need to happen, which is like a big part of what happens at big companies, especially. They'll sit in a big room and try

12:43 Warren Kucker: Yeah.

12:43 Neha Gupta: to calibrate across managers. And you don't have to do that with AI. Like it's already calibrated because it's the same thing happening again and again, you know.

12:52 Warren Kucker: Yeah. I think there's two points there. I want to make sure I get back the second point, but on the the first side of the fence. Obviously the sports metaphor of like no one goes out and just plays the game without practicing is is an easy one and it's how you kind of motivate reps. I've actually like kinda parsed it down into it's actually one of the only business dip disciplines where you don't spend any time practicing what you're gonna do and the most important aspect of the the business. So like engineers don't code on production code. It'd be like a f a CFO trying to put together their financials in a board meeting. Marketing's not trying to one shot their marketing. materials, right? Like, but sales people walk into a call with some like baseline. I understand sandwich methodology and I've been doing this for a while, but never actually

13:31 Neha Gupta: Yeah.

13:32 Warren Kucker: kind of being like, I have an important call. This is a $500,000 deal. I mean, everyone's walking into that call cold compared to other disciplines, right? Like it's a it's a weird because it's so intrapersonal and the most important of the job is that discussion that is variable. Plus role playing just hasn't

13:48 Neha Gupta: Mm-hmm.

13:49 Warren Kucker: been fun for a long time. it's it's

13:51 Neha Gupta: Yeah.

13:57 Warren Kucker: on the second piece I thought it was interesting, the calibration piece you talked about, right? Cool, so there's calibration, that's wonderful, right? Like I think we underrate how like a manager coaching a sales rep, like if the sales rep's great, cool, that was a great role play. If the sales if the real sales rep isn't so good, meh, that wasn't a great one. Regardless of the actual outcome, you have that bias at a minimum of their performance and how you think they're doing, right? then you have calibration

14:18 Neha Gupta: Yeah. That's right.

14:21 Warren Kucker: across managers. Let's say you do great, you calibrate well for the product, you're providing a similar experience for all the reps. And coaching them through their parts of their rough spots. How do you make it so it's not too repetitive? How do you make it so it actually mimics a conversation and that there's a a decent amount of variation within the good parameters that actually help reps get better at managing calls?

14:43 Neha Gupta: Yeah. That's actually been like a real art that we've been really just trying to fine-tune ourselves over the last few months, especially because we've been now ramping up, getting a lot more customers, and we want to make sure that the experience does feel as realistic as possible. So we're updating our prompting all the time around trying to make sure that there's curve balls thrown in, that the AI is not literally saying the exact same thing every single time, but like kind of give bringing in different dimensions so it feels a little bit more human-like that way. we also have people. people practicing with a lot of different personas. So the demanding one is going to feel one way, but then the analytical one is going to feel a little bit different. So by having 11 personas now you're able to practice that same thing, but from someone having a different personality basically on the other end, right? So you kind of can f have that feel for what that might you know, how it might be different. So there's a lot of different things that we're trying to think of that make that a lot less repetitive because you can only do a role play so many times and then it starts feeling sort of predictive, right?

15:43 Warren Kucker: Right. Awesome. That's great. Is there anything in terms of like the role playing that's kind of counterintuitive? I I think actually this role playing is a great example for the bideverse build discussion. If a s sales team or any team wanted to

15:57 Neha Gupta: Yeah.

15:57 Warren Kucker: build their own rate role-playing action, they could do it in ChatGPT, right? There's the real-time agent and voice agent that you can have a chat with and you give it system instructions. But there's a lot of work to make it actually like usable, right? Like give it the personality types,

16:11 Neha Gupta: Yeah.

16:12 Warren Kucker: vary, get the user feedback and analyze that data. What do you find that's It's like kind of counterintuitive in how you kind of like build this and actually how like folks have conversations with these role-play agents.

16:23 Neha Gupta: Yeah, I mean I think there's a couple of things that make It's different from just doing like your own prompting with an LLM. One is the personas, right? So it's gonna take it took us 12 plus months with a lot of data to make sure that the LLM is actually gonna stick to that persona. So even if you tell ChatGBT, hey, be a demanding customer for me, it will the first one or two backs and back and forths, and then it just stops. It becomes agreeable. Like I've literally tried this several times in different ways, and it becomes agreeable. It is just ultimately programmed. To be that way. So that's the one main difference. The second one is you don't really know what everyone's practicing. Like you can give the same prompt to Chat GPT over and over again, and it will say different things each time. So you may think that everyone's practicing the same thing, but you don't truly know that. So if you really want to make sure that everyone's having this consistent practice with the objections that you actually hear in the real world, situations that are like role plays that are based on your real situations. doing one prompt or doing a series of prompt is not gonna be helpful. the last thing is like the rubric. There's no built-in rubric. So you don't again, you can tell chat here's the rubric, here's what I want you to base it on. But like now the manager is gonna have to what ask somebody to send them their results every time they do a role play. We have that all transparent. So like the manager can go in and be like, okay, this is the area that they're really struggling with. Now I can go in and actually do some coaching around this, right? You're not gonna get all of that sort of trend analysis across many different role plays from Chat GPT without it being super manual, right? So there's that's why it's sort of like more of a system that we're building rather than just access to it yet another AI chatbot type of thing.

18:08 Warren Kucker: Gotcha. That's great to hear. Okay, cool. Let's say let's walk through something specific. on the leadership side. Let's say great, I've I've bought Breakthrough, I want to implement it on my team. What are some like best practices and like starting a program around some like formal role playing? How should you roll it out? What's the pacing? What are the types of things if I look at the actual like tactics of using role play

18:31 Neha Gupta: Yeah.

18:32 Warren Kucker: to better my sales org?

18:34 Neha Gupta: Yeah, so we've seen a few things. One is if you just roll it out and don't really like tell anybody about it too much. Then it doesn't get usage, right? Like this is true for any training. You kind of have to make it an event, honestly. So you want to like have an onboarding where people are invited. You know, right now it's me who's walking everyone through the product. how do you actually use it? What we've been doing is leaderboards, also, that has really made it a big difference to see like how much are other people actually using the product so that you know, okay, I if they fit it into their day, I should be able to fit it into my day. The most successful pilot that we've done is where they actually She made it a competition, which is something that sales teams do often is competitions. So

19:14 Warren Kucker: Mm-hmm.

19:15 Neha Gupta: a competition around who can do the most overall. We do we've done one with streaks. how there's different levels. So obviously level one, the customer, the client is like nicer. They get more and more demanding as the levels go up. How many like ones versus twos versus threes have you done? that type of thing. The competition really helps. People really like having to like win something, even if it's like a a token prize type of thing. but definitely making it like a training period. People don't want to train forever. So I think what's happened in the past is like either you have like a one-time workshop. And that's great, but it's only one time. Or you have something that's like video training that you just assume people will do on their own time. They never do it, honestly. Like there's always so much more work to be done. so you have to do something in the middle where it's not one time, but it's also not forever, where it's maybe like a six week period where everybody is like really practicing and getting better. And then we kind of end that period. And then if you someone wants to keep going, they could, but like really let's like focus on it, let's talk about it in team meetings, let's like make this a thing. We also send like weekly updates to try to get people to really keep engaging every week, but they're just not going to. It's training at the end of the day and nobody wants to train forever. So that's what we've seen has worked the best.

20:34 Warren Kucker: That's really yeah, it's so some of the fundamentals there, right? Obviously you have to make it a thing and you know what anyone in sales is like looking how do you make it competitive, right? How do you make it into a leaderboard, how do you turn it into like a quota esque number, so that's that's really beneficial because then eventually what you want to make sure is that role play is just something an org does. It's not something extra. This is how this is how we practice, this is how we warm up, this is how we provide the best experience for our buyers when we're popping on a call. so

20:57 Neha Gupta: Exactly.

20:58 Warren Kucker: that's really helpful. let's maybe get into some a specific here. I you mentioned

21:03 Neha Gupta: Okay.

21:03 Warren Kucker: in a pass call a pricing negotiation discussion. Tell us

21:07 Neha Gupta: Mm-hmm.

21:08 Warren Kucker: like how do you structure the prompt or the agent to manage that? How do you grade it? Like w give me some details around let's say pricing and negotiations has a specific in this case.

21:18 Neha Gupta: Yeah. So pricing negotiations, we have built like a scenario where the person is like, everything looks great, right? So you go in, what's gonna happen, your actual experience, let's say you get in, you put hit I wanna practice. It's actually gonna give you a little bit of a framework. On also how to handle this. Let's say you're someone who's brand new and has never done a pricing objection type of role play. We don't want to get you in there practicing the same bad habits over and over again. We want to actually give you some sense of how would best practices look like to handling this type of thing. So we're giving you a little bit of a framework, then we're putting you into the role play. So that's also a little bit of difference in ChatGBT in that it's not necessarily going to give you some of that framework that we've been building based off of like past experience. So you go in and then the role play that we have around price objections, like everything looks great, but can you give me a 10% discount? So, not based

22:08 Warren Kucker: Right.

22:08 Neha Gupta: off of like anything that's like wrong with the proposal or any issues. It's just, hey, this feels a little higher than we wanted to spend. I want a 10% discount. You're coming up, you know, I love everything you're doing, but you're on the high end of others that we're looking at. So there's different ways that we're having the, you know, the agent basically like asking for a discount. And what you're practicing is how to not just give in right away in order to make the sale, but really to talk about the value of what. you're delivering, right? And like why that's the price. So making sure that you're both are aligned on here's how we built this, here's the expertise that went into that. Here's the value that you're getting out of this. And that's why this is the pricing. Because if you focus straight on pricing, then of course everybody wants a b a good deal. Like everyone's gonna ask for that discount. And the way to counter that is not just by giving in and giving control of the call over to the buyer right away, but how do you make sure that they understand the value and stuff like that. So that's really what you're practicing. And then after you've gone back and forth a few times, I d you know, we do have it programmed well, it will it will be a positive solution if you're doing all the right things, it will back off, for example, in a price situation. It will say like, okay, I see the value, like let's move forward. You know, so like the it that is how we programmed it so that you feel like you can be successful in the situation. Then the rubric, there's a feedback agent. It's looking through the transcript, it's looking how you responded. And it's looking at things like how did you manage pushback? And it's looking for micro behaviors on that. Like, did you acknowledge their concerns? Did you bring back like specific data points or value points to like manage that pushback versus just saying, no, we can't do it or something like that. Did you end the call with some sort of momentum? So, like, how are you actually ending the call? Is it just like waiting for them? Are you giving them timelines? Are you talking about next steps? Like, how do you make sure that there is some momentum at the end? how clearly are you explaining the product and the value, right? Like, is it easy to follow? Am I having a hard time following you? Are you rambling too much? How much are you talking versus listening? You know, so those are the type of things that we're looking for. So there's different rubrics depending on if it's a sales track. or you're like a new manager and then what the AI agent is doing is it's looking for all of those micro behaviors again. So like the more we define it, the more objective the the AI feedback ends up being.

24:36 Warren Kucker: I think it's really good highlight of why practice is important. and why rehearsal of these types of conversations is important. A discount conversation actually has a monetary value to it, right? So you can almost peg how many discount conversations, how many like pricing negotiations am I having, how many deals am I closing, what's a ten percent discount on every deal worth? And then if I don't practice for that at all, am I more likely to give that ten percent discount? There's like a a funny math parallel. And I actually think the behavior

24:59 Neha Gupta: Yeah.

25:01 Warren Kucker: that would happen w happens when you don't practice versus when you do is really stark here as well. So most of the time what ends up happening in pricing negotiations as a sales leader, I give one training on how to negotiate on price on onboarding, right? Like that's it. So like that's what you get, right? Like and then I talk about it during coaching calls and people ask me questions. or they'd say they're gonna have the call that I bring up and try to bring up the framework that we use in order to manage that.

25:24 Neha Gupta: Yeah.

25:25 Warren Kucker: Without practice, what ends up happening is like someone gives the same old answer, which if you're gonna guess is like, listen, I'm not I'm not authorized to give you a ten percent discount, but let me talk to my boss and see what I could do. Right? Like this is whatever eighty percent of salespeople are saying, right? But the frameworks

25:35 Neha Gupta: Yeah. Yep. Yep.

25:38 Warren Kucker: are really easy and clear, right? But it's not if it's not repeated, it's not noticed noticed. So the price in negotiation, you say, like, great, totally understand. Is it that you're in a specific budget that we're trying to hit? Is it that you don't see the value in the ROI that we discussed on our last call? Or are you just trying to be a good steward of your company's capital and you're asking for For a 10% discount to ask. And it's perfectly fine if it's the last one, I would do the same. And that really disarms people, but that gets talked about once or twice or three times a year. That doesn't become a learned behavior. The default behavior becomes the behavior. No one recalls that. If you can really

26:08 Neha Gupta: Exactly. That's right.

26:10 Warren Kucker: spend your time role-playing, be like one, have role-played that 20 times this year. I don't expect everyone's role-playing a pricing discussion every week. But then before that call, have that role play. It really reforce reinforces core core instincts, right? rather than trying to like f push a behavior that maybe you learned once or twice. So I think that's really kind of like this pricing discussion c discussion here is really stark. Yeah.

26:32 Neha Gupta: Yes, exactly. It's that muscle memory. Yeah, that muscle memory you're trying to build up. So then you go in also not having this sort of desperate energy of like really wanting to close a deal, because we know that doesn't work either. You need to come in, you know, feeling calm and confident. And that's what really like buyers want to buy from people they like and they trust. So you have to come across like likable, confident, calm, all of these things. And that happens by reducing the anxiety, right? So we've seen like in our last pilot that people felt forty percent less anxious after they did these role plays. So that then you're not coming in with all this anxious energy, right? Which is really important in sales.

27:13 Warren Kucker: Yeah. that's awesome. Maybe one s second to last question before I ask my typical last question. From your perspective, you're selling your own product. How have you found that your role playing tools helped you in certain directions beat old bad habits or find new ways to get more creative on your calls?

27:31 Neha Gupta: Yeah, I definitely find that I talk too much. so I get that feedback from it. I I'm not as short in my sentences. So I've been trying to get to the point quickly and not reiterate again and again. I've also found that I need to just ask more discovery type questions, right? Like really engage the the buyer in what is really important to them. And then the sale actually becomes much easier if I'm actually engaging them and getting them to talk about themselves. So that's doing this role play has really helped me to sort of stop my instinct of wanting to just talk about the product and all it's great features and why you should buy it, but really like stepping back, taking a chance, you know, time to like ask questions, listen, and then having something to reiterate back to the buyer, right? Like their own words almost back to them on why this is like a good fit for what they what they need. So those are the things that I have sort of picked up from just doing this where it's become a little bit more natural for me to do that and not have to think about it because I'm already thinking about a million things. Like I don't really want to have to think about exactly like how to control this call as well. I've also learned a lot about like there's different styles of sales, right? So there's there's the hunter style, there's farmer, there's gatherer. So like also practicing different modes of that in sense in terms of like, is this a worm lead? Is this like a connection of somebody that I know? Is this someone totally cold? So like even that been really helpful because those relationships and those conversations can be a little bit different, you know.

29:03 Warren Kucker: Awesome. That's great. Cool. All right, last question. Let's fast forward three years. Obviously we're at the early stage of this generative AI journey here. Three years

29:11 Neha Gupta: Yeah.

29:11 Warren Kucker: from now, what do you think we're doing in this role play space that we couldn't even fathom two to three years ago?

29:18 Neha Gupta: Yeah, I mean, I think like it is just gonna become more and more realistic. I mean, definitely you see ours is very audio focused because we feel like that feels the most realistic right now. we've seen a lot of the ones that are more video based and they just feel sort of chunky and clunky and like the lips aren't moving exactly right. And I find that to be more of a distraction personally than like getting to really focus on the role play. But that's what's gonna improve,

29:42 Warren Kucker: Mm-hmm.

29:43 Neha Gupta: right? Like that ability to actually feel more realistic video-wise. I also think Think the interesting thing, I read this study about how we're gonna be able to use AI to actually detect a lot more around tone. Right now,

29:56 Warren Kucker: Okay.

29:56 Neha Gupta: it's like very word-based. Maybe you can count the number of words someone speaks, so like their speech, like how fast they speak, or something like that. But we're gonna be able to eventually be able to say, hey, you sounded a little sarcastic at that point. Or you sounded like not as excited or something like that. Like are we really gonna be able to get into people's tone, which is a big part of communication. And you

30:19 Warren Kucker: Yeah.

30:19 Neha Gupta: don't see anyone doing that really that well yet. So that I have

30:22 Warren Kucker: Interesting.

30:22 Neha Gupta: read, I think it was like something out of MIT that's experimental, but they're really trying to use AI to detect not just positive and negative tones, but like actually what was the tone of the person speaking. So I think that'll be really cool to see.

30:35 Warren Kucker: That's awesome. Yeah, that's the tone piece is a really interesting part. I hadn't really thought about that. It's still guessing syllables, right? So like it's not it's not picking up on tone, right? Which is there's a lot there's a

30:44 Neha Gupta: It's not picking up on Tone yet, yeah.

30:46 Warren Kucker: lot of depth to tone that you're trying to and then

30:48 Neha Gupta: Yeah.

30:48 Warren Kucker: it'll be like again, there's a good biverse build. If you're doing a role play and you're trying to incorporate tone You wanna make sure you're s serving up that tone and that context in a way that's actually useful and not giving someone wide varieties of different tones within a call that aren't logical, right? So like and throwing too much at someone, when you're looking at tone. it's also

31:04 Neha Gupta: Yes, yes, exactly, exactly, yeah.

31:08 Warren Kucker: the repetition n t tone is subconscious, right? So like hearing the same tone over and over and how you respond and having that response be well received is the value, not the they're sarcastic, so I'm gonna say this sentence kind of like rote response kind of scenario.

31:22 Neha Gupta: Yes,

31:22 Warren Kucker: f fascinating. Well super excited to see what you

31:22 Neha Gupta: yes, exactly. Yeah.

31:25 Warren Kucker: we continue to do with breakthrough. before we go, tell us where our listeners can find you. You post on LinkedIn, any blogging, go to any conferences.

31:32 Neha Gupta: I post, yeah. I post on LinkedIn all the time. So definitely find me there. I usually post post under myself, so under Neha Gupta, and then Breakthrough AI is reposting. Our website's join breakthrough.ai. You can always reach me there. My email is just [contact details removed]. So anywhere, yeah,

31:50 Warren Kucker: Perfect.

31:51 Neha Gupta: you yeah, you can find us. LinkedIn is definitely the best place though. I'm I'm on there all.

31:56 Warren Kucker: Great. Well it's been an absolute pleasure. Looking forward to seeing where you go through in this space and best of luck.

32:02 Neha Gupta: Thank you. Yeah, it's been a pleasure. It's fun talking about it. The stuff I can talk all day on this. Yeah.

32:05 Warren Kucker: Same here. Yeah, same here. excellent. Thank you, Nayha.

Account ExecutivesSales ManagersVPs of Sales

About the Host

Selling AI is hosted by Warren Kucker, founder of Topiq. Each week he sits down with revenue leaders, founders, and operators to unpack the strategies, tools, and stories behind selling AI products and using AI to sell smarter.

Warren Kucker — Host of Selling AI

Warren Kucker

A revenue leader and entrepreneur who founded Basiq.work, focusing on conversational AI for sales teams. Previously served as VP of Sales and CRO at Series B companies, scaling revenue to $2M+ quarterly bookings. His exits include Shelly.ai (ML email automation) and Boxton (digital freight platform). Started his career at Apple managing a $100M shipping efficiency program.

Related Episodes

EP. 33 37 min

Your Buyer Doesn't Know What to Ask For Yet

with Michael DeLeonardis

Michael DeLeonardis, chief revenue officer at Verity, explains how selling autonomous warehouse drones differs from selling software, and why AI has pushed enterprise sales back into a demand creation cycle. He and Warren Kucker discuss deep discovery around physical constraints, win-loss analysis from call transcripts, deliverable-based selling that guides the buyer through each decision, a client success role that spans pre-sales and post-sales, and why industry expertise and a LinkedIn presence matter for sellers.

EP. 32 38 min

"You're Inspiring People to Mute Your Logo" — A RevOps Leader on AI Outbound.

with Jimmy Downing

Jimmy Downing, VP of Revenue Operations at NextGlass, explains how he went from an accidental Salesforce admin to running RevOps, and why the AI tools that work are the ones that solve one simple problem well. He and Warren Kucker discuss why AI-driven outbound damages brands, where to layer AI on clean data, why soft skills now matter more than ever for sellers, and the risks of features built too quickly with vibe coding.

EP. 31 31 min

Build Your Campsite Where the Buyers Live

with Morgan J Ingram

Morgan J Ingram, who went from sales development rep to sales trainer and now runs AMP Social, explains why outbound has shifted from sheer volume to creativity and intention. He and Warren Kucker discuss in-person events and experiences, pattern interruption in an AI-saturated inbox, how market shifts affect SDR and AE careers, building industry credibility, meeting buyers where they live online, and why deepening relationships will matter more as buyers grow less trusting.