Salesforce Admins Podcast How Admins Can Embrace the AI Shift
Today on the Salesforce Admins Podcast, we talk to Prag Ravichandran Kamalaveni, Founder & CEO of Skilled Cohort and the Founder & Co-Chair of Salesforce Saturday Cohort Canada.
Join us as we chat about his Dreamforce presentation on AI readiness and how to be a better storyteller.
You should subscribe for the full episode, but here are a few takeaways from our conversation with Prag Ravichandran Kamalaveni.
Why AI is moving technology from clicks to chatsI caught up with Prag fresh off his Dreamforce presentation about AI readiness. The idea for his talk came from thinking about what people were saying years ago, when he started attending the conference, compared to now. If we're talking about Agentforce this year, what will we be talking about in five years? In ten years?
"We are moving away from clicks and moving towards chats," Prag says. AI currently sits in the application layer, working as a tool on top of your org. But as these functionalities become more deeply embedded in everything we do, Prag predicts that an "agentic layer" will sit directly on top of your data.
What this all means is that data cleanup and data quality need to be top priorities for your organization to get the most out of AI.
How to find presentation topicsAs a 10-time Dreamforce speaker, Prag has a simple and effective approach for how he turns ideas into presentations.
It starts with the topics—pick something you're excited to talk about. The best topics are in areas where you have some experience but want to learn more. "Curiosity is fundamental for success," Prag says. Write out a list of topics that you're curious about, and then look at the big picture of how they might fit together in a presentation.
Prag also points out that you don't need to write a fully finished 20-minute script to submit for a conference. Technology moves so quickly that by the time you're actually giving your talk, half of the information will be out of date. Write a good abstract and focus on the core concepts.
Finally, keep a sense of perspective and don't get discouraged by rejection. As an event organizer myself, I see plenty of great presentation ideas that just don't fit with the event as a whole. It's all about persistence, so keep at it.
Forget the script—go for flexibility and rehearsalSo you've submitted some great topics and booked that speaking gig—how do you prepare?
As he's gained more experience as a speaker, Prag has stopped trying to write a detailed script. Instead of focusing on the exact words you have to say, think about what you want your audience to understand.
Practice makes perfect, so make sure to take the time to rehearse your speech in front of friends, family, and anyone who will give you input. Prag has found ChatGPT to be effective for doing a few practice runs, once he prompted it to stop being so complimentary and give him direct feedback.
Listen to the full episode for more insights from Prag about AI readiness and how to be a great presenter. And don't forget to subscribe to the Salesforce Admins Podcast to catch us every Thursday.
Podcast swag Learn more-
Salesforce Admins Podcast Episode: Essential Tips for Creating Effective Presentations
Full show transcript
Mike: This week on the Salesforce Admins Podcast, we're joined by Prag to talk all things AI readiness, from Dreamforce stages to real world slide decks, so if you've ever wondered how to prep for a major presentation in a fast moving AI world, or how tools like ChatGPT can play coach, creator and confidence booster, I promise you, you're in for a treat. Now, Prag shares what it takes to go from idea to impactful delivery, plus the pivotal role that data plays in getting AI ready. Now, whether you're planning your next big talk or just curious how AI fits into your workflow, this episode has something for you. So with that, let's get Prag on the podcast. So Prag, welcome to the podcast.
Prag: Pleasure to be here, Mike, with you.
Mike: Let's get started, for those people that didn't see your presentation at Dreamforce or haven't bumped into you in the community, let's do a little bit, tell me how you got started in the Salesforce community and what you presented at Dreamforce this year.
Prag: Wow, okay. So you are backdating probably 15 years ago, that's when I started Salesforce. This was the pre-Salesforce classic version is when I started, and Salesforce evolved a lot, and I couldn't believe that it's already 15 years past, and I think this Dreamforce is my sixth or maybe seventh.
Mike: Wow.
Prag: But every Dreamforce for me is pretty amazing because of the energy and people around the globe who could able to come join us, exchange ideas, mostly the problems and the pain they're going through, it's fantastic. So I always look for Dreamforce every single year, and the energy that I take away from Dreamforce is obviously the biggest driver for me. I know people sometimes complain that they become tired by end of day four. I think, me, on the opposite, I allow to steal energy from others, so I take more energy, if it is crazy week like Dreamforce, and that's fantastic feeling for me.
Mike: So you're taking other people's energy, oh no. I think this year you presented about getting AI ready. Can you talk to me? I mean, if you think about it, even in the short time that you've been, if you go back six or seven Dreamforces ago, I promise you we probably weren't talking AI, and now we are. How was that presentation and what did you talk about?
Prag: So it's a very, very new topic to a lot of us. We are consuming AI in a lot of ways, and I'm pretty sure almost everyone who is listening to this podcast might have used some version of ChatGPT or Gemini at some point, but I started believing in my co-speaker, like we both started believing that we all are going to start using AI as part of our work. So the presentation, we started in a way that in the next five to 10 years, what we predict is we are going to move from application layer to data centric layer. What it means is the AI or the agent layer is going to sit on top of the data layer, and the phrase we actually mentioned at our session was, we are moving away from clicks and moving towards chats. So that's going to be a very interesting starting point when it comes to AI. How are we going to consume AI?
Mike: Yeah, I mean, I feel like having just come off of [inaudible 00:04:26], even if people don't want to talk AI, they're still talking AI about just about everything. I even had an AI conversation with my Lyft driver, which I will tell you, that was probably the first time I've ever had that. Yeah.
Prag: That's nice. Yeah, I think none of us could able to avoid the topic.
Mike: So what did you see at Dreamforce this year that you're excited maybe to have come GA or just the next advancements in what could come out for Salesforce in 2026?
Prag: Wow, there's a lot of things, but being very honest, I enjoyed this Dreamforce because the message was clear. If your AI needs to work, you need to have the right data. I think that message was little bit under the hood for a while, but I think it's been clearly communicated this time that any kind of AI content that you consume during Dreamforce, it's not even disclaimer, it's very clear that, "Hey, you need to have the right data to make sure that your AI works the way you want it to work." So first of all, kudos to the team who made this approach of making the message clear about the data layer, which is super important to make your AI works. But I also super excited about the naming convention and the name change of all the products.
Mike: Oh?
Prag: It's not easy to move from sales cloud to agent force sales. That's a big difference. That means that there is an organization switch at the company level, or at the product level, there's going to be a lot of agent force presence going to come along that we are going to start using probably much faster and much sooner. Those are all kind of an indications that I got, but at the same time, at the keynote, I was super surprised to see Patrick doing a demo of agent code wipe. That's beyond what I would expect to see at the keynote floor.
Mike: Well, that's the whole reason we do keynotes is to wow and amaze you.
Prag: Yeah, it is definitely worth getting the goosebumps moments there, for sure.
Mike: Yeah. Well, let's talk a little bit about, so you presented at Dreamforce and you talked about getting AI ready, I think absent of talking about your presentation, one thing that's really relevant having just come off of a community conference, and I know quickly the holidays will be here and gone and it'll be January, February and community conferences, user groups, Trailblazer DX will be coming back. I'd love to hear your perspective on if you had some advice or things that you would love to share about how you got ready to present at Dreamforce that would be helpful for others.
Prag: I think what I have done in the past as a mistake was, oh, I want to make sure that my presentation is ready or the content is completely ready before I even submit my topics to talk about. I think that's a biggest mistake that I have done in my own life and the moment I started coming out of it, because the things are changing much faster. You cannot have something ready assuming that you can present that in three months or four months, then you're going to be old dated content. So my suggestion would be, start picking up the topics that you are excited to talk about. You have experience, but also wants to gain more experience. So the curiosity is the fundamental success for you to have a successful content delivery and for that, pick a range of topics. So I'm always the security guy. I love talking about Salesforce security. It's just my thing. And now I love data. I started talking about data layers in different angles, and since I talk about security and data layer, I could able to start adopting AI topics because when you talk about anything AI, you need to have two most important things, which is the data and how secure that AI have access to the data. So it's kind of putting those puzzles together is going to be phase two of how you can start thinking of presenting to either the community conferences or trial ideas or even at Salesforce World Tours. So these are the two fundamental things that I would be picking before I start worrying about how to deliver that presentation.
Mike: Yeah, I completely agree, I mean, selfishly, the event organizer in me would love to have everybody have all their topics ironed out, but in reality, if you have a really good idea and a really thoughtful abstract that explains it, it makes me think of the Wayne Gretzky quote, you just have to skate to where the puck will be. Which is you have to know that you're going to create that presentation even though it's not created right now, so I think that's really good advice because sometimes people can think they have to sit down and have a fully vetted idea before they submit, and you don't, you just have to know that your idea will come together and be incredibly helpful for others.
Prag: Exactly. Again, I was one of those folks who made that mistake, but the moment I moved out of it, now I start my presentation with just a piece of paper or a Google Doc. I won't even go to Google Slide and have four or five bullets. I would, "Okay, this is the four or five things I want to talk about that. What could be a potential title?" And then I pick a title and I usually submit at least three or four different kinds and versions of the content, because again, don't give up if your topic didn't get selected. I have seen so many people, if they don't get selected to speak within the first two or three times, then they started believing that they are not the good presenters or they don't know how to submit their topics, and they started giving up on not taking that approach of submitting the topics. I want to tell them that my topics has been rejected many times compared to the times of the topics that's been selected. So rejection is good at times to make sure that it's not your turn yet doesn't mean that you never get your turn.
Mike: Yeah, I mean, having been on both sides of the fence on that, it's amazing, but if you think about it relevant to other things like applying for a job, you apply for jobs when you're searching for a job. And yes, there's jobs that you really want, but you also apply knowing you're not just the only person applying, and to your point, it's not like you find one job and apply for it and then go, "Oh, well, I guess nobody ever will employ me again, just because I got rejected from that one." From an event standpoint, it's a combination of, what are the contents being submitted, what is the mix that the event is looking for, and what do I have to pick from? And sometimes you can be one of a thousand submissions on a certain topic, and it's not that your content isn't good, it's one of a thousand, and there's maybe only two or three slots. And other times it can be, boy, I've had it where, like in your case, I've needed one or two sessions on a certain topic, and thankfully there was one or two submissions on that. And so I hate to say it's luck, but a lot of it is. It's being open and flexible and kind of making presentations that fit for what the organizer's looking for.
Prag: Exactly, exactly. So I think if people get clarity around that, I think they won't get demotivated for not getting selected.
Mike: No, I mean, it's also, how many times do you buy a lottery ticket and you don't win the lottery, right? If you bought a lottery ticket, you're like, "Well, I'm going to win." And then you don't win and you're like, "I'm never buying a lottery ticket again." Come on. I mean, I hate to say it's the same way, but part of it is that's kind of just how, when you're in a big pool like that, it works, and sometimes the reverse also holds true. You can submit something and it may be a great presentation, it just doesn't fit for what the company or the organization's looking for. I mean, I've gone through lots of call for presentations for other events, and I'm like, "This is a really good subject." And from my perspective, it's really good, but from an organizer's perspective, it's good, it just doesn't fit what they're looking for. And so it's like being cinnamon and somebody's trying to make a ravioli dish, and I'm not saying cinnamon's bad, cinnamon's really good with sugar and frosting. You can make cinnamon rolls, it's just, we're trying to make ravioli over here and cinnamon doesn't work out.
Prag: Yeah, yeah, no, I one hundred percent agree with that. So yeah, these are all something that I suggest people to think before submitting their topics. But once it gets accepted, I think the best thing that you can do as a presenter is not to have a pre-written script.
Mike: I would agree.
Prag: And I have tried it multiple times and almost every single time I failed because if things are not going as per the plan, which 90 percent of the time it doesn't, then the moment you get into the panic mode, the script that you remember magically disappeared and you started blanking, you started blinking, you started filling with words that is, even though it's appropriate, but it doesn't match the exact tone that you're trying to say certain things. So what I started practicing for the last decade is, "Okay, I'm going to talk about the things that I do know and I'm not going to stick with a specific script." And if I'm speaking with my co-speaker and if my co-speaker took more time, then if I don't have a script for me, I could able to shorten the version of the content that I plan to deliver. So there are a lot of flexibility and lot of freedom that you get if you do not have a script that's pre-written and you want to stick with it, which is not easy for people who doesn't have English as their native language because I'm coming from that background. But once you started practicing by delivering presentation without scripts, it's easy, it's comfortable, and you won't go into a panic mode any time.
Mike: I couldn't agree more. As somebody that presents a lot and helps prepare presenters, there's definitely a point in time that a script is necessary to help you kind of vet out what the talking points are, but you're not in a play, you're not in a musical, you're not in a blockbuster movie where you have to say it word for word. And the thing that I always remind presenters when they get up on stage is, people will remember how they feel when they saw your presentation. People will remember the passion and the information better if you just speak to, what is the point you're trying to make across. And a script is definitely a step in that, you need something to kind of get your framework about it, but I think too often people do get hung up. Even English speakers, where we're native language, they get hung up on, "I have to say these exact words."
Prag: Exactly.
Mike: And you don't, you have to convey the meaning. What is the point of this slide? The script got you there, but in your own words, what is that slide? And that's often when the presentations, I can always tell when a presenter goes what we call off script, and generally the first time they rehearse it and they're off script, they go long because the need to over explain, it's in the back of your head and you're like, there's this little person inside that's like, "You didn't say enough. You got to keep talking. Say more words. They don't understand."
Prag: Exactly.
Mike: And the irony is, it's actually fewer words, but with more intention, because if you choose your words carefully, then you can really convey that meaning. So yeah, I mean, boy, and I've seen, you can also tell as an audience member, when somebody is up there and they're going through the script in their head and they're looking up and then they forgot something and they don't know what to do, and that's where rehearsal comes in. Just rehearsing for that. I always tell people, "You should rehearse in front of your family, your friends, and periodically have them unplug the computer or let something go wrong." Because it's like why sports and athletic people practice all the time, because the more you go through the motion of doing it, the more natural it is when something comes up. I've walked up on stage and had microphones fail, I've had screens go blank, and there's a moment where you're like, "Oh, I've practiced this. I'm just going to keep with the talk. I'm going to let the AV people do their thing, and if it comes back, great. And if it doesn't, then I just realized nobody can see what I saw and I'm going to have to describe it in my head." But if you're on script, you're like, "I don't know what to do because the script doesn't say that."
Prag: Yep, exactly. No, I one hundred percent agree with that point, right? It's the pattern, and I think I've been told once that you should also explain in a way a ten-year-old should able to understand, so that you kind of avoid all the complex words and terminologies, to even explain a complex concept, but in a simple way.
Mike: Yep, I agree with that.
Prag: Yeah.
Mike: Interesting, so I'm kind of curious, this is sort of Salesforce related, sort of not. But what role did AI play in getting you presentation ready? Did you lean on it to help create some scripts for you? Did you lean on it to help create some visuals for you?
Prag: Oh, a combination of both, being very honest. I did use AI to generate specific graphics. I used to have to go rely on a graphic designer to do some graphics that gets aligned with Salesforce color factor, which changes every Dreamforce. So until last year, I was relying on another human who could able to do the magic and then get it plugged into my presentation, but this time it wasn't required, actually. I could able to give the prompt, give the exact font style I want, give the exact place of the image that I want, and it could able to get back to me with my expectation. Even though it does not meet exactly how it used to be done in the previous times, but this is pretty amazing, I took advantage of building images for sure. The other thing I used it for is subtitles. For each and every slide, I always try to avoid the tiny little subtitle under your title section on each and every slide, because I don't know what I could put there as an subtitle, and then I tried it with AI and it could able to process my content and say, "Oh, this is something you can put under your title." Which is very interesting way of seeing it. And I started almost for all my slides, I used AI with my own content that I put it on that slide and asked it, "Hey, how can I frame it under subtitle?" And it could able to give me some really interesting titles.
Mike: Yeah. I got to say, for probably this year, for me, it's been so incredibly helpful to generate images that I needed to really drive home a point in a presentation without worrying about copyright or having to figure out where I was going to get an image because it could generate the image for me, and I saw that across other presentations as well, because that was always something you run into, is the use of images that were under copyright.
Prag: Exactly. Oh, a hundred percent, I agree. I also remember, I want to remove a certain content from the image, and I asked AI to remove it. It did a decent job, but the job it does is all I needed, and that was an a-ha moment for me. Rather than just generating one, it also fixes the image in a much faster and a much efficient way.
Mike: Yeah, yeah, absolutely.
Prag: The other thing, I know you ask about the presentation and you mentioned about presenting it to your friends and families, but you wouldn't believe that I presented to ChatGPT, actually.
Mike: Tell me more.
Prag: Yeah, so it's the voice conversation and I gave inputs that, "Hey, these are all the topics. These are all the content that I'm going to present on, and I want to do a trial run with you." And it was listening for the entire 20 minutes, and then the feedback it gave me was phenomenal. It was like, "Here you went faster. You have to slow down. Here, your content is repeating, try to address. Here you are using lot of ums. Here you are using a big pause. Here you are running because you notice that you're running out of time." This is amazing. So the active feedback layer that I received from the conversation I had with AI tuned me, trained me in a much faster way compared to doing it with a human.
Mike: Yeah, I mean, that's really amazing because if you think about it, part of presenting in front of others is just to get used to being in front of people and talking, but rarely do you have the opportunity to present in front of a talented speaker coach, and you did that because ChatGPT has access to, I don't even know how much information, and it can be that talented speaker coach that normally speakers pay thousands of dollars to interact with, and you got to do that. It's really an interesting use case, I really like that.
Prag: Oh, yeah. And the feedback it gives, it was unbelievable. So I did at least a couple of times. The first time, it was more soft acknowledgement, I got it. And then the second time before I even say, I said, "Hey, be rude. Try to be a coach. Don't give me polite responses. Be straight and help me understand because you are preparing me and not judging me." And then you won't believe the acknowledgement and the response I received was more straight, there is no polished way of saying, it says like, "Prag, this is where you made this mistake. This is how you need to fix it." It was more, I think I would say commandable in a good way, which I really liked it.
Mike: That's interesting that you had to tell it to be a little more straight with you as opposed to probably understanding and empathetic.
Prag: Exactly. It's not a human. So for me, why would I expect it from a mission, right? It's like, okay, I'm here. I want to use you for this particular reason, this is the use case for me to use you as an application. I don't expect you to be nicer to me. I want you to be the reason for why I'm using you for.
Mike: Right? Yeah. I know, maybe this is a positive thing, but often I run across a lot of prompts and different settings for different AI because they're built to be so helpful, and you often have to kind of like, "No, dial it back. You can disagree with me." And I run into that where anything that you use, the AI is almost like, "Oh, well, that's a really great idea, Mike, and let's do that, and would you like me to do this?" And you're like, "But it's not always a great idea. You can tell me it's not a good idea." You can go to an AI and be like, "Hey, I've got some cinnamon. Would this taste good dusted on ravioli?" And I mean, I guess to our benefit, we've built AI systems that are like, "Absolutely, Mike, you are the Bobby Flay of the next world, but [inaudible 00:29:52]." You're like, "No, you can say it's going to be bad." But we almost have to tell it that. I guess it's somewhat commendable that we've done something that we have to say, "No, you can tell me I can be wrong."
Prag: Exactly. And I am one of an example that I tried it and the responses are different. So if I am having any serious thought process that I have, and if I want to get an AI, not opinion, but how it would answer it or how it would handle the situation, before I ask them, I will say like, "Hey, make sure that you don't need to be kind or soft as part of your response and being very straight. I want more factual information." And the response you get is definitely different.
Mike: Yeah. Wow. I love your tip on using AI and presenting to it and having you give you feedback, because I think that's incredibly useful. Well, this has been a fun conversation, Prag. I appreciate it.
Prag: Thank you, Mike. It's always fun connecting with you, and I'm happy that we are talking about the future of how we are going to use tools that's going to help us becoming more and more efficient in what we are doing. That's my excitement.
Mike: I know I often think back to what'll happen in 10 years if somebody's listening to this podcast and be like, "Oh, you remember that time that Prag and Mike were talking about presenting to AI, and they thought that was such the future, and now it's bloody [inaudible 00:31:51]." I think of that because I go back and listen to old episodes. I mean, there's over 12 years of the podcast episodes available.
Prag: Wow, wow.
Mike: And I go back and I remember talking about things like dynamic lightning components, like a page rendering dynamically, and now we're just like, "Well, yeah, obviously you would do that. Obviously you would do that." But back then it was like, "Oh my goodness, you can do that now, and you can make this setting." And then I remember putting together a demo and like, "Watch, we'll change this stage, and then this field will appear." And then everybody's quiet, and then the page refreshes and the record's there, and then the field shows up and you're like... I wonder if we're not going to sound like that in 10 years when we're talking about AI.
Prag: But I wish we be like that in 10 years, right? That's where the technology advancement is going to take us.
Mike: Yeah, hopefully we still have to keep telling it it can disagree with us because I kind of like that part.
Prag: I agree. I agree.
Mike: Well, thanks so much for being on the podcast.
Prag: Oh, thank you so much for having me, Mike. It's always a pleasure.
Mike: Big thanks to Prag for joining us and sharing so many insights from crafting compelling content, to using AI as a trusted prep partner. If you're thinking about submitting a talk or just want to sharpen your storytelling, this episode was packed with some wisdom. Now, do me a favor, share it with a fellow trailblazer who perhaps needs that nudge. Until next time, we'll see you in the cloud.