From Static to Dynamic: How to Turn AI Into Your Sales Enablement Cheat Code w/Michele Nieberding

Speaker A: This is Wes, and you are listening to project marketing to you.

Speaker B: Welcome back to another episode of, uh, Product Marketing for you. Today's guest is my friend Michelle. Now, I'm excited for a couple of reasons. One, she's obsessed with AI to how it actually works for PMMs. Not the buzzword version, but the actual real version. Two, she spent her career figuring out how and and why your sales teams ignore all the enablement content you build and she has a game plan on how to fix it. And three, she's got zero patience for bs, which means this is gonna be a really good conversation. Might be a little bit therapeutic for some of us. Now, Michelle's worked at some solid companies. She lives and breathes this stuff, and she helps other PMMs figure out how to do this AI thing in a productive manner. But here's what matters. She actually knows how to take all the garbage static content we've created and turn it into something your sales team will use without you having to beg them to do it. If you've ever spent two weeks building a beautiful one pager, a deck, a resource guide or whatever, and then the rep hits you up on slack asking for the exact same thing you just built, Michelle's got the playbook to help you out. And this episode is for you. Okay, let's take a break, and when we're back, I'll get into my conversation with Michelle. This season's partner is PMM Camp, the number one resource for product marketing leaders with access to newsletters, leadership frameworks, and a vetted peer community. It's where smart senior PMMs like you go to become indispensable. And most of this season's guests are actually campers themselves. So consider this your sneak peek into what life is like as a camper. Go to pmmcamp.com and join me and 9000PMMs and counting on on our journey to become the next wave of product marketing leaders.

Speaker A: Product marketing for you. Product marketing for you.

Speaker B: Michelle, how are you?

Speaker A: I'm five cups of coffee and. But who's counting? So we're doing great today. This is going to be a vibe. I'm excited.

Speaker B: It was. Our last conversation ended up being like a micro therapy session, so hopefully this is somewhat kind of like that. So we're going to get the best of each other right here. But. But we're not here to talk about my mental situations. You are arguably known, at least in my circle, as one of the predominant voices in the PMM community when it comes to AI as well as, dare I say, an enablement queen. Which is why I'm super hyped that I, uh, somehow taught you into joining this one. Because we're talking about a couple different things. More specifically, you've got a playbook that is designed to help PMMS take static content and get that into actionable, useful, dynamic enablement resources. But the first question I have for you, Michelle, you've mentioned this thing called an enablement black hole. You spend weeks and we've all been there, we've as PMMs, we've spent weeks building this beautiful content to support launches, to support efforts, to support ask from our sales teams. But the rep still slacks us at the end of the day, asking for key differentiators between a specific competitor and you, right? And it drives us crazy because we're like, yo, my guys, it's. And ladies, it's been right there this whole time in about five different places. But there's a specific reason why the traditional sales element is so broken. What's actually causing this disconnect?

Speaker A: It's so funny because I think as uh, PMMs, we're on the hamster wheel. And then we build content that goes into this black hole. It's like a weird cycle that happens and I think for uh, so many different reasons, right? Like sales lives in this high velocity, like we need this now state. You can tell them a hundred times, but until they're in that moment that they needed, it's almost out of sight, out of mind. And if they don't have the resources like the things available to them to get what they need in that moment, that's when you get that like panic slack. It's like right before a big call, right? And they're like, I've known about this prospect and this competitor for how long? But nope, now 10 minutes before the call is when I need your help. And you're like, guess what, I'm in a meeting too. It's this chaos that happens. And I think we're in this black hole because we for so long have seen sales enablement almost like a checkbox, right? Like you do your enablement, you build your one pager, you make your slides, whatever. It's like boom, boom. This is a static checklist. And yes, okay, we've done it. But how can sales actually use that in real time and be self sufficient and find these resources? And I think we've come a long way with AI that we can actually like build things for sales to be self supportive, which is what a concept. And yeah, it's maybe easier than you think so that's, we'll get into that today.

Speaker B: Um, moving from stack to dynamic content can mean a lot of. So what does that actually mean? Because if we do it wrong as PMMs, that could be uh, our ticket out of, out of town. Right. And not a great way.

Speaker A: I think there's so many, uh, you have to define what does dynamic mean and how to move there. I was at a previous company and our sales enablement was every week, one hour you're going to have a sales enablement session. Who knows what it's about, who knows what you're building all as a product marketers, you got to have that deck ready and the story. And so I think when you're trying to figure out how to like really scale in a way that's like fun and interesting for your reps that they can actually remember, there's a couple things I think to, to keep in mind to how to make it dynamic. So one of the things that I do now is I build a notebook. Lm, um, for all of my different products and we'll talk about kind of maybe what goes into that. What are some best practices, maybe some things you haven't thought of. But I build a notebook for each of my products. So if I'm launching something new, sales need to learn about or ask questions like literally Q and A as maybe they're having conversations on a call. What can this feature do not do? What can it replace all of those things. They can chat within that notebook because it is specific to that product. Now you can do that for competitors too, right? If there's nuances about a deal, it could be like a rip and replace or what else is in their stack or what customers like them in their industry, their size, whatever. What have they done with our tool against a competitor? Uh, who within that list has maybe replaced a competitor, things like that. So I have different notebooks for my key products and for my key competitors that I throw everything into again so that sales can have that like interaction of hey, I have this question, this moment, I'm panicking. I don't know what the answer is and get that kind of feedback much more, more quickly.

Speaker B: Thinking about this idea of random acts of enablement. That's that this tool or what you're talking about is going against that, which is phenomenal, uh, curiosity here for me, like what made you choose notebook as kind of the resource. I know there's a lot of different ways to do with so many different tools.

Speaker A: So I think it's better at synthesizing things that uh, is more, I don't know, research specific or whatever. And the way that you can interact with notebook kind of changes based on how the rep wants to learn. So yes you can do written Q A but within notebook you can also do that like podcast style audio version of this stuff. So if you have all the content, you have your brain, your library, whatever in that page. If a sales rep wants like an auditory learning version again about a specific question that they have or a use case perhaps like they can put that question in and build an audio file from that information based on what the content you've put in there. So it like adjusts to how they want to learn. I m am not personally an auditory person so like that doesn't work for me. But some people love that it can be podcast style and you can change like the tone of what the rep one. So again back to this like self service idea. They can do this themselves. You don't have to curate the podcast and have two sales reps listen to it. Whatever it might be, they can do it themselves. And again it's just like a good. The way that it centralizes information by product, by competitor I think is a little bit easier than ChatGPT. I guess the other thing to note is if your organization has an enterprise license for another LLM, I'm not saying it has to be notebook, but I think it's just a little more flexible in terms of outputs and how you organize information.

Speaker B: And there's some strengths, some have better strengths than others too, which is a big thing. Uh, are you piping any of your like your call recordings through this as well? Or is this strictly content that you've created and you're uploading for easier access?

Speaker A: I get this question all the time. They're like, what do I put to build this brain, to build this context? That's actually interesting because if you think about it, if you don't have any very differentiated content in like this brain that you're building, you're going to get the same output. I've actually seen situations where you're saying in chatgpt, whatever and there's. You haven't given it any context and it's okay, how does my product compare against this competitor? It often pulls from like probably their competitive landing pages, right? And it actually says, oh, the competitor is better in these different ways because they have a landing page or they're sourcing this information from other stuff a competitor has built. So it actually worked not in that person's favor. It was very interesting because again, the competitor just had more content about how they're better than the other company. The other thing that's interesting is again, if you don't have this, I don't know, like hot takes it piped into your system is that you're going to get the same answer or similar answer. Not always exactly the same, but again, it's, uh. I don't know, uh, we'll talk about competitive for a second. If you're typing in how do I compare? And you're typing in the same thing, it's pulling from a similar set of sources. Right. You're not giving it anything interesting to pull from. So everyone's going to get the same answer. And so your competitive differentiator all kind of blends together because there's no difference in what it's pulling from. It's. It's literally the same output that just. It melts into each other. It's very silly. So back to your main quest. Oh, go ahead. I was like, you've got some Apple.

Speaker B: It's an Ableman slop, basically.

Speaker A: Uh, yeah. Oh, it's the same crap. And you're like, oh, yeah, here's a differentiator sales team. I'm like, that's not actually a differentiator. Like, where did this come from? And transcripts can help anything that you have that is voice of customer, like I call it voc to pov, voice of customer to a point of view like that obviously has to be in there, but that's almost. To me, that's like too obvious. Of course I'm going to upload my customer transcripts. Pros and cons. A lot of LLMs limit how many transcripts you can upload. Sometimes it's only 10 and that's not enough to get what you need. So if you can prioritize an LLM that has a connection to your gong, to your chorus, whatever, that's super helpful because there are weird limits that happen like that. But yes, in my head, the obvious answer is, of course you're going to upload those transcripts because there's nothing better than the voice of the customer. I see. What I think is maybe, I don't know, things you might not think about. A little bit different than what you would consider your normal stuff is this. I call it like high signal, low noise documents. And let me explain what I mean by that. So we'll go back to competitive intelligence, the under the COVID stuff, the like hot takes or the ugly competitor intelligence that you have now. I probably shouldn't say it that way. But like again, the LLMs are going to pull anything public facing. We know that cool customer reviews, like competitive comparison landing page is fine. What you need to pull are those like weird, interesting hot takes. I think you had someone else on who talked about like secret shoppers. Right? Are finding people that had worked at the company before. That's the information you want because that's not the information. That's like very specific to you. The more you can get granular into like the under the covers, like the ugly competitive intelligence. Do it in like a responsible way. But it's just, I uh, call it my dirty laundry list. Again, maybe not the right phrase. But like exactly where your competitor is breaking that's knife you want to twist to like really make it something different than what you're going to find on just oh, hey, do competitive research and go through customer reviews. I'm like, yeah, of course. But you can dig deeper than that. There's probably four. I don't know if we have time. There's four other things that I do that's different, but I don't know, do we want to go through them?

Speaker B: Eh, let it fly. Uh, why not?

Speaker A: Uh, screw it. Why not? So many thoughts. How did I sign up for just 20 minutes? This is what you get.

Speaker B: You know, we'll let the content dictate the length. You know, it's, it's my podcast. I could do whatever I want.

Speaker A: You do it. God, isn't that nice? This is this the moments. We love being product marketers. We do what we want. I'm a control freak, so I appreciate that. Okay, that's one. Find the ugly side of competitive intelligence that you can put in that an LLM wouldn't have or wouldn't pull from some public facing documents. Another thing that I think about a lot is I call it the if then objection vault. Because a lot of people will put in like an FAQ, right? But the way that LLMs kind of work or the way that they're structured, it's. They need an objection and a response matrix. Like they need that structure to really do something. Well, when you think about like how a sales rep is talking, almost like an if then statement, if you remember, like old school Excel, like if something happens, this happens. Sales is kind of the same way. It's like very logical. And so AI needs that branching logic. The same way you build an agent, the same way you build a customer journey, like needs to think like step by step. So I have a table and so my Table says if the prospect says objection, says this objection, then pivot to this certain value prop and provide the specific pain point. So it's. I don't know if they say we lack a certain feature, acknowledge it's on the roadmap for Q3, but then pivot to how this other feature solves the underlying business problem today, right? It's like that. If then statement becomes so clear for the LLM to give a response that actually helps sales in that moment. It's not just, oh, hey, we have that coming in Q3. It's no, we, yes, we have that in coming in Q3. And here's what you can do right now based on what products we have or how we can solve the pain point. Right? Because it's not just about the products we have, which sales loves to talk about. It's how we solve an actual problem. And so that kind of way that you set it up makes the LM really understand how to give a response in that moment that isn't just what's on the roadmap. Another one I think about, uh, uh, I realize I have such weird names for these things. Like, you know, you're. You've probably gotten some professional advice at some point. Keep the receipts. This is similar. So, uh, AI is like really good about sounding confident. Like, how many times have you asked it? I don't know, cite something, cite some sort of resource. And it's here you go. And it's, oh, is that link, like legit? Is that a real link? And it's, no, just kidding, sorry. Or asking a question. Oops, you're right. Didn't mean that. It's good at sounding confident, but it's not. So when you think about like, uh, maybe they're before and after. So you'll upload your customer stories, right? Yes, again, of course, Voice to customer. All the transcripts, customer stories, of course. But instead of a lot of customer stories are written pretty vaguely like, it depends on how yours are written. So like, for example, we helped company a improve efficiency, right? What customers case study doesn't say that, but instead you want it. You have to be specific now, whether you're rewriting this or like doing a separate document that's more specific, you have to say, we helped this company reduce time to value by 14% in the first 30 days by doing this. Like, how did you help them do this? So in some cases I've seen like spreadsheets where it's, here's the industry, the problem, the metric that was improved and then the hero quote again. And you'll see alums ranks this stuff like they love these like matrixes or they'll spit them out themselves. So if you have all of your customer stories in that kind of format, they're more likely to give the how and the why this happened. How did, what did I say? Improve efficiency, right? Like it will pull that better than just saying hey yeah, a customer like you like improved efficiency 20%. Okay, great. Again, you're. You've already lost the sale because that means nothing. It's just again, yeah, enablement slop is like the best way to say it.

Speaker B: Being more specific with the numbers in case studies. I know just in writing in general is one of the biggest things I picked up and learned. So I'm glad you brought that up. It's. It's amazing what human. The human brain will latch onto much quicker and accept as viable as opposed to just completely discrediting it from off the bat.

Speaker A: And you have these stories, they're hard, we know they're hard to get. Make sure that the LLM can actually pull the right stuff out of them. Like it's a thing. Last year.

Speaker B: Keep cooking, keep cooking.

Speaker A: Yeah, last year, I promise. And this is maybe no surprise, but a lot of people don't think about this like the anti Persona, anti sea icp, who not to sell to. I say this because in so many cases AI will hallucinate and say, oh, your product is perfect for this person. You're like, no it's not. So that's really important. Then finally, this one's kind of easy. But if you have any kind of like internal vernacular. So if you're putting in these customer transcripts and like your rep calls a feature the rocket, I don't know, they have their own like way of saying it, but the website calls it like, I don't know, an automated scaling module. Right. That doesn't connect within the LLM. They can't say that this equals this, that the LLM doesn't know that. So if there's internal lingo that you use, it's almost like translating into a different language. Put that in as well so that it catches that in the transcripts. And if you have something like glean Slack channels, whatever, so it can parse what those things are and think about

Speaker B: this like, and you've done all this and I'm assuming multiple times and it's shaped into almost like a seven step process. I'm sure you have a dope name for that too. But like tactically walk us through At a highest level of like how you're thinking about actually turning all of this into a usable playbook for your teams.

Speaker A: Yeah, if I would say like how to start tomorrow? And yeah, of course we're product marketers. Of course you have a framework or a matrix, whatever. So I feel if I were to get started tomorrow, I call it the three by three. Here we are. So first is start with the three. I uh, call them top three scariest questions. But you can ask your sales leader what are the three questions that your reps always freeze out when they hear this question? Like where do they get stuck? You can tell on some of the transcripts that there are like long pauses or I'm not sure or oh, let me get back to you. You can also mine that from like your customer calls if you have gonger course. But if you ask your sales leaders what are the three questions that come up that your reps just freeze at, like start there. Then you can write a three sentence like response to each of those questions. Like very force yourself to be concise, force yourself to be very specific and then build uh, a like custom GBT or whatever you want with all the context that we said. Give it those questions, give it your response, those three sentence responses. So three scariest questions, three sentences per question into your like again, custom GPT or notebook, lm, whatever it is. And then give your three champion reps, your three favorite salespeople, ideally reps that are doing really well at this and let them use that, that like custom GPT for a week to prep for calls, to answer slack messages, emails, whatever it is. And this is like your minimum viable product, right? This is like your test bed for something that needs to turn into long form enablement. So start there. Very low level lift. You're not having to buy new tools, you're not having to do anything crazy. It's just a really good place to start so that when you get comfortable with like what you're uploading and thinking about how you're uploading. But sales is also getting comfortable with this process of hey, I can also self serve. I have a custom GPT that I can use to start answering questions on these three specific things that I get questions about all the time.

Speaker B: And then the other question I always have is because I've dealt with this for two plus years now, hallucinations and just uh, garbly junk. Is there anything that as you're building this framework from the ground up that you're doing specifically to make sure you can avoid as much of that as possible.

Speaker A: Yeah. So I do have prompts and I'm happy to share them that are like, very specific. Once you feel good about the content and the context that you have in your notebook or whatever it is, you can give it very specific prompts to say, hey, do not leave the context of what's in here. Don't look at outside research. You can add links, like, if there are very specific links that you want to include, great news, articles, whatever. But when you tell it to not leave the confines of the information that you have added, it tends to not hallucinate so much. Again, like the frameworks that you're giving it, the way that things are formatted that you feed into, it does make a difference. But again, something as simple as a prompt or you can even push back, right? And say, hey, this doesn't sound right. Like, where did you get this information? And guess what? If they like went outside the parameters that you gave it, be like. And I haven't had that happen often. I've had to write these prompts because of the hallucinations. Right. Like, I've had that test, bubby. But you can always push back on your LLM nicely in case like Terminator happens and AI takes over. You can nicely ask, hey, friend, where did you get this information? I'm not sure every.

Speaker B: Every now and again, like, if I haven't used ChatGPT for a minute, but it would say like, hey dude, like, this is perfect. This is this. And I'd be like, no, it's not at all correct. Like, oh, you're right. Like, no, stop. So I love that I started to learn very quickly that keywords put in there not to use, like anti words was huge when I built custom gbts. Michelle, as we're around the corner, there's a lot of good stuff and I know that you had mentioned that you potentially have some resources that I could hyperlink in this episode before we kick things can and say we are doing good on a Friday afternoon. We're recording this. What are the three major key takeaways if a PMM is listening to this, nodding their heads, actively saying, wow, Michelle gets me. Like, what are the three things you want them to go? Basically write down and think about tomorrow.

Speaker A: And I'm happy to share what I start your notebooks, like throw stuff in there, everything. You can start centralizing your data. You will thank yourself down the road when you get that request. And you have sales like again, just playing around with something like custom GPT, a notebook, whatever that will save you so much headache and so many panic slack. So as you're building this stuff, think about it. Make your own checklist and start building that now. It'll absolutely help you in the future. Number two, I would say experiment. If we had an hour, we could go into all these different things. But there are so many cool ways that AI is changing how we do sales enablement. This is very foundational, like a very easy way to get started. There are role playing for reps. There is like the podcast, like I said, like, there's so many things to do and there's different ways people learn. So experiment, use AI, try. There's so many free tools. Like I said, give it a try and see what works. It can't hurt. And you might be surprised. Like I said, role playing has been really big for us with what works and, like what your reps get excited about. And that, uh, start to remember things. Because it's one thing to see something on a slide, it's another thing to actually say it. And the role playing tools I used, reps actually have to speak, speak to the AI. And then three, probably most importantly is don't forget the customer. We are so needed. Even in this crazy world of AI, there nothing beats like actually talking to a customer, actually sharing their story. Bring your customers onto enablement calls. I think my most successful enablement sessions has been a customer actually joining us and saying how they're using the tool. Because reps will remember that and remember stories. Like, at the end of the day, it's human to human. Like, and stories are so important about everything when it comes to enablement. Like, nothing, nothing beats that human to human conversation. So as. As cool as AI is, as much as I love it, do not forget it's like a low level or low impact, whatever. For high good results, ways to bring your customer onto a call.

Speaker B: Low hanging fruit.

Speaker A: I was like, don't do it, Michelle.

Speaker B: It's at the intersection of AI and enablement. Yeah, I've done really good at not saying buzzwords until this episode.

Speaker A: I've ruined it.

Speaker B: For me, the key takeaway was, I think, garbage in, garbage out. And you really, really did a great job at making sure to set guardrails to prevent that from happening. So thank you for reminding me to make sure that my alum doesn't keep calling me dude and giving me completely weird answers. Michelle, even though technically it's a little bit longer, I still think the gems you dropped are well worth it. Thank you for hanging out with me and really being generous with sharing that and some of the resources that you have to help every PMM that wants to take that next step in AI enhanced enablement and going from static to dynamic structures. A leg up above everyone else.

Speaker A: Love it. Yeah. Experiment, have fun, enjoy. Share what you're learning. Because we're all in this together. So thank you for having me. This has been fun.

Speaker B: Sweet. Michelle, have a good one. And to those who are still listening, this has been another episode of product marketing for you. Have a good one.

Speaker A: Sam.