In short
Using AI to run a multi-layer message gap analysis for a specific underperforming sales segment, turning call/win-loss/NPS/market signals into updated messaging and recommendations.
Guest
Emily Pick, product marketing leader with 10+ years in SaaS (Series A–F), now at Docebo (enterprise learning platform). Background includes translating messy cross-functional signals into clear PMM stories; known for practical LinkedIn posts on competitive differentiation, buyer expectations, and AI-assisted PMM work.
Key claims
- A message gap analysis that used to take weeks/months can be completed in ~2 hours with the right data + AI tooling.
- Control hallucinations via audit trails (linking outputs back to sources and transcripts) plus spot checks.
- Validate messaging internally with stakeholders, then rely on market testing (ads/top-of-funnel) and iterate.
Notable examples
- At Docebo, a segment closed 6–7% below benchmark; drop-off after demos led to the hypothesis of mismatched messaging vs demo expectations.
- Inputs included Gong call recordings, Glean agents for sentiment trends, win-loss data (segmented by persona/use case), and Perplexity with a curated influencer/analyst source list.
- Output: updated messaging by persona/vertical with value drivers and KPIs, shipped to her boss same day; then used by demand gen/BDR to update ads.
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Chapters
Tap a time to open that second in VOThe Pre-AI Era and Its Challenges
0:46 to 1:40
A reflection on the time-consuming processes in product marketing before AI.
“No, I'm kidding, but I really do think you'll remember what this feeling is like.”
Introducing Today's Guest: Emily Pick
1:41 to 2:29
Meet Emily Pick, a leader in product marketing with a unique background.
“an accidental marketer who, while starting with an education in human physiology, followed opportunities and landed as a leader in product marketing.”
Understanding Decebo: The Learning Platform
2:30 to 3:54
Emily explains what Decebo is and its role in learning and development.
“Also, I want the audience to know that Elle actually spelled out phonetically how to say DeCebo.”
The Importance of AI in Message Gap Analysis
3:55 to 5:00
Discussion on how AI can enhance product marketing strategies.
“We primarily serve enterprise audiences, internal and external.”
Identifying the Performance Gap at Decebo
5:01 to 6:34
Emily shares insights on performance gaps in sales processes at Decebo.
“And I feel like this is such an important one with all of the tools that are available to us.”
The Steps for Conducting an Analysis
6:35 to 7:50
Exploration of the first steps in conducting a message gap analysis.
“So the hypothesis that we came up with from that is that, hey, we must not be telling the market the right message and we must not be meeting their expectations when they do come into a demo with us.”
Gathering Insights and Data for Analysis
7:51 to 10:10
Emily discusses how to gather relevant data and insights for effective analysis.
“and talking through what was happening at Jochevo, like let's actually pretend for a moment that you've now started, we're in the future, and you've now started at a new job and the same kind of signal is happening.”
Utilizing Technology for Efficient Analysis
10:11 to 11:40
The role of technological tools in streamlining the analysis process.
“or Reddit, we're seeing now, especially with ChatGPT and all the LLMs, they're really surfacing Reddit and having that, I guess, almost unfiltered voice of the market coming through.”
Introduction to the Audience Insights Agent
14:03 to 15:11
Learn how the audience insights agent works with Salesforce data for analysis.
“Our demand operations team built us an agent using, I think it's called Unity apps, but we call it our audience insights agent.”
Gathering and Defining Insights
15:12 to 16:19
Discover the process of defining and gathering insights from calls.
“Yeah, you've got to have a full seat, which unfortunately I do not have.”
Show all 30 chapters
Controlling for AI Hallucinations
16:20 to 17:24
Understand the importance of monitoring AI outputs for accuracy.
“The first thing that I want to dig into is that there's a call out between your step one and step two.”
The Importance of Spot Checking Data
17:25 to 18:36
Learn why frequent checks are necessary for ensuring data integrity.
“And so I had to do spot check throughout.”
Proofreading and Data Analysis
18:37 to 19:49
Explore the parallels between proofreading and validating data insights.
“I'm going to click on where it came from.”
Alternative Tools for Data Analysis
19:50 to 21:27
Discover manual workarounds for data analysis when tools are limited.
“Or didn't you always proofread a paper before you turned it in in school, right?”
Using AI for Custom Analysis
21:28 to 23:24
Learn how to leverage AI tools for custom analysis and reporting.
“I actually have another podcast episode on that topic.”
Conducting Messaging Audits
23:25 to 26:38
Understand how to perform a messaging audit to align with customer expectations.
“not sure what you call them in quad, but kind of use your tool of choice just as far as creating a custom GPT.”
Final Insights and Implementation
26:39 to 28:00
Learn about the implementation of insights and messaging adjustments.
“are these things that we're actually connecting with them.”
Initial Insights on Tool Validation
28:00 to 29:25
Learn about the importance of validating assumptions with experienced stakeholders.
“I was like, tell me if it's garbage or not.”
Navigating Tools for Effective Messaging
29:25 to 31:30
Explore how understanding the full capabilities of tools can enhance messaging.
“I thought you'd get back to us in a week.”
Analyzing Tool Usage in Marketing
31:30 to 34:18
Discuss the importance of identifying secondary use cases for marketing tools.
“You've gotten, also gotten some trainings, which you're lucky to have.”
The Continuous Cycle of Marketing Validation
34:18 to 36:09
Understand the ongoing process of testing and optimizing marketing messages.
“Also, this is the direction that we're going as a company.”
Leveraging AI in Product Marketing
36:09 to 38:48
Learn how to effectively use AI tools for product marketing tasks.
“And we'll see like what worked, what didn't work.”
Advice for Building AI Agents
38:48 to 40:55
Get practical tips for leveraging AI tools in product marketing strategies.
“Like I said, I recently joined Echebo in the interview process.”
Critique of Granola's AI Note Taker
42:04 to 43:38
Discussion on the usefulness of Granola's AI note-taking capabilities.
“So they are an AI note taker and I was really introduced to them since I joined Echebo and I no longer know what my days would look like without Granola really listening in on most every call.”
Evaluating Granola's Messaging
43:38 to 45:51
Analyzing the effectiveness of Granola's marketing message and its emotional resonance.
“I love just how simple and straightforward their message is.”
Unique Selling Points of Granola
45:51 to 48:31
Exploring what makes Granola's note-taking service stand out from competitors.
“I think that's such a clever tip hitting in on the emotional aspect of it.”
The Importance of Clear Messaging in Marketing
48:31 to 50:43
Discussing the significance of straightforward messaging in product marketing.
“But like, I think these are the kind of stories that you you don't necessarily drag out of people from a LinkedIn post or from a Twitter post or whatever.”
Moments of Gratitude in Product Marketing
50:43 to 52:40
Acknowledging the importance of mentorship and community in product marketing.
“So Emily, one thing I like to make space for and this podcast is a moment of gratitude because in product marketing, we never get to where we are alone.”
Influential Leaders in the PMM Journey
52:40 to 55:51
Sharing experiences and gratitude towards influential figures in the speaker's career.
“Who has brought you to this point in your career?”
Gratitude for Influential Leaders
56:01 to 56:55
Reflecting on the impact of great leaders in one's career.
“I know I've said this a few times, but you are so lucky.”
Transcript
Automatic transcript. May contain errors.0:02Emily Pick:Hi, friends. Welcome to Product Marketing Adventures, the only show that takes you beyond theory and into the real execution of product marketing. My name is Elle Grossenbacher, your trusted guide and host of this show. In each episode, I bring in expert product marketers to co-host two segments of this show. First, a case study example of their work, followed by a messaging critique for companies we admire. So if you're a PMM looking for practical guidance and inspiration, you're in the right place. Are you ready? Let's go.
0:39Emily Pick:Gather round PMMs. Let me take you back to a moment in time, back to the pre-AI era. No, I'm kidding, but I really do think you'll remember what this feeling is like. Remember back in the day when you would decide to test out some messaging or maybe run an analysis on customer feedback. And it would take like weeks or even months. I remember being a PMM at Twilio and putting together a quarterly customer insights report. And it was quarterly because that is how long it took to gather the insights, perform the analysis and all of that. Well, not anymore. So now I'm going to put on my Captain Obvious hat and say, AI has changed the game.
1:20Emily Pick:And our guest today has the perfect example. She used AI to run a multi-layered message gap analysis in, silent drum roll please, two hours. Not two weeks, not two months, two hours. With that, it is my pleasure to have Emily Pick on the show. Emily is what some people might say, an accidental marketer who, while starting with an education in human physiology, followed opportunities and landed as a leader in product marketing. With more than 10 years in SaaS, she has worked through the gamut of series A through F companies and now sits in her first enterprise-sized public company at Tochebo. And if you haven't caught one of Emily's viral posts on LinkedIn, let me fill you in.
2:07Emily Pick:She's known for translating messy cross-functional signals into clear stories that help PMMs win, whether that's competitive differentiation, buyer expectations, or how to use AI to be a stronger PMM. She's so generous with her time and shares practical guidance and real world learnings every week. Emily, it's amazing to have you on the show. Elle, far too kind, first of all. Also, I want the audience to know that Elle actually spelled out phonetically how to say DeCebo. That has been probably the number one thing since I joined the company. It's spelled D-O-C-E-B-O. And the very first time I heard about it, I was actually working for another company where Decebo was one of our customers.
2:54And I was getting ready to interview them for a customer story. And I was chatting with our head of content. And I made a comment about how I was going to be interviewing someone at Decebo. And he was like, absolutely not. You are not allowed to walk into that conversation and say Decebo. I can tell you that I was going to say that's probably the number one thing that I've done since I've been here is just teach people. It's as my former head of convent said, Decebo. So it is an Italian based or Italian founded company. So great job on the pronunciation. And I actually think one of the things that I made me smile when I read our notes getting ready for the show.
3:33Emily Pick:Thank you. I love it. Yes, full disclosure, I totally admit I spelled it out phonetically so that I wouldn't go Docobo? Docobo? I mean, we've heard every iteration. I know. But Docebo is really fun to say. So let's start actually by giving the listeners some context. What is Docebo exactly? Yeah. So Docebo is a learning platform. We primarily serve enterprise audiences, internal and external. So if you think about there, there is a bit of the connotation of the old school LMS. We are taking a step beyond that, though. So we not only are serving your internal audiences, doing things like employee training, compliance, onboarding.
4:19We're also serving external use cases like customer education, partner training, things like franchise training, really the full spectrum of learning across any enterprise. So if you are getting ready to onboard into a new company, there's a strong possibility you've got Decebo and you don't know it.
4:39Emily Pick:How fun. Yeah. With my experience at Cisco, I know they were like huge in courses and certifications and like thousands and thousands and thousands of IT professionals and developers and software architects would always go in and have to go through some kind of learning platform. So very cool. Thanks for the background. So the topic of today's episode is about using AI for a message gap analysis. And I feel like this is such an important one with all of the tools that are available to us. And I feel like a lot of PMMs out there don't know just how to navigate that. So for the first segment of our show, I want to start with that as our case study and how you used AI to uncover that gap in your own messaging.
5:25Emily Pick:Tell me more about what was going on at Decebo when you kicked off this analysis or you realized that like this analysis had to happen. Yes. So I actually have only been at Decebo two, two and a half, close to three months now. So still pretty fresh into the org, still fresh. This is my first time in the learning and development space. So lots to learn. And we'll talk about how AI has been truly helpful in that regard as well. But just from a situational context, our CRO, he has a former mentor that he likes to bring in to come in and do a full end-to-end pipeline analysis. Where are we sourcing our leads from?
6:03How are they moving through the pipe? Also, how are they moving through on the back end, all of our lead gen systems? Really a truly inclusive process. That's really cool. Yeah, it's pretty great. The takeaways so far have been pretty dynamic. But one of the things that she really found is that we had a specific segment that was closing at a rate 6 % to 7 % lower than our benchmark. And so one of the things that we had talked about was, okay, where are we really seeing this start to fall off in the pipe? And it really was after that demo stage. So the hypothesis that we came up with from that is that, hey, we must not be telling the market the right message and we must not be meeting their expectations when they do come into a demo with us.
6:49So that's the task that I needed to investigate. What we needed to understand was, hey, when this segment comes in, what are their expectations? When we win opportunities, why is that? Why is it that we lose? Does it change based on the personas that are involved? Does it change based on the size of the company or the industry? And most importantly, what has happened in the past with our product and where are we heading in the future so that we know where we can lean into or out of based on the output of these other investigations?
7:23Emily Pick:So basically, you have the segment that you realized was not performing at the same level as all of the other segments. and the task at hand was like, hey, let's go figure out what the heck is going on with all of the leads that are falling off for this one particular segment. Obviously, you used AI, right, to go forward and do some of that investigation to figure out what the heck is happening with this segment and why isn't it performing well. So while we're kind of going back in time and talking through what was happening at Jochevo, like let's actually pretend for a moment that you've now started, we're in the future, and you've now started at a new job and the same kind of signal is happening.
8:06Emily Pick:You've got a segment that's underperforming to expectations, but you have the playbook now. So let's talk about it. What does step one look like now that you've identified that signal that you've got to go do some exploration? First and foremost, I'm going to identify what are the inputs that I need to do the analysis. Some things have not changed from the before times, from the pre-AI times. We know we need to have our win-loss data. We know that we need to have access to things like NPS scores so we can track how things have changed over time, which direction we're trending, just to see if we can validate some of those assumptions that we're making.
8:46We also want to be able to kind of pull insights from our customer conversations. I think we all at this point have had access to some sort of call recording tool. We use Gong here at DeCebo. In the before times, I would have gone through, I would have probably done a keyword search. I would have tried to pull together any insights that I could from conversations that I thought were relevant, download the transcripts, just really being able to pull together those insights. Yeah, that's like a full day's work right there. At least.
9:20Emily Pick:And you haven't even done the analysis yet. You're just gathering the information. Yeah, I was going to say, like, there's some things we'll talk about as we get into this, but just the advancements of technology. For context, this request came to me around noon on a Friday, and I signed off around 2 p.m. having it fully done and out the door to my boss for first rounds of reviews. So a little different than the way it used to be. But I also think just as far as trying to pull together insights as far as what are trends within the industry, I mean, I think about where do we source our insights from now?
10:00Obviously, we go to analyst reports and we go to those trusted voices within the space. If there's an influencer on LinkedIn or a creator that we have a lot of respect for, maybe that's someone whose insights we go try to pull from. or Reddit, we're seeing now, especially with ChatGPT and all the LLMs, they're really surfacing Reddit and having that, I guess, almost unfiltered voice of the market coming through. So things like G2, things like Gartner Peer Insights, those are all places that we would be going after just to pull together, hey, how are people talking about us online? And then also, what are their expectations for the industry and moving forward?
10:42So speaking of amount of time to go through for that, I mean, I don't know. I don't know. Even today, like how often are you trying to find a stat for a sales deck or something? Right. Oh, yeah. Right. I need this stat. Do we have any stats on this thing? And like you'll go through how many hoops to jump through and be like, oh, I have finally found this Forrester report. Oh, it's gated for three thousand dollars. I cannot pull that from there.
11:08Emily Pick:so the time that it would take previously is quite intense so that and that's just the information gathering stage we haven't even gotten into the analysis yet so step one find all of the data that you need based on the ask so pre-ai i would have done all of this yes totally exactly and that alone would have taken maybe a few weeks hopefully less there's identifying what you need. And that's like, obviously a quick list, but actually like gathering the information can take a long time. If things are gated, as you said, if it's just like, you got to chase people down, you got to, it's just, it just takes a long time to gather that information.
11:50Emily Pick:Okay. So step one is define your inputs. What's step two? Oh man. Step two is actually start performing some of the analysis or maybe not even actually that yet because now we're taking those insights and we're going to curate them, collate them together in a way that's going to be usable for the analysis. So for me, in this situation, I am incredibly lucky to have Deceva's tech stack at my disposal. So first and foremost, we have automations already set up that all of our NPS scores are automatically being surfaced. You are You're so lucky. So lucky. Like all of our NPS scores are automatically surfaced within a Slack channel.
12:31I was able to create a Glean agent. Again, we also have Glean at our disposal, which is huge. I was able to create an agent in Glean to pull those insights and look at, hey, over the past six months, how have we seen trends and sentiment relative to this segment and the message we're trying to solve for? How has that changed over time? Have we seen positive or negative improvement in these different areas? What has changed that people maybe had feedback for us from six months ago versus how did it look three months ago versus how does it look today? So understanding how that sentiment is changing based on how we've been evolving our roadmap product.
13:12Then going and pulling all of our win-loss data. We have a wonderful competitive intelligence manager here. His name is Ben Cherry, and he came to us from Clue, which also helps explain why our Clue instance is really top-notch and out of the world. Shout out Ben.
13:29Emily Pick:Great job. Ben is incredible. We love Ben. But he also just incredible record-keeping as far as he actually goes out and performs the individual interviews with customers that we have won or lost. And the way that he has it segmented the data, I can see. Cool. Exactly. Which segment, which personas were involved, which use cases. That's really cool. I can slice and dice that data in so many different ways. Then I have a tool that is beyond my expectations as a PMM. I kind of brag about it all the time. Our demand operations team built us an agent using, I think it's called Unity apps, but we call it our audience insights agent.
14:12And what it does is it looks across all of our gone calls. So based on that Salesforce data, how it's all kind of organized in there, again, being able to slice and dice. But before, as you would have had to go through and individually identify different calls, now I can just do a natural language query and say, hey, I'm looking for this segment, this industry, these personas. I want to know what they were looking for when they came in. I want to know how they reacted to our demo. I want to know what kinds of questions they were asking. I want to know whether or not we were able to adequately answer those questions.
14:53Emily Pick:And I want to know if they moved on to the next stage. I try not to say game changer very often because I feel like it got pretty cliche there for a while. But this is a legitimate game changer. It truly is. I know Gong recently launched their AI builder, I think it's called. I haven't had a chance to use it. You've got something. I've seen it. Yeah, you've got to have a full seat, which unfortunately I do not have. But the audience insights agent really alleviates any concern there. Pretty much anything from that I can also do within the audience. And we also have it attached to all of our benchmark data and our value trees and our testimonials.
15:32But I was able to pull all of that data. And then the final stage before I started the analysis was I was able to go in and I created a doc with parameters for perplexity. And luckily, because I've been a newer hire, another new hire on the team, she and I have been working on a doc that outlines who are our influencers in the space, who are the analysts that we trust, who are the voices that we listen to, where do we go for information in Reddit, all sorts of different channels. We really just outline the sub stacks we listen to or we watch the podcast we listen to. And that was actually great as far as giving perplexity parameters when I said, hey, I want to understand the state of the industry.
16:15I want to understand what people expect in this segment. I want to understand what they're trying to solve for and where they think that the entire industry on a whole is falling down today. The first thing that I want to dig into is that there's a call out between your step one and step two.
16:31Emily Pick:So in step one, you were gathering the insights. And then step two was, oh, sorry, sorry, sorry. I mean, step one was defining the insights. And then step two is gathering the insights. And part of that gathering, it sounds like, was like cleaning and organizing the data as well. Which is a step that I kind of forgot about, actually. When you're like synthesizing like different types of data, but from different sources, and then you're trying to put it all together, it's a huge headache and hassle to do it. I mean, you have like what else is there to do it? Well, now you can use AI. Thank God.
17:10But with AI, you also have to watch out for the hallucinations, which is why I'm glad you brought this up. That's also true. Because when I did post about this on LinkedIn, that was one of the number one questions, which was how are you controlling for hallucinations across this process? process. Completely fair and valid because hallucinations are inherent to AI, especially LLM models. And so I had to do spot check throughout. I was spot checking. One of the things is that between Glean, our audience insights agent, perplexity, one of the great things that they do is just the automatic audit trail.
17:45So being able to link out and say, hey, where did this stack come from? Where did this quote come from? With audience insights, We can go and look at where specifically within what call did you pull this information out of. So that was also a part of the process. I was telling a friend the other day because she also works at DeCebo and we were talking about the audience insights agent and how much we love it. We were like, ah, sometimes, though, it takes a while to pull some of those insights, which is fair. It's cross-referencing a ton of data. But I was like, hey, it's actually great because in this project, in that in-between downtime is when I was like, okay, I've put this query into the audience insights agent.
18:25It's going to get me this stuff eventually in the interim. I'm going to review the things that I pulled out of this other information. I'm going to do a spot check and just make sure that everything in here is coming out. And there's nothing that I'm like, ooh, big questions or ooh, red flag. Like, hey, here's a stat. I'm going to click on where it came from. And how many times does this happen? Even in the best LLMs where you click it and you're like, hey, this stat isn't anywhere on the page. Where did you come up with this? Luckily, I did not run into a ton of that or at least nothing where I was like, oh, this is going to completely derail this project.
19:04Nothing alarming. A lot of spot checks. I had a few people who did leave comments and were like, why didn't you run this as one giant agent? and I mean, I would love to, number one, but also I will say from a confidence in the inputs that we ultimately use to analyze, I feel more confident in them than I would have if I had allowed every single thing to run via an individual agent. Hopefully that changes one day where hallucinations won't be the problem that they are right now, But from a, is this data reliable? Are these insights reliable? You really do have to go through and spot check.
19:49Emily Pick:Yeah. Which, when you think about it, I mean, don't you always proofread a paper? Or didn't you always proofread a paper before you turned it in in school, right? Absolutely. You would never not proofread something and double check that your sources are accurate. And that's something that I think really high performers do anyway. Yeah. Listen, I'll take, spend an extra hour, you know, proofreading or checking for hallucinations. Versus, this would have taken three months. I was going to say, like the amount of time to track down all of these individual resources and insight on my own. And that's pretending that you can sit down and actually work a full day and dedicate a full day of time.
20:34Emily Pick:Like, when are you ever going to do that? Right. Exactly. It's rare. Very cool. I have another question for you about, and this may be difficult to answer, but I'm thinking, again, imagine yourself. You're at a new company now. We're in the future. You're at a new company and a very similar situation has popped up. You're recalling back on this playbook. Maybe tools have evolved a little bit, but maybe not. Let's say for the sake of argument, they have not. What if you don't have an audience insights app or tool or whatever, the one that was built internally? What if you don't have that? I mean, I think we have come a long ways with tools like Gong.
21:14I came from Clary. We had a very similar tech to their call recording or call intelligence software called Copilot. One thing that is great, maybe they're not fully at the level of sophistication that I would need for this type of cross call analysis, but they do have things that you can do like set up trackers. You can set up keyword trackers. You can set up. I actually have another podcast episode on that topic. Look at me leading the conversation. I love it. Thank you.
21:48Emily Pick:Anyway, sorry. I guess lagging the conversation. Okay, so there's the trackers that you can use. There's absolutely workarounds and they will be more manual, but you can get really the same level of output. Or at very least, you can get the outputs that you need to perform your own individual analysis or to use a tool like ChatGPT, Claude, Gemini, whomever it may be. You can get the transcripts, pull them, and then analyze them in another LLM. Yeah. Okay. So you're still using AI. It might take a little bit longer. Right. But still shorter than three weeks. Yeah. Yes. it just is gonna it's gonna like i said it's gonna that manual tracker they're they're good they're not great i'll be very transparent about that um especially relative to the degree of granularity that i'm able to get into with this specific agent but again incredibly privileged to work with a team who's fully uh dedicated to building those kinds of tools that's really cool but yes set the bar on set the way to set the bar they're gonna they're gonna ruin me They're going to ruin me.
23:00I'm going to go somewhere else one day and it's going to be like, I'm going to feel so dumb. I'm going to be like, what do you mean?
23:07Emily Pick:What do you mean I can't just automate this? Okay. So step one is define the inputs. Step two is getting access to the data, cleaning the data. What's next? All right. So finally, once we have all of this, it's getting into the actual analysis and kind alluded to it earlier, you can just use a chat GPT, create your own custom GPT or a gem, or I'm not sure what you call them in quad, but kind of use your tool of choice just as far as creating a custom GPT. So one of the things I did was I consolidated all of the insights from across all of these different sources into a single doc that I fed to a custom GPT that I was able to use to provide the relevant context for the discussion at hand.
23:54We were very fortunate when I worked at Clary as well to have someone who was an expert in ChatGPT come in and talk to us about best practices. We got some really great insights about prompting and about as you're setting these things up, really making sure that you have super tight, clear prompts up front because of how quickly it loses its memory across a lot of your conversations. So I went through and I followed some of the best practices that we had been given just as far as how to prompt this thing. And then I kind of went through and just started asking questions about all of the data that we had.
24:33Like, I didn't go into it knowing exactly what I was looking for. And exactly. I knew that I wanted to create some sort of audit. I knew that I wanted to have an output at the end that was recommendations for how to update our messaging in this segment. And I knew that I wanted to be able to put it into a format that was going to be consumable in particular by our demand gen and our SDR BDR team because they are the primary source for how we generate that top of funnel pipeline.
25:05Emily Pick:You evolve that as you are working with the data. And I know we chatted a little bit about this as we were preparing for this conversation, but I see a strong comparison to even like when we would do analysis before we ever had AI. Sometimes you would go in, and I'm just throwing this out there, like we get a combination of qualitative and quantitative information. But like if you're jumping into a spreadsheet that has like thousands of rows of data or hundreds or whatever, like sometimes you just start manipulating it. And as you get your hands on it, then you start like, oh, well, what happens if I look at it from this angle?
Read the full transcript
25:39Emily Pick:And what happens if I do this? And what happens if I play with it like that? And I feel like it's really analogous to like when you're molding clay or I don't know if you've ever mixed paint before, but when you're mixing paint with some of the most basic colors, Like, well, what happens if I add a little bit more blue or a little bit more green? And sometimes you go a little too far in the wrong direction. You're like, oh, not that. Because that accidentally happens. We'll pull a Bob Ross and be like, make this a happy tree here. Yes. Okay. So once you've done the analysis, like what happens then?
26:14Yes. So once I got all of the output that I was looking for, which was essentially a full audit of our messaging across this segment. of how we were positioning today, how much that actually resonated with the expectations that our prospects had coming in. If that was being reflected in the questions that they were asking, the things they wanted to see, really the pains that they expressed during discovery are these things that we're actually connecting with them. And we did find some disconnects. So it was not a fruitless effort. Certainly there were some things where we're like, oh, okay, that's actually not our product strength today.
26:53It might be in the future based on where we know our roadmap is going. But for what we can do today and how far ahead we're positioning, let's maybe not lean into some of these directions because that's actually, say, one of our competitors' differentiators. So we don't want to go head to head on something that we know we're not the strongest in. So instead, how do we pivot the conversations so that we're talking about the things where we lead and that we know that they care about? So ultimately what I pulled together was a document with updated messaging, with updated use case needs, who the personas were, what the KPIs that they are generally metriced on, what are our value drivers associated with each of those, how do we message those things, and then how does that change by industry or vertical.
27:45And all of that within two hours.
27:48Emily Pick:Two hours. Yes. My boss, I shipped it to my boss at the end of the workday because I'm in Pacific time zone and everybody else is in Eastern. And I was like, it's amazing. I was like, here you go. I was like, tell me if it's garbage or not. I've been here like a month and a half at this point, a month. So in my first time, again, in the industry. So I was like, I'm going based on what I think I know, but I need somebody who's a little bit more seasoned to tell me if the things that I think I know are correct. He was like, this is not the level of depth I expected or the time frame we expected.
28:26We thought you were going to be stuck on this for an entire week. And so I handed that off to him for validation. we also validated with a few of our key stakeholders across the business that, hey, these are things that we would talk about. These are things that we hear about from our customers. This is where we are strong and where we get the best reaction. And then we're able to feed that back again to our BDR and our demand gen teams. And we are pretty early in the process, but we are starting to update a lot of our ads, a lot of our top of funnel messaging just around these value props and these pinpoints and these kind of solutions focused messaging that we put together.
29:07So the output or at least the results as far as the quantitative metrics on the results of this DBB, but really the time saving from a PMN perspective, incredible.
29:21Emily Pick:Can I just say you mentioned this and I don't know, maybe you're just kind of like speaking off the cuff a little bit, but like your manager saying or your boss saying, yeah, I thought you'd get back to us in a week. I'm like, a week? Without any of these tools, try a month. Yeah. No, we've got a week with the tools. We did not think. But again, so much hinged on having the right information, having the right tools in the tech stack, and then being able to, and this is a great, I saw somebody post about this on LinkedIn recently where they were like, hey, it's really great that we have the tools.
29:56How do we start training the systems thinking for how do we tie these tools together and use them in a way that we're A, leveraging their strengths and B, we're coming up with kind of a trusted output because a lot of people may not seem, may not think to combine, you know, the four or five different tools that I was immediately like, Ooh, I know I have this. I know I have this. I know this is what it's good so that I can get all of that data I need for the analysis. And then to even be able to be like, oh, I should make sure that I have these custom prompts set and that I have a custom GPT set up with all of the context immediately woven in.
30:41Emily Pick:With experience, a seasoned PMM can figure this out, right? And I say that because with the experience, you know exactly what the journey should look like. And when I say journey, I mean like the journey of like your playbook that you just spelled out for us. You know exactly what inputs you need. You know exactly what the questions you need to ask, regardless of whatever tools you have at your disposal. You as an experienced, seasoned PMM know how to do the analysis to get the information or to get the result that you need or to answer the question at hand, right? And for your case, it was, you know, do we have a messaging gap for this particular segment?
31:21Emily Pick:And then the other experience that comes into play is just experience with the tools. So now you know what tools are out there because you've educated yourself and you've, you know, played around with them. You've gotten, also gotten some trainings, which you're lucky to have. I think that's kind of what it comes down to, at least for now in terms of, you know. Yeah, I agree. I think also one thing it comes down to is truly exploring the tools in your stack and not just thinking about them from a surface level. A good example is Glean. It's one of my absolute favorite tools. I actually have a friend who works there as an AE.
32:03We work together at Clary. He and I will send text messages just kind of bouncing ideas off of one another as far as what are things that we can do. when most people think about Glean, they think about enterprise search behind the firewall. And that is like truly my number one use case. It's funny, my friend Harjip, he's like, so many other use cases. I'm like, yeah, I know that. But this is the one that I use 90 % of my day. But like the ability to set up agents in Glean, I think is one of the biggest differentiators and some of the most savvy with AI PMMs that I know have been really diving into these tools beyond those like surface level use cases.
32:45Like we have a tool called OneUp that we use, our RSC team uses it for RFPs and they fully manage it. And it's a wonderful tool that we can also use and that I think people don't always think about what other tech do you maybe have in your company that you could get a seat in? Because there are secondary use cases that are incredible for PMMs.
33:11Emily Pick:Yeah. And that was part of your step two process and suggestion, right? When you were saying to identify the, um, like, so your first step was to identify the inputs, but then the second was to understand, like accessing the data, but you said to understand the tools and to ask whoever, ask your employer like what tools do I have available for this or and if you don't want to if that's too broad of a question maybe it's like you know what categories like you as the PMM should know what all who all the players are in those particular categories like for example call recording and ask do we have a gong or a clary or xyz yes and that was a lot of times it's I'll get like a random octa notification that will be like you have been added to this tool and I'm like well tell me more.
34:00Yeah. Ooh, what's that one? Yeah, exactly. Right. Exactly. Like even little things like I'm like, Oh, you gave me Tableau access.
34:07Emily Pick:Yeah. Well, do you know how I'm going to slice and dice that data? You don't even know I'm going to be using all this.
34:16Emily Pick:It's so funny. Okay. So back to your process. I think you might've gotten into this already, but you, now you have this beautiful report and then is the next step to like shop it around you just sent sounds like you just sent it off straight to your boss obviously right away but then is there a like a process or part of the step that you would say is to like meet with the stakeholders like with demand gen did you actually like physically get in like a virtual room and talk about it we shopped it around and said hey uh we also uh met with some of our friends over on the sales side of the house and also on the product team just to be like, hey, do these hit?
34:55Can this resonate? Also, this is the direction that we're going as a company. So let's hope it resonates. But yes, we got in the room. We got their blessing. We really didn't even make too many tweaks, like pretty minor, which is fantastic. So once we got that approved, now it's with the demand gen team. And like I said, they're starting the process of updating our ads, of getting some other top of funnel assets out the door that are kind of associated with this. And we know that it's going to be a while typically before you start to see any results off of that. But we're hoping that as we influence those opts as they come in, we'll see a bit more qualified buyer.
35:41Yeah.
35:42Emily Pick:So would you say that that's the last step is just to validate after you? So the validation actually isn't necessarily the blessing from your stakeholders. It's like the feedback from the market. Like, OK, well, now let's test it. This is the thing, right? Like marketing kind of never dies. It's never done. Never done. It's never done. It might be off your plate for a minute, but I'm going to have to. A campaign will end, but the story is still kind of evolving. It will. Yeah, exactly. And we're going to continue. And we'll see like what worked, what didn't work. Realistically, we changed nothing else in this process aside from the message.
36:20So if you're looking at an A-B test, the before versus the after, we should get some pretty solid results. Obviously, everything is not completely same, same, but we will continue to see as the results come in. Once we start seeing those results, we'll start tweaking, we'll start optimizing. optimizing we'll probably start this process all over again um yeah if you can use that agent and
36:45Emily Pick:the the process with your all your tools that you've already followed yeah i it's now that i know how to do it um probably hey who knows maybe we'll get it down to an hour and 45 minutes instead of two hours right you just keep i keep getting it faster and faster i love it it's so it's so modern PMM. Every day I am in awe. I'm simultaneously terrified of AI and so in awe of it. So I hate it for so many reasons and love it for so many reasons. I know. I think you are not, you're in company in those feelings for sure. Okay. Well, this is my last question for you before we move on to the next segment of the show.
37:28Emily Pick:What advice do you have for a product marketer who's trying to build an AI agent or like leverage some of these tools specifically to do like a message gap analysis? Just start. Like I said, I loved your clay analogy. Sometimes you don't know where it's going until it starts going. Simultaneously, don't be afraid to advocate for yourself to get a seat if there is a tech within your stack that you want to use. I think that there is immense value here. And if you can take it and prove the case of, hey, this is something that we can use to think of that increase of 6 % to 7%, even if we got it down to 3 % to 4%, what does that 3 % look like in terms of value?
38:22What does that 3 % of additional revenue to look like for this company. I think being able to frame it in the types of outcomes that you're trying to drive can help you get access to the tooling that you need. But like I said, also just dive in, just get going. The longer you wait to learn how to use these tools, the longer you wait how to use AI, the further behind you get somebody else who is learning how to do this already.
38:48Emily Pick:Exactly. Yeah. And that's not to be foreboding. It's just the way the market's moving. Like I said, I recently joined Echebo in the interview process. I was talking to a couple other companies at the time as well. And the number one question everybody asked was, how are you using AI? This isn't just something that can help you today. It's something that can help you with your career down the line. Absolutely. I love the way that you phrased that. And it can sound really daunting for someone who's never done it before or seen how it's done to, hey, go build an AI agent. I was like, what? A Glean is, I won't say it's shockingly easy, but they have so much documentation and their team is so wonderful.
39:31If you can find a contact there, if you can speak with your, if you can get your team to connect with the CSM for your account. Like we had multiple, like we had the Glean CMO come and chat with us at Clary, which isn't to say they'll pop in for everybody, but they saw the value in marketing teams building their own agents and they have plenty of people on staff who are willing. And I'm sure that's the same for pretty much any of your AI partners.
40:02Emily Pick:Yeah. And what I think, what I'll compare it to is like with any fear that you have, you've built it up in your head to be bigger than it actually is. So like my - your agent doesn't work, it's just going to pop off and say, does it work? That's what I was going to say. What's the worst that could happen? Oh, that didn't work. Yeah, it's fine. It's okay. Try again. Yeah. You're like, oh no, now I need to submit a ticket somewhere because I don't know what I'm doing. Whatever. Yeah. Whatever. Yeah. That's fine. Career is not over. And yeah, you still have time left in the day because it's AI and it only took an hour.
40:40Right. Exactly.
40:41Emily Pick:Okay. Well, this was so fun. Thank you so much for sharing such an amazing use case. And I'm so excited for the validation for you and the message gap. I hope it delivers some great results. All right. Now it's time for the next segment of our show. This is the messaging critique. It's so fun. It's where we as product marketing experts get to analyze real world messaging. And the fun part is Emily as my guest. You get to pick the company that we critique. So before we get started, I'm just going to list a few ground rules for anybody who may be new to this segment or new to the show. First, Emily is going to share the company that we're going to talk about.
41:24Emily Pick:And for this one, I try to focus on a company where either we are the customer or we know the ICP really well because it wouldn't be fair to critique messaging on a customer with an ICP that we don't even know. Like I could never critique like cybersecurity, a cybersecurity product because I don't really know that use case. Not a way. Yeah. Emily, you're going to tell me something that you're loving about it, something you wish the PMM would have done differently. And then we're just going to creatively talk about it in ways that that PMM can take the narrative to the next level. So without further ado, do you want to share the company that we are critiquing today?
42:03Emily Pick:I want us to critique with love Granola. So they are an AI note taker and I was really introduced to them since I joined Echebo and I no longer know what my days would look like without Granola really listening in on most every call. Truly. It's also to be perfectly honest, again, I work, I'm in the Pacific time zone. All of my team is in the Eastern time zone. And then also the, all of my product team is in Italy. So from a hours perspective, I tend to be a bit behind that I knew coming in at some point I was going to sleep through a 6am meeting. And that happened yesterday. It happened yesterday.
42:49Emily Pick:So it has happened to all of us. Trust me. There's not a human being on the face of the earth. I realized my mistake when I was like eating a bagel at 645. And I was like, looking at my calendar and I was like, hold on. There was a meeting 45 minutes ago. And I was like, oh, that was actually a part of that. And so I kind of went to my team's lecture and I was like, hey, did anybody record that call? And they're like, no, but we have the granola notes. And so sent me the notes. I love it. I didn't, I didn't need a recording. I got everything that I needed out of the granola notes. That's awesome.
43:23That's awesome. Okay.
43:24Emily Pick:So for those of you who want to follow along, we're going to granola.ai. That's G-R-A-N-O-L-A.ai. Okay. So now that we have a good sense of what the product is, walk us through their messaging. What's standing out to you? I love just how simple and straightforward their message is. having come from the rev tech industry and everybody has some version of a call recorder these days to be perfectly frank it's pretty commoditized there's so many options to choose from and what I love about granola is that they're literally like hey we're an AI note taker for people in back-to-back meetings I don't need any more information than that that's exactly the scenario that I am in.
44:14And that is exactly what I need help with. I have so many back-to-back calls. I think also one thing I really value when I'm in calls, when I'm in a customer call, when I'm in a call with my product team, I like to be present in the moment. And I like to be able to ask questions without being distracted by taking my notes over to the side of the screen. Every so often I will, when it's something super important, we're like, oh, I really need that nuance or that, that context. But for the most part, I love that with granola, I can be present in my meeting. And I think that that's something that's actually underrated that people don't think about.
44:52And if I were kind of angling the message at granola, I think that there are some very more emotionally resonant messages that you can add because a I love, I'm not also, I mean, I do generally have a harder time critiquing other product marketers' outputs just because, like we said, we don't have the full context. We don't know what they're being asked to do. We don't know what their mandates have been from leadership. And I think for what they have on the site today, it's great. I think that they have options to deepen it. And I think that there are some of those, like I said, emotionally resonant angles that they could take around things like, like, hey, be present in your meetings.
45:38I think that's where when I look at what they have today, I'm like, yes, but also why you versus a competitor? Because again, everyone seems to have some sort of call recorder, some sort of note taker. What is it about granolas that is special or unique?
45:55Emily Pick:I think that's such a clever tip hitting in on the emotional aspect of it. I'm imagining like, as someone who's also been in like back-to-back meetings, how I want to be present and how different, how some of those meetings, it's really hard to like context switch as well. So often when I'm in a meeting, I'll have missed the last one. If it's like a recurring meeting, because I had a different meeting that I had to attend instead. So if it's like, Oh, I didn't have time to watch the recording because that was, if I watch a recording, then I might, that's like having another meeting. And if I already have back to back meetings, then I need a meeting to watch the meeting recording.
46:32Emily Pick:And it's just Like basically, right? Because I have to block off my calendar so that I actually have time to watch the recording. And yeah, I can like scan through the notes. Like there are so many great angles that I feel like they could play with in their messaging. Like what are the scenarios in which you would be very thankful to have a tool like this? I think we're obviously thinking it from the lens of the PMM. But what is it from the lens of the sales guy or the sales gal, the sales they? uh, whoever it is that, um, like they're in back-to-back meetings. What are they missing? What are those key components?
47:10I think about if I'm going to go into a conversation, I'm in back-to-backs with different prospects. Like I have forgotten, or, oh, this is a great one. Think about like a recruiter. Think about a recruiter who is, or a hiring manager who is out there trying to like keep track of all these people.
47:30Emily Pick:Right. Like how do you, I will say one of my favorite things and like i don't know if if everybody's seen like the granola unwrapped they're kind of their take on the spotify yeah yeah crunch on their home page on their yes yes yes which is again such great i'm like please continue to enjoy like i think one of the things that makes me so sad for a lot of the enterprise companies out there is the way that even leaves creativity because you must sound buttoned up and professional. And I love that granola gets to keep the creative. You get to have crunched, like wonderful, how fantastic that you get to play with things and be like that unique.
48:12I think that when I see my notes from granola,
48:16Emily Pick:they are shockingly accurate. They summarize conversations so that it's more human and less like robotic yeah i feel like granola really understands the context of the conversations and understands the sentiment and intonations and really where are we pointing the emphasis of these conversations which i haven't necessarily seen on a lot of other tools so i was i was just gonna say if they like really like start pulling more from that like i think they have a wall of like testimonials, which is great. But like, I think these are the kind of stories that you you don't necessarily drag out of people from a LinkedIn post or from a Twitter post or whatever.
48:58I think that they have an opportunity to like really tee in on how are the users using you today? And how is that special? Or what is it that makes it uniquely relevant to their day? I totally agree with you.
49:14Emily Pick:Yeah. I mean, overall, they do such a great job telling their story. I mean, like even on their website, how they show kind of like the before and after and the way they talk about how it works. And, um, you know, basically they're doing a really good job of show, don't tell, or, or actually, which I prefer show and tell, because I think in marketing, you have to do show and tell, got to make it like stupid easy for your customers to know what you, you don't want to leave. You don't want to get your customers do the work, you know? Yes. And I think there's also like so many companies are so eager to expand and become platforms.
49:48And like when that happens is when your messaging tends to get diluted. And I think that's one of the perks of solely focusing on, you know, call notes today. You do one thing at a time. I would say who knows how they're going to evolve in the future. I see they've got some agents and workflows in the works. but I'm like when I go to granola's website I know what they do immediately sometimes the
50:15Emily Pick:simple message gets overridden in favor of butt words um and unleashed and unlocked and it's you know powered by something yeah they they tell you a lot without telling you anything and I feel like I went to granola and I was like yes AI powered note-taking for people in back-to-back meetings. Thank you. Done. Stupid. Simple. Done. Easy. I get it. Yeah, I know. I love it. Well, great job, Granola PMMs. We love your narrative. We're big fans. We're big fans. Yes. Great job. All right. So Emily, one thing I like to make space for and this podcast is a moment of gratitude because in product marketing, we never get to where we are alone.
50:59Emily Pick:We're always stealing each other's playbooks and iterating on them and learning from each other and we're all better for it. So before we wrap up, I just want to say a quick thank you for taking time to do this episode with me. It does take prep work beyond this conversation and that's you being generous with your time and willing to share with the PMM community. So thank you so much for one, just being a badass and then two, being willing to share. Well, obviously thank you for creating a space for PMMs to share their stories. I know you and I chatted when we finished up with one of our prep calls the other day.
51:37How do people learn? How do we pass on things? Over generation, that's through storytelling, right? And one of the reasons I got started posting on LinkedIn in the first place is because I was a founding PMM at a Series A startup. And I just wanted to know if I was doing things right or if people had better ways. And I really hadn't found a place where I was able to do that to that point. And LinkedIn became that place. And I love seeing other people's stories. And I love podcasts like yours where people are coming in and they're telling you, I tried this this way. It did or didn't work. And being able to really share the learnings and kind of be, hey, here's how I failed.
52:25So you don't have to. It's that building in public, but really for your own career.
52:30Emily Pick:I love it. Yeah. It's my favorite part of being a part of the PMM community. Everyone is so willing to share and teach. It's just so lovely to be here. Who has brought you to this point in your career? Who are the PMMs? I literally just wrote a LinkedIn post recently about the people who've shaped my career. PMM. And I am very fortunate to have worked with some incredible leaders, people who have kind of given me the real talk I've needed when I was early on in my career and still trying to figure out what to do with the people who have helped guide me, who've given me best practices, resources.
53:11So I want to first, sorry, it's going to be a list. First and foremost, David Tishgart and Taizzi Skogstad. They were my PMM leaders at a little company at Series B, became Series C while I was there, a startup called Data.World. Really, I had been a solo PMM before that and had not had direct guidance on how to become a PMM. And those two were the ones that really gave me the first foundation. I transitioned from content marketing into PMM. And Taizzi really gave me a kick in the pants where she was like, hey, you're trying to be a content marketer with a product focus and you need to become a real product marketer.
53:50And I remember just being a little bit dumbfounded that day. But then I was like, she's right. Like, I got a there's a lot more that I can and should be doing rather than just writing content. So I'm very thankful for them. Then Sonia Mowry, she is actually currently the director of product marketing over at skill jar which is one of the chamber's uh competitors to agree so it's been fun um now that we now that we get how funny against each other but um sonia was my uh vp of pmm at clear bits and we only got to work together for a short time but no one has probably given me more playbooks and frameworks than sonia and she had a lot of belief in me and a lot of uh like freedom she gave me a lot of freedom, but also gave me a lot of guidelines to work within.
54:41So my friend Maggie, she started out as my boss at Rattle.
54:45Emily Pick:We weren't together very long there either. I might be the angel of death. I'm not sure. But after she left, she became one of my really close friends and mentors that I've been able to bounce things off of. So we're still in touch. We worked together for maybe two months. Well, she lives in the Bahamas, which is not a bad connection to have. So I've been able to go visit her there. And then also just at Clary, I had the opportunity to work under quite a few different PMM leaders that super appreciative of Julian Savage, Ryan Boe, Ava Cavell, April Rossa. Each of them gave me something that I have walked away with.
55:24And I am so appreciative of all of them. And we are who we are because of the people who put their faith in us and teach us and train us. And now I'm currently reporting into Ben Battaglia at the Chabo. And I love Ben. He's wonderful. He is an incredible advocate. He, again, somebody who gives me a lot of freedom and also isn't afraid to give me feedback. So I think like sometimes you get bosses that just kind of pat you on the back and tell you to go on your way. And he is not that person. He will absolutely be like, hey, I think we should do this this way. And I'm like, great, cool. Let's do it.
56:00but yeah, always the people who I've been very fortunate to have leaders who push me to be better. And I just have to say, thank you.
56:08Emily Pick:I know I've said this a few times, but you are so lucky. I truly like I put that list together for that LinkedIn post. I was like, damn, I had some great leaders. Yeah. I am an amalgamation of all of them. So that's, that's how I got here today. And I wouldn't be here without them. oh i love that yeah big hug and shout out to all those amazing pm leaders and friends um and i know we've mentioned linkedin a few times now as well is that the best place for our listeners to find you that would be me i am like not on any other socials which is weird to be like linkedin is my only single media and tiktok but nobody follows each other on tiktok tiktok is for aimless scrolling of strangers.
56:52Emily Pick:So yes, LinkedIn would be LinkedIn would be the place. I love it. Well, thank you again, Emily. This has been such a wonderful conversation. And thank you PMM listeners for coming on this adventure with us today. I hope this episode leaves you with inspiration to take in the next step of your own journey.
57:20you
From the publisher
Product marketers, you know this pain. You need insights yesterday, but the data is scattered across calls, surveys, dashboards, and internal docs. In this episode, Emily Pick shares how she used AI agents to run a multi-layered messaging gap analysis in just two hours, turning what normally takes weeks into a repeatable process. The big lesson is not magic prompts. It is getting clear on inputs first, then using automation to pull the right data, and finally using AI to pressure-test narratives and messaging angles fast.
We also cover why feedback loops matter if you want the insights to stick. Emily shares how she validates findings with product, demand gen, and stakeholders so the work gets sharper and earns trust. Then we finish with a messaging critique of Granola.ai, an AI note-taking tool with a refreshingly clear promise for people in back-to-back meetings, plus a smart “Crunched” campaign. The opportunity is emotional differentiation. In a crowded market, the story is not just better notes. It is being present in the meeting.
Key Takeaways
- Start with inputs, not tools. Define the data that actually matters before you touch AI
- Automate collection so you can spend time on narrative, not admin
- Use AI to explore and test messaging angles quickly, not to replace strategy
- Build feedback loops to validate insights and increase stakeholder buy-in
- Messaging wins when it is consistent, specific, and emotionally resonant
LINKS
Messaging Critique:
https://www.granola.ai/
Connect with Emily:
LinkedIn: https://www.linkedin.com/in/emilypick/
Connect with Elle:
LinkedIn: https://www.linkedin.com/in/elle3izabeth/
