Acceptable Mistakes & Ruthless Prioritization: How Top PMMs Are Winning in AI GTM

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
.999Hey crafters.
Just a reminder, this podcast is for informational entertainment purposes only and should not be considered advice. 3 00:00:08,330.0000000002 --> 00:00:18,10.001 The views and opinions expressed our own and do not represent those of any company or business we currently work with are associated with, or have worked with in the past. 4 00:00:18,610.001 --> 00:00:20,530.001 for tuning in to the FutureCraft podcast. 5 00:00:20,800.001 --> 00:00:21,770.001 Let's get it started. 6 00:00:21,770.001 --> 00:00:29,562.81052381 uh, uh, uh, um, uh, Hey there. 7 00:00:29,562.81052381 --> 00:00:32,322.81052381 Welcome to The Future Craft Go To Market podcast. 8 00:00:32,322.81052381 --> 00:00:38,892.81052381 We're exploring how AI is changing all things go to market, from awareness to loyalty and everything in between. 9 00:00:39,132.81052381 --> 00:00:42,222.81052381 I'm Ken Roden, one of your guides on this exciting new journey. 10 00:00:43,167.81052381 --> 00:00:49,407.81052381 And I'm Erin Mills, your co-host, and together we're gonna unpack the future of AI and go to market. 11 00:00:49,707.81052381 --> 00:00:57,417.81052381 We're gonna share some best practices, how-tos and interview industry pioneers and leaders who are paving the way in AI and go to market. 12 00:00:57,877.81052381 --> 00:01:10,277.81052381 Ken, how are you paving the way? So Erin, something a little different than maybe I've talked about in the past, I've been talking to marketing leaders over the past couple weeks, and one of the challenges they're having is even though they have. 13 00:01:11,212.81052381 --> 00:01:19,552.81052381 The discussions and they're thinking about how to go agentic and really going into workflows, it's a bit intimidating and the teams just aren't ready. 14 00:01:19,552.81052381 --> 00:01:21,202.81052381 Not everyone on their team is technical. 15 00:01:21,382.81052381 --> 00:01:24,732.81052381 Someone said I don't have the Google marketing team. 16 00:01:24,732.81052381 --> 00:01:34,872.81052381 I have my team who's already maxed out, and so to help with that, I created something called the AI Sandwich Framework, and you know me very well. 17 00:01:34,902.81052381 --> 00:01:38,652.81052381 Sandwiches are my favorite food group They may be your love language. 18 00:01:38,952.81052381 --> 00:01:41,202.81052381 They might be a club sandwich. 19 00:01:41,252.81052381 --> 00:01:44,882.81052381 What's great about a club sandwich is it has layers and so does this framework. 20 00:01:45,212.81052381 --> 00:01:52,622.81052381 The idea is taking the best parts of what AI can do and allowing humans to do what they do best. 21 00:01:52,922.81052381 --> 00:01:54,512.81052381 So it breaks down. 22 00:01:55,182.81052381 --> 00:02:02,262.81052381 Step by step how to into a workflow that's not a specific person's workflow, but think more a program. 23 00:02:02,472.81052381 --> 00:02:12,992.81052381 So competitive intelligence is an example that I'm using, and the idea is it's going to save time, but also help train your team on how to think about incorporating AI while keeping the Human Element center. 24 00:02:14,47.81052381 --> 00:02:15,187.81052381 Ken, that's interesting. 25 00:02:15,307.81052381 --> 00:02:24,637.81052381 And I know you've been working with a lot of different companies, getting them at that like a hundred, if the team is at zero to 50 and you're working with teams to get them to a hundred. 26 00:02:25,37.81052381 --> 00:02:27,527.81052381 Gimme a little sample of what does it look like? What are some of the input. 27 00:02:28,157.81052381 --> 00:02:28,487.81052381 Sure. 28 00:02:28,487.81052381 --> 00:02:36,317.81052381 So if we're looking at competitive, which I think is a great one to do because it can be really impactful for a marketing team as well as a sales team. 29 00:02:36,317.81052381 --> 00:02:37,817.81052381 But there's just so much that goes into it. 30 00:02:38,147.81052381 --> 00:02:40,817.81052381 So first we start with the human element. 31 00:02:40,867.81052381 --> 00:02:42,637.81052381 And that's where we have the. 32 00:02:43,97.81052381 --> 00:02:45,107.81052381 Inputs come in to steer the conversation. 33 00:02:45,327.81052381 --> 00:02:56,667.81052381 The competition, the priorities, able to look at previous analysis, maybe it's key battlegrounds, so what landscapes, segments or geos you might be focusing on. 34 00:02:56,737.81052381 --> 00:02:59,827.81052381 Then AI can jump in and start doing the deep research. 35 00:03:00,7.81052381 --> 00:03:02,997.81052381 They can look at SEC filings, recent hirings. 36 00:03:02,997.81052381 --> 00:03:05,247.81052381 Funding customer reviews. 37 00:03:05,397.81052381 --> 00:03:11,337.81052381 They can also analyze your sales calls, chat pt, the team and enterprise function. 38 00:03:11,567.81052381 --> 00:03:24,447.81052381 Or fun has functionality that's compliant, meaning you can upload your gong calls or your ICP information in a way that's a secure and you can get reliable research from that's has sources linked to it. 39 00:03:25,197.81052381 --> 00:03:33,347.81052381 And then there's a chance for the human to come back in where they can review findings, look at the positioning gaps, they can upload their positioning to have it incorporated. 40 00:03:33,347.81052381 --> 00:03:35,57.81052381 And then you can set a messaging strategy. 41 00:03:35,247.81052381 --> 00:03:37,977.81052381 From that, I think the magic happens. 42 00:03:37,977.81052381 --> 00:03:39,117.81052381 You can get battle cards. 43 00:03:39,457.81052381 --> 00:03:43,447.81052381 Personalized objection handling for different geos for different buyers. 44 00:03:43,627.81052381 --> 00:03:51,457.81052381 You can have a custom GPT built that the seller can interact with and ask questions about, maybe product specifics versus a competitor. 45 00:03:51,607.81052381 --> 00:03:56,437.81052381 And then the human actually comes in to do the training, the rollout, and do the relationship building. 46 00:03:56,677.81052381 --> 00:04:06,592.81052381 And that's how we've seen organizations cut time by 30 to 40% which is huge because that allows the employee, the human to do meaningful work. 47 00:04:06,652.81052381 --> 00:04:08,842.81052381 So I'm really excited about it. 48 00:04:08,872.81052381 --> 00:04:11,332.81052381 I want people to give it a try and let me know what they think. 49 00:04:12,37.81052381 --> 00:04:16,302.81052381 Should people just reach out to you on LinkedIn or what's the best way to get in touch? Reach out to me on LinkedIn. 50 00:04:16,362.81052381 --> 00:04:21,752.81052381 I am happy and eager to talk to anybody and nerd out on human AI sandwiches. 51 00:04:23,257.81052381 --> 00:04:24,7.81052381 Sounds tasty. 52 00:04:24,242.81052381 --> 00:04:33,502.81052381 What about you? What have you been doing with ai? I'm gonna go a little bit lighter fair and talk a little bit about a book series I've been working on. 53 00:04:33,722.81052381 --> 00:04:35,102.81052381 I love my niece. 54 00:04:35,337.81052381 --> 00:04:38,787.81052381 She lives far away, which is a real bummer for me. 55 00:04:39,217.81052381 --> 00:04:41,707.81052381 One of the things she loves is reading. 56 00:04:41,707.81052381 --> 00:04:49,147.81052381 And so for her birthday, made a little book that she was featured with one of her favorite characters, and she absolutely loves it. 57 00:04:49,177.81052381 --> 00:04:52,980.31052381 We got it laminated and she takes it in the bath and takes it every, it's insane. 58 00:04:53,107.81052381 --> 00:04:53,857.81052381 It's very popular. 59 00:04:54,367.81052381 --> 00:04:58,27.81052381 But I thought what would be even more fun is to provide book. 60 00:04:58,377.81052381 --> 00:05:03,857.81052381 Books about us and our books about our family doing things with her. 61 00:05:03,857.81052381 --> 00:05:14,427.81052381 So she's gonna be go going golfing with uncle, she's going to be cooking with cousin and a bunch of other activities just to, keep it really fun. 62 00:05:14,427.81052381 --> 00:05:21,577.81052381 And I was able to use ChatGPT to create the characters and the illustrations, which, would've been costly before. 63 00:05:22,677.81052381 --> 00:05:24,297.81052381 That's so cool. 64 00:05:24,347.81052381 --> 00:05:25,97.81052381 I have a question for you. 65 00:05:25,97.81052381 --> 00:05:33,497.81052381 I've tried to use chat bet's image creator, and I get weird things including hands that are one hand with no fingers. 66 00:05:33,687.81052381 --> 00:05:36,392.81052381 Do you have any tips for me to get a quality image. 67 00:05:36,392.81052381 --> 00:05:42,272.81052381 I can't remember the example, but you were designing something in ChatGPT, bt, and it came out very like not. 68 00:05:44,267.81052381 --> 00:05:48,47.81052381 The design you were hoping for, or it was like very cartoonish. 69 00:05:48,47.81052381 --> 00:05:56,6.81052381 Do you remember it? I think I've gotten some weird abstract dolly type outputs when I've been trying to get a new headshot. 70 00:05:58,466.81052381 --> 00:06:04,841.81052381 So the tip I would give you is like anything AI related context is the most important piece of it. 71 00:06:04,841.81052381 --> 00:06:11,301.81052381 So give us the ai, something that you think is really good and the style, describe the style that you want. 72 00:06:11,331.81052381 --> 00:06:17,851.81052381 Maybe take an image that you really like from something else and upload it and say, I want my images to look like this. 73 00:06:18,61.81052381 --> 00:06:30,391.81052381 And then what I did with the book was took pictures that gave, really head on, really good quality photos that AI could read and then translate into the illustrations I wanted it to. 74 00:06:30,721.81052381 --> 00:06:36,631.81052381 And so again, if you can get the context, make sure it's high quality, you might get a really cute children's book out of it. 75 00:06:37,486.81052381 --> 00:06:50,136.81052381 I am thinking this might be a good idea, so I'm gonna try it out, but I gotta know, is Auggie getting a book? How could she not? For the French Bulldog lovers out there you all know they take center stage. 76 00:06:50,136.81052381 --> 00:06:54,561.81052381 So Augie is gonna be featured in the first book, but I'm thinking she probably gets her own series. 77 00:06:55,776.81052381 --> 00:07:01,86.81052381 Okay, it better be on shelves by holiday time 'cause I'm gonna be ready to buy for every kid I know. 78 00:07:01,96.81052381 --> 00:07:02,841.81052381 I'll fall in love with ugly just as much as I am. 79 00:07:03,116.81052381 --> 00:07:05,306.81052381 Hey, it may be Augie and Uncle Ken. 80 00:07:05,586.81052381 --> 00:07:05,981.81052381 Watch your mailbox. 81 00:07:07,656.81052381 --> 00:07:16,641.81052381 So Erin, who are we talking to today? We've got Rebecca Shaddix on the show today, and some folks may recognize her from a lot of really great thought leadership. 82 00:07:16,641.81052381 --> 00:07:17,751.81052381 She puts out on LinkedIn. 83 00:07:17,961.81052381 --> 00:07:19,461.81052381 She's a Forbes contributor. 84 00:07:19,761.81052381 --> 00:07:30,31.81052381 I met her about a year and a half ago and interviewed her for a role on my team she always stuck with me as somebody who was really interesting and had thoughtful things to say and innovative in ai. 85 00:07:30,241.81052381 --> 00:07:32,491.81052381 So really excited to catch up with her today. 86 00:07:33,151.81052381 --> 00:07:33,811.81052381 That's great. 87 00:07:33,981.81052381 --> 00:07:34,971.81052381 Let's go talk to her. 88 00:07:34,971.81052381 --> 00:07:42,779.13363458 uh, uh, uh, um, uh, And we're back. 89 00:07:42,779.13363458 --> 00:07:49,409.13363458 Today's guest is Rebecca Shaddix, A trusted mind in go-to-market strategy At the intersection of AI, product and marketing. 90 00:07:49,679.13363458 --> 00:07:57,829.13363458 Rebecca is the head of product and lifecycle marketing at Garner Health, where she's helping redefine how AI can power smarter healthcare decisions. 91 00:07:58,249.13363458 --> 00:08:00,829.13363458 Before that, she led go-to-market strategy for high growth. 92 00:08:00,829.13363458 --> 00:08:10,459.13363458 Ed tech brands advised dozens of SaaS teams throughout her consultancy strategica, and built a reputation for turning complex data into clear, repeatable go-to-market systems. 93 00:08:10,974.13363458 --> 00:08:12,864.13363458 Rebecca writes frequently on strategy. 94 00:08:12,894.13363458 --> 00:08:19,874.13363458 Highly recommend checking out her Forbes columns, and is known for frameworks that move fast without losing trust. 95 00:08:20,234.13363458 --> 00:08:21,404.13363458 Welcome to the show, Rebecca. 96 00:08:21,404.13363458 --> 00:08:23,114.13363458 I'm super happy to have you on the show. 97 00:08:23,164.13363458 --> 00:08:24,424.13363458 So excited to be here. 98 00:08:24,964.13363458 --> 00:08:31,289.13363458 Hey, Rebecca, to start us out, one of the things you've written about is this concept of acceptable mistakes. 99 00:08:31,469.13363458 --> 00:08:44,254.13363458 How does this framework help go to market leaders, navigate the AI implementation when rapid experimentation is so critical right now? I think it really helps make the right experimentation faster and clearer. 100 00:08:44,414.13363458 --> 00:08:52,359.13363458 it's moving so quickly and the concept of an acceptable mistake is essentially the opposite of a risk and something I like to put at the forefront of go-to-market. 101 00:08:52,664.13363458 --> 00:09:01,94.13363458 Planning, what are the things that we're willing to trade off because they don't jeopardize our ability to hit our goals and we wouldn't waste time optimizing for them. 102 00:09:01,454.13363458 --> 00:09:09,525.13363458 And because things are moving so quickly, I think it's even more important than ever to be really clear and intentional internally with the team of where we're willing to. 103 00:09:10,489.13363458 --> 00:09:11,449.13363458 Not invest. 104 00:09:11,539.13363458 --> 00:09:17,239.13363458 And that makes the experiments that are worth running, the hypotheses that we're actually testing that much clearer. 105 00:09:17,609.13363458 --> 00:09:24,719.13363458 the way I like to think about these acceptable mistakes is the way we would define experiments is really we would have some problem statement. 106 00:09:24,719.13363458 --> 00:09:25,799.13363458 The biggest business. 107 00:09:26,154.13363458 --> 00:09:31,914.13363458 Risk problem statement, et cetera, defined upfront that'll make the acceptable mistake really clear. 108 00:09:32,244.13363458 --> 00:09:37,314.13363458 From there, we'll have two to four hypotheses about what's potentially driving these problem statements. 109 00:09:37,554.13363458 --> 00:09:45,834.13363458 They should be explicit and clear, and then from there, anywhere from five to six to 12 experiments to seek to disprove those hypotheses. 110 00:09:46,234.13363458 --> 00:09:47,734.13363458 where I think a lot of. 111 00:09:48,644.13363458 --> 00:09:52,124.13363458 motion is going without momentum is not having that order. 112 00:09:52,154.13363458 --> 00:10:13,754.13363458 if you see people just running experiments like, Hey, what if we make the CTA button on our homepage green? What is the hypothesis that we're testing with that? What's the problem statement? How do we know that's a priority? And so I think I see a lot of just activity and motion and sort of chaos that because the floor to enter using AI is now so low. 113 00:10:14,84.13363458 --> 00:10:14,984.13363458 Kids can do it. 114 00:10:15,824.13363458 --> 00:10:22,214.13363458 It means that if you don't have a really clearly defined framework there, we start with our problem statement that makes the acceptable mistake clear. 115 00:10:22,334.13363458 --> 00:10:26,84.13363458 We then prioritize which of these problem statements we're working on. 116 00:10:26,114.13363458 --> 00:10:28,4.13363458 And to be clear, there's no limit to how many there could be. 117 00:10:28,4.13363458 --> 00:10:30,734.13363458 There could be 80 problem statements. 118 00:10:30,764.13363458 --> 00:10:32,564.13363458 We prioritize the ones that matter. 119 00:10:32,894.13363458 --> 00:10:37,94.13363458 Then we have hypotheses about the four problem statements we're tackling this quarter. 120 00:10:37,839.13363458 --> 00:10:39,399.13363458 Then we start running experiments. 121 00:10:39,639.13363458 --> 00:10:56,379.13363458 So everything ladders really neatly back to why we're actually doing it, and a lot of the downstream problems are solved, misalignment, lack of clarity about why we're making the investments we're making, that all goes away when at the onset, all of the strategic decision makers have been involved in that. 122 00:10:56,964.13363458 --> 00:11:00,264.13363458 So now when we report on our experiments, it's clear why they matter. 123 00:11:00,264.13363458 --> 00:11:06,174.13363458 It's not just marketing or product saying look at this change in some metric that the rest of the team doesn't understand. 124 00:11:06,474.13363458 --> 00:11:11,784.13363458 They've all bought in and there's a lot of cross-functional alignment between tackling those problem statements. 125 00:11:12,4.13363458 --> 00:11:15,184.13363458 every department's role is clear and we just go from there. 126 00:11:16,709.13363458 --> 00:11:25,199.13363458 Yeah, it's interesting because one of the things that I'm finding as I'm talking to more people using AI is the skill they really need right now is self-regulation. 127 00:11:25,379.13363458 --> 00:11:35,679.13363458 To not go try 80 experiments because AI can do so much, and it's actually staying focused on the hypotheses that are going to drive things forward And not every experiment should be run. 128 00:11:36,49.13363458 --> 00:11:42,604.13363458 Sometimes it's so clear to you based on all of the data in your brain and your understanding of the market, that some change should be made. 129 00:11:43,294.13363458 --> 00:11:43,684.13363458 Great. 130 00:11:43,864.13363458 --> 00:11:44,314.13363458 Make it. 131 00:11:44,554.13363458 --> 00:11:56,34.13363458 We may not have, statistically significant volume of traffic to test something It may not be worth slowing down some load time to put some more tool in, There may be any number of reasons why running the experiment doesn't matter. 132 00:11:56,334.13363458 --> 00:11:58,734.13363458 It could be that it's just not an important enough change. 133 00:11:59,19.13363458 --> 00:12:01,509.13363458 It could be that we're not going to get actionable data. 134 00:12:01,719.13363458 --> 00:12:11,339.13363458 There's lots of reasons to make changes that don't have to be experimented on, but I think when you're clear about what you're actually testing and why those resources become much more obvious, That makes a lot of sense. 135 00:12:11,999.13363458 --> 00:12:17,429.13363458 I am one of the victims of vibe, experimentation for sure. 136 00:12:17,819.13363458 --> 00:12:26,89.13363458 what do you think is the moment for you that sparked that AI could really reshape marketing and not just automate it, beyond sort of the experimentation phase. 137 00:12:27,319.13363458 --> 00:12:28,99.13363458 Oh man. 138 00:12:28,429.13363458 --> 00:12:35,9.13363458 seeing that ChatGPT launch was a game changer because when I first got introduced to it, it was just predicting. 139 00:12:35,354.13363458 --> 00:12:47,504.13363458 The next character, which was fun and cool, then, I don't know, two years later, it's actually synthesizing complex processes that are really complicated questions that as a human would take a ton of data to input. 140 00:12:47,984.13363458 --> 00:12:57,254.13363458 That was a really big one, but I also think just starting with a messaging idea or framework or copy and saying, here's my audience. 141 00:12:57,474.13363458 --> 00:12:58,49.13363458 This is the target. 142 00:12:58,779.13363458 --> 00:12:59,589.13363458 Conversion. 143 00:12:59,679.13363458 --> 00:13:10,679.13363458 Here's the copy I have drafted Can you make it more engaging? Come up with an alternate subject line that could drive up open rates, and then running both and saying A, it sounded great. 144 00:13:10,949.13363458 --> 00:13:17,369.13363458 B, some of the variants that AI had come up with were actually converting better than the ones that humans had. 145 00:13:17,699.13363458 --> 00:13:19,109.13363458 That was the wow. 146 00:13:19,409.13363458 --> 00:13:25,649.13363458 There really is potential for this tool in a language processing capacity to be. 147 00:13:26,189.13363458 --> 00:13:36,849.13363458 Even more effective with some controlled inputs For you did you have an aha moment? I'm curious, did you have a, oh wow, this could change everything. 148 00:13:38,654.13363458 --> 00:13:42,513.13363458 For me, the hardest thing is looking at a blank page. 149 00:13:43,3.13363458 --> 00:13:51,433.13363458 the aha moment for me was being able to use AI as a thought partner really accelerated what I was able to achieve in a shorter amount of time. 150 00:13:51,583.13363458 --> 00:13:59,23.13363458 I think of myself as a really good editor and being able to build on great ideas, but the blank page would get me hung up. 151 00:13:59,28.13363458 --> 00:14:02,8.13363458 that has been game changing overall. 152 00:14:02,248.13363458 --> 00:14:04,648.13363458 Ken, what about you, you got on this so early. 153 00:14:05,878.13363458 --> 00:14:06,808.13363458 Yeah, I agree. 154 00:14:06,808.13363458 --> 00:14:13,28.13363458 Like seeing ChatGPT I was whoa, this could help with, I quickly foundationally copywriting, but what really. 155 00:14:13,778.13363458 --> 00:14:24,948.13363458 Got me excited and a little scared honestly, was when you're able to start uploading your own documents and PDFs to ChatGPT and I was like we're not just like again, vibing here. 156 00:14:25,38.13363458 --> 00:14:33,638.13363458 We're actually able to iterate, like Aaron said, and take something and, triple its impact or multiply a content asset, personalize it. 157 00:14:33,643.13363458 --> 00:14:37,318.13363458 I was just thinking marketing's gonna be redefined in the next 10 years. 158 00:14:37,468.13363458 --> 00:14:39,868.13363458 It's gonna look totally different and very exciting. 159 00:14:40,358.13363458 --> 00:14:40,768.13363458 Yeah. 160 00:14:41,158.13363458 --> 00:14:41,728.13363458 I love that. 161 00:14:42,658.13363458 --> 00:14:46,858.13363458 Erin, I think you hit on something I hadn't thought of or articulated, but I think it's really worth. 162 00:14:47,998.13363458 --> 00:14:48,628.13363458 A conversation. 163 00:14:48,628.13363458 --> 00:14:50,278.13363458 I don't know that I've heard people talk about it. 164 00:14:50,548.13363458 --> 00:14:57,928.13363458 There's different human personality types and strengths that make using the same AI tool totally different. 165 00:14:57,928.13363458 --> 00:15:17,368.13363458 I think that's probably something missing from the current prompt engineering conversations, I have a friend who would never trust AI to generate a first draft, but she really loves having a second set of eyes on, is there a better way to convey this point, this order, this organization? So for her first draft is very human, but second draft. 166 00:15:17,368.13363458 --> 00:15:17,728.13363458 Sure. 167 00:15:17,728.13363458 --> 00:15:21,208.13363458 As she takes a third and fourth pass, I totally relate to that. 168 00:15:21,238.13363458 --> 00:15:22,198.13363458 Oh, this is blank. 169 00:15:22,258.13363458 --> 00:15:30,388.13363458 How do we make some of these ideas into something? for you as a first draft, I think it's really interesting to think about how people will find their groove. 170 00:15:30,808.13363458 --> 00:15:34,708.13363458 With how they use even the same tools totally differently. 171 00:15:34,708.13363458 --> 00:15:36,928.13363458 I don't know if I've heard anyone talk about that just yet. 172 00:15:38,353.13363458 --> 00:15:48,318.13363458 It's funny, I haven't heard it talked exactly like that, but what I have been observing and hearing people mumble about is when SaaS software came out, we got really specialized software. 173 00:15:48,318.13363458 --> 00:15:52,598.13363458 There was compensation software, optimized for SEO experts. 174 00:15:52,698.13363458 --> 00:15:58,899.13363458 There was really specific buyer journey content machine software, but this has opened up a. 175 00:15:59,428.13363458 --> 00:16:09,978.13363458 A Jack and Jill of all trades era where you can be, a marketing generalist who maybe is a T-shaped marketer who has, one area of expertise, but you're dangerous enough to know the rest of it. 176 00:16:10,38.13363458 --> 00:16:23,118.13363458 And you can actually be almost a full stack marketer and I, and that's where I think the people who are really thriving are the people who, know a little bit about marketing from a well-rounded perspective and can say I need product marketing here. 177 00:16:23,118.13363458 --> 00:16:24,498.13363458 I need go to market strategy here. 178 00:16:24,588.13363458 --> 00:16:26,28.13363458 Here's how I'm gonna drive pipeline. 179 00:16:26,28.13363458 --> 00:16:28,488.13363458 Here's my social strategy, and here's how it's all connected. 180 00:16:28,708.13363458 --> 00:16:30,838.13363458 they're able to operate much more effectively. 181 00:16:30,998.13363458 --> 00:16:41,18.1336346 I think the discernment about what's coming out because it is so ubiquitous and easy to access, discerning what makes mediocre versus exceptional strategy at this point. 182 00:16:41,988.1336346 --> 00:16:53,993.1336346 even the best model still hallucinates in ways that may not be accurate or may just be regressing to some mediocre mean, and so you have to have enough critical thinking, discernment. 183 00:16:53,993.1336346 --> 00:17:04,743.1336346 Does that make sense? Is it exceptional? Is it good? I think it's very much a junior level thought partner at this point, as opposed to something we would delegate and expect good results. 184 00:17:06,633.1336346 --> 00:17:11,223.1336346 Going back to that earlier point of people using it in different ways. 185 00:17:11,223.1336346 --> 00:17:21,753.1336346 We did on my team the exercise around the working geniuses and you could see where people are really good at ideation or there's like tenacity or things of that, and I think. 186 00:17:22,193.1336346 --> 00:17:31,133.1336346 To your point, the idea of somebody using it differently, but being able to use the same tool effectively is actually what probably makes it most powerful. 187 00:17:31,463.1336346 --> 00:17:44,713.1336346 But going back to the idea that you've got folks who think about things and tackle problems in different ways, but then can really get meaningful outputs, get to the same end point, but be able to supplement their weaknesses. 188 00:17:44,813.1336346 --> 00:17:53,103.1336346 So Rebecca, one thing that you and I have in common is that we are both product marketers and you talk about. 189 00:17:53,913.1336346 --> 00:17:58,83.1336346 Product marketing being the nerve center of a go to market strategy, which I wholeheartedly agree with. 190 00:17:58,773.1336346 --> 00:18:11,528.1336346 How is AI transformation giving product marketers competitive advantage specifically in B2B growth right now? I really think the thought partner ideation is awesome. 191 00:18:11,648.1336346 --> 00:18:16,448.1336346 There's definitely limitations to this chat-based interface, which is ubiquitous right now. 192 00:18:16,778.1336346 --> 00:18:21,798.1336346 But as far as thinking through something, being able to upload a defined, trusted set. 193 00:18:22,213.1336346 --> 00:18:37,933.1336346 Of assets and inputs, letting that be a training focal point, thinking through problems of, given what you've seen, how their campaigns perform, what we've uploaded about our ICP, given what you know about these defined sets of inputs, incorporate that. 194 00:18:38,633.1336346 --> 00:18:43,13.1336346 Help me think through how this persona would respond to this message or this offering. 195 00:18:43,223.1336346 --> 00:18:58,318.1336346 How would this compare when we are thinking about our competitive differentiation or whatever we've talked about, how does this compare and that real time thought partnership and then ability to talk through other inputs that may not be forefront in the product marketer's mind. 196 00:18:59,148.1336346 --> 00:19:01,578.1336346 Could be driving a lot of the output. 197 00:19:01,938.1336346 --> 00:19:03,138.1336346 That's really powerful. 198 00:19:03,188.1336346 --> 00:19:13,748.1336346 when I think about these guiding sets of documents that are really important for product marketers to be intentional about, super important to train our tools on it, but really important for our teams too, to be intentional. 199 00:19:14,78.1336346 --> 00:19:24,668.1336346 the benefit that AI has is it can synthesize a lot of these very different inputs and weight them in ways that it can then explain largely if you ask it in the right way. 200 00:19:25,683.1336346 --> 00:19:28,743.1336346 we can forget, as we have recency bias, we have all these things. 201 00:19:28,743.1336346 --> 00:19:32,793.1336346 We have our own biases that can be controlled with AI strengths. 202 00:19:33,33.1336346 --> 00:19:38,673.1336346 Letting it leverage that to think through elements we may not have thought of as being so important. 203 00:19:38,673.1336346 --> 00:19:42,3.1336346 Some performance trends that we may not have connected the dots through. 204 00:19:42,243.1336346 --> 00:19:48,3.1336346 Just bringing back more of these defined inputs that should be part of the decision making process. 205 00:19:48,273.1336346 --> 00:19:58,283.1336346 A document or an experiment that you ran six months ago may not be as top of mind as a focus group from yesterday, but it could be more important and could have more powerful connection. 206 00:19:58,283.1336346 --> 00:20:00,383.1336346 And that's the power I think of. 207 00:20:01,223.1336346 --> 00:20:03,173.1336346 Tools like AI to really bring that forward. 208 00:20:03,173.1336346 --> 00:20:09,983.1336346 we think about differentiating, leaning into what humans do well, training and being intentional with the outputs that we want. 209 00:20:10,343.1336346 --> 00:20:16,373.1336346 And then leaning into what AI does well, I think is really important, but never mistaking where any of those are. 210 00:20:16,373.1336346 --> 00:20:39,553.1336346 I wouldn't, at this point anyway, fully delegate any kind of decision making, And I think that's where some people are getting into trouble Not only do things have to be double and triple checked, but really just that the critical thinking layer of is this the right thing One of the things we talk about a lot on this podcast is AI and marketing in particular really allows marketers to spend more time being strategic and thoughtful. 211 00:20:39,743.1336346 --> 00:20:49,133.1336346 one of the biggest impacts I've seen AI have for product marketers is allowing them to outsource some of the tactical work. 212 00:20:49,233.1336346 --> 00:20:59,783.1336346 One of the most impactful things AI can do for product marketers is being a content multiplier for the marketing team and the voice of product marketing when they're not in the room. 213 00:20:59,883.1336346 --> 00:21:08,953.1336346 I've seen brand teams leverage product marketing, positioning, messaging frameworks when they're doing their global campaigns, going down to regional specific strategies. 214 00:21:09,73.1336346 --> 00:21:09,793.1336346 I've seen. 215 00:21:10,3.1336346 --> 00:21:22,863.1336346 A ton of personalization across multiple personas for the same product release because they were able to leverage a marketing program and strategy developed by the product marketing team that customer marketing and demand can use. 216 00:21:22,953.1336346 --> 00:21:29,253.1336346 And that allows them to go do all the stuff you were just talking about, like actually think about what's meaningful. 217 00:21:29,258.1336346 --> 00:21:30,843.1336346 And it's a lot of it centered around the customer. 218 00:21:30,873.1336346 --> 00:21:31,773.1336346 I loved you saying that. 219 00:21:32,393.1336346 --> 00:21:33,83.1336346 Yeah, for sure. 220 00:21:33,83.1336346 --> 00:21:37,238.1336346 And the specificity of those guiding docs I think is even more important than ever. 221 00:21:38,33.1336346 --> 00:21:44,513.1336346 Brand guidelines, voice and tone need even more examples now, so it's clear when we're looking right. 222 00:21:44,513.1336346 --> 00:21:45,324.1336346 For a brand team who's not. 223 00:21:46,18.1336346 --> 00:21:50,488.1336346 Necessarily product marketing, maybe not as involved in the research that went into that. 224 00:21:50,758.1336346 --> 00:21:55,948.1336346 I think they need even more examples of this is what it sounds like in a DM reply for customer success. 225 00:21:55,948.1336346 --> 00:21:58,888.1336346 This is what it sounds like in an email. 226 00:21:59,188.1336346 --> 00:22:00,868.1336346 This is what it does, this is what it doesn't. 227 00:22:00,868.1336346 --> 00:22:04,378.1336346 And the nuance of why those inputs have to be even more explicit. 228 00:22:04,378.1336346 --> 00:22:10,798.1336346 And again, the beauty of large language models is they can ingest a lot of that in ways that our human brains might tune out. 229 00:22:10,848.1336346 --> 00:22:12,48.1336346 I have a human. 230 00:22:12,633.1336346 --> 00:22:14,913.1336346 Version of what these guidelines look like. 231 00:22:15,43.1336346 --> 00:22:28,688.1336346 when we train our models they need to be really explicit and specific, and I always want AI tools to explain the thinking of why I usually follow up with, write an email to an executive explaining the inputs you use to make this decision and why you made it. 232 00:22:29,18.1336346 --> 00:22:33,548.1336346 Because then I executive can look at it and say, that's not the input. 233 00:22:33,548.1336346 --> 00:22:35,978.1336346 I want it to wait, do it again this way, et cetera. 234 00:22:36,158.1336346 --> 00:22:48,208.1336346 The same way we would train a very high potential Junior employee, they need those inputs to be really specific because the patterns they're picking up on may not be the ones that we want, and they might be obvious to us. 235 00:22:48,268.1336346 --> 00:22:55,138.1336346 So all of that is good, for more effective onboarding, but then making sure the models are even more reliable and they don't tend to degrade over time. 236 00:22:56,813.1336346 --> 00:23:01,143.1336346 I think that's a great tip for folks, the asking, doing the executive summary. 237 00:23:01,143.1336346 --> 00:23:04,843.1336346 How did you do it? Why did you think through it? I'll definitely be borrowing that. 238 00:23:04,843.1336346 --> 00:23:10,243.1336346 I think even the weighting of why did you make this decision over that is fascinating. 239 00:23:10,613.1336346 --> 00:23:13,563.1336346 Another thing you said was interesting was the recency bias. 240 00:23:13,593.1336346 --> 00:23:17,103.1336346 And I think you're spot on there because we tend to. 241 00:23:17,593.1336346 --> 00:23:30,8.1336346 Think about the last great as being the greatest thing, but being able to balance some of that deep research that you can do with ai with some of the speed that, you can also benefit from ai. 242 00:23:30,208.1336346 --> 00:23:49,623.1336346 Can you talk about, balancing the training of the AI with the deep research, with being able to test and learn? How does that look for you? It's bi-directional for sure, and involves as many touch points of alignment between the internal stakeholders who need to use it as possible. 243 00:23:49,673.1336346 --> 00:23:58,193.1336346 Weekly revenue team meetings, biweekly product marketing team meetings to talk about, go to market strategy, road mapping, inputs, et cetera. 244 00:23:58,383.1336346 --> 00:24:03,3.1336346 we on marketing, are running our experiments, consolidating the data that's important. 245 00:24:04,398.1336346 --> 00:24:08,838.1336346 Presenting in a way that A, the product and other teams wanna use, but B also can be used for our models. 246 00:24:09,138.1336346 --> 00:24:17,178.1336346 And the same thing with product, right? They are running their evaluative product, research on how features are being used, what's driving satisfaction, et cetera. 247 00:24:17,538.1336346 --> 00:24:26,628.1336346 So when that's a bi-directional, very frequent touchpoint, it just reinforces that this is iterative and customer data and insight is going to be at the forefront. 248 00:24:26,958.1336346 --> 00:24:29,628.1336346 We can upload direct transcripts to compare. 249 00:24:29,628.1336346 --> 00:24:32,268.1336346 This is how we thought customers would react to it. 250 00:24:32,448.1336346 --> 00:24:39,738.1336346 This is how Alpha users that we've intentionally chosen because they're representative of the buyer that we are wanting to represent. 251 00:24:39,928.1336346 --> 00:24:40,978.1336346 actually reacted to it. 252 00:24:40,978.1336346 --> 00:24:48,598.1336346 it just refines those insights and distills them more and more over time in a way that's really hard for us as humans to comb through hours of. 253 00:24:49,138.1336346 --> 00:24:55,838.1336346 Gong or chorus call data the biggest difference for me with this data is pre ubiquitous ai. 254 00:24:56,108.1336346 --> 00:25:07,968.1336346 I defaulted to three or four power users that were power users at the beginning, were less representative of the target market as we scaled, but we still had such a good rapport, relationship, et cetera. 255 00:25:07,968.1336346 --> 00:25:10,368.1336346 And so I think they overweighted before ai. 256 00:25:11,403.1336346 --> 00:25:20,223.1336346 Decisions I would always run through, like, how would Jay respond to this? How would Carrie respond to this? And that was really easy to guess over time, but new users would come up. 257 00:25:20,223.1336346 --> 00:25:29,883.1336346 There'd be some patterns that I wouldn't think to identify or connect with that AI tools are really good at figuring out, this is a pattern of what's driving the adoption that we actually wanna see. 258 00:25:30,93.1336346 --> 00:25:36,813.1336346 These are some of the early signals of behavior that we either wanna replicate or intervening quickly because it's important. 259 00:25:37,128.1336346 --> 00:25:40,248.1336346 For the long term viability and LTV of the customer. 260 00:25:40,498.1336346 --> 00:25:51,188.1336346 with each function being the respective owner of that data and those inputs, they'll have a high quality bar for what should be ingested and a critical lens for interpreting what's coming out of that. 261 00:25:51,488.1336346 --> 00:25:55,388.1336346 When they share that out, there's often a different PM who rotates. 262 00:25:55,388.1336346 --> 00:25:57,608.1336346 This is the research they found on their users. 263 00:25:57,923.1336346 --> 00:25:59,933.1336346 This is a customer marketing manager. 264 00:25:59,983.1336346 --> 00:26:07,633.1336346 We go in these iterative cycles, so we don't get stale with the data and we don't ever really think we figured it out because we know things are moving. 265 00:26:08,263.1336346 --> 00:26:14,143.1336346 That's the same human ai, symbiotic, bidirectional feedback loop that again, really has to be. 266 00:26:14,683.1336346 --> 00:26:28,643.1336346 Controlled for the inputs that like they have to be good or else the outputs are not actionable and just reinforces that we're continuously ingesting more and more customer data and insights as customer success teams surface interesting call recordings. 267 00:26:28,693.1336346 --> 00:26:35,143.1336346 AI tools are picking out keywords that we didn't even identify were common in different points of the same sales pitch. 268 00:26:35,143.1336346 --> 00:26:39,13.1336346 I may say, Hey, we're having the same objection happening at this slide. 269 00:26:39,388.1336346 --> 00:26:40,348.1336346 We didn't pick up on that. 270 00:26:40,718.1336346 --> 00:26:49,358.1336346 identifying things that may be worth double clicking into, figuring out what's causing those patterns and then acting on them ongoing is really powerful. 271 00:26:49,358.1336346 --> 00:27:06,723.1336346 And I also think it's powerful in a way that overcomes a common problem that a lot of teams don't realize it's the competitive signal noise, right into the competitive intel channel that's usually just some haphazard Slack channel where like anyone who sees any press release that any competitor does, just drops a link in there. 272 00:27:08,58.1336346 --> 00:27:12,318.1336346 no context of who they want to do what with it, and then the sales team goes nuts on it. 273 00:27:12,318.1336346 --> 00:27:12,798.1336346 We've seen that. 274 00:27:12,798.1336346 --> 00:27:25,68.1336346 And so I think that the big, the reason I'm bringing this up is because I think competitive noise is more commonly, frequently talked about than customer intelligence in many organizations. 275 00:27:25,368.1336346 --> 00:27:34,408.1336346 And the beauty of the tools we have at our disposal now is that it can unseat a lot of that the frequency with which we're talking about new customer quotes, new customer use cases. 276 00:27:34,618.1336346 --> 00:27:44,428.1336346 It can be a lot higher when we have tools to identify the ones that are relevant and then go back into, the strategic human decision making, which is discerning what's worth discussing and why. 277 00:27:46,198.1336346 --> 00:27:46,733.1336346 I love that. 278 00:27:46,733.1336346 --> 00:27:51,893.1336346 I think like you, I can think of the customers that I'm close. 279 00:27:51,893.1336346 --> 00:27:57,803.1336346 With that I am constantly getting feedback from, but it really is those people who are not necessarily your power users. 280 00:27:57,803.1336346 --> 00:28:07,243.1336346 How do you get them to become your power users and the signals that they're dropping? I would be curious to see if you guys are able to get more power users by, engaging those folks. 281 00:28:07,443.1336346 --> 00:28:07,923.1336346 Totally. 282 00:28:07,933.1336346 --> 00:28:32,263.1336346 being more intentional with the segmentation of them, they may look similar from the defined characteristics we already have of our ICP, but there could be even more granular insights these people tended to engage with the product in a similar way that we wouldn't necessarily have grouped now we can say, okay, this group actually wants something different out of the product, or their time to value looks different because they're going for something different. 283 00:28:32,563.1336346 --> 00:28:37,63.1336346 It's hard for humans to discern that, especially in a sales-led growth motion. 284 00:28:37,453.1336346 --> 00:28:48,818.1336346 It's a lot faster to start identifying these patterns I wanna shift gears a little bit and talk about AI application, especially right now when companies are. 285 00:28:50,23.1336346 --> 00:28:56,553.1336346 Talking about investing in ai, but honestly, everyone's also dealing with tight budgets right now. 286 00:28:56,553.1336346 --> 00:29:13,243.1336346 So for leaders who are facing tight budget constraints, what criteria determines whether an AI tool makes it into your tech stack that you could share with us? I actually think this is one of the ways that AI is less different or no different than things we've used in the past. 287 00:29:14,53.1336346 --> 00:29:27,63.1336346 Clear problem statement, clear hypothesis, clear, ROI How does any tool, hire, conference, et cetera, make it into the budget? I think AI tools are exactly the same in this way, and this may be somewhere we overcomplicate things. 288 00:29:27,63.1336346 --> 00:29:34,623.1336346 We have a clear use case, which is part of a priority because of this clear problem statement, mapped these goals, here's how we're measuring it. 289 00:29:34,653.1336346 --> 00:29:42,153.1336346 I don't see AI tools as being any different really than non-AI tools in the budgetary planning capacity. 290 00:29:42,978.1336346 --> 00:29:48,493.1336346 I'd be curious if I'm missing something or you no, I think it's really interesting because there's. 291 00:29:50,163.1336346 --> 00:30:04,773.1336346 A clear path, right? If you know what you're doing with ai, if you're getting the mandate that you need to go ag agentic, which we've heard leaders being told they need to do, it makes it a little bit harder because there's not really a play to go. 292 00:30:05,448.1336346 --> 00:30:14,908.1336346 Buy a piece of software that can even claim to do the most amazing thing and have the highest ROI at the pace that you might need to get it at for it to be valuable. 293 00:30:15,38.1336346 --> 00:30:19,448.1336346 what I found really interesting too is I'll just use ChatGPT and Gemini over the last three months. 294 00:30:19,578.1336346 --> 00:30:21,313.1336346 Enterprises have gone with Gemini. 295 00:30:21,503.1336346 --> 00:30:26,213.1336346 if you talk to someone who's used Gemini in the last two months, they've said it's gotten a lot better. 296 00:30:26,343.1336346 --> 00:30:27,543.1336346 But three months ago it sucked. 297 00:30:27,693.1336346 --> 00:30:32,13.1336346 people are starting to say chat, GPT is for whatever reason, not working for them right now. 298 00:30:32,143.1336346 --> 00:30:37,443.1336346 AI is moving so fast that putting your eggs and your investment in one basket seems risky. 299 00:30:37,443.1336346 --> 00:30:38,788.1336346 it's making it harder to know what. 300 00:30:39,543.1336346 --> 00:30:40,803.1336346 Horse to bet on. 301 00:30:40,963.1336346 --> 00:30:43,783.1336346 that's one of the challenges even we have using our podcast. 302 00:30:43,843.1336346 --> 00:30:46,863.1336346 Tech Stack sometimes we're like, oh, this tool isn't that good anymore. 303 00:30:46,873.1336346 --> 00:30:47,653.1336346 I agree with you. 304 00:30:47,653.1336346 --> 00:30:53,23.1336346 People get stuck in all of this stuff, but looking for business impact would be the obvious thing. 305 00:30:53,53.1336346 --> 00:30:56,563.1336346 Or trying to improve a specific goal rather than just saying, going AG gentech. 306 00:30:57,313.1336346 --> 00:31:03,278.1336346 A hundred percent right? It's the same, right? Honestly, this is so much lower risk than picking a marketing automation platform. 307 00:31:03,818.1336346 --> 00:31:13,268.1336346 The migration of any of them to any of them, picking any of them is like a six, seven figure decision. 308 00:31:13,268.1336346 --> 00:31:18,308.1336346 Often by the time you factor in that transition, the consultants has set it up, the integrations, et cetera. 309 00:31:18,768.1336346 --> 00:31:19,488.1336346 Yeah, I don't know. 310 00:31:19,488.1336346 --> 00:31:26,393.1336346 I come back to this being much less different and also less risky than, here are your priorities, these are your goals you wanna accomplish. 311 00:31:27,318.1336346 --> 00:31:34,398.1336346 Work backwards into budget headcount, tools that it takes to get there and then do it manually. 312 00:31:34,398.1336346 --> 00:31:42,438.1336346 First of, I don't think there's any point in jumping head fast into some set tool with some big budget and some big commitment. 313 00:31:42,768.1336346 --> 00:31:43,548.1336346 Do it manually. 314 00:31:43,548.1336346 --> 00:31:46,628.1336346 Decide what works, then put your eggs in a little more. 315 00:31:46,968.1336346 --> 00:31:52,938.1336346 at least for now, it's still a smaller decision than a lot of the big decisions in marketing. 316 00:31:53,403.1336346 --> 00:32:03,983.1336346 I would pick an AI tool to bet on for the quarter and decide a hypothesis so much more easily than oh, are we switching off of this automation platform to another one? So yeah, I just come back. 317 00:32:03,983.1336346 --> 00:32:10,223.1336346 I think this is still less different and easy to overthink at this point. 318 00:32:11,843.1336346 --> 00:32:15,143.1336346 Problem statement, hypothesis, experimentation. 319 00:32:15,653.1336346 --> 00:32:16,463.1336346 Then we ramp. 320 00:32:18,643.1336346 --> 00:32:22,228.1336346 Yeah, it's like the $20 a month subscription Yeah. 321 00:32:23,278.1336346 --> 00:32:24,838.1336346 I'll throw 200 your way. 322 00:32:24,838.1336346 --> 00:32:26,278.1336346 If you really want the deep insights. 323 00:32:26,278.1336346 --> 00:32:28,888.1336346 Asana can cost more than that per license. 324 00:32:28,888.1336346 --> 00:32:29,908.1336346 So yeah. 325 00:32:29,908.1336346 --> 00:32:32,908.1336346 I think this is one of those ones where we iterate. 326 00:32:32,938.1336346 --> 00:32:33,658.1336346 You wanna use. 327 00:32:34,63.1336346 --> 00:32:37,3.1336346 Claude free version and ChatGPT Pro. 328 00:32:37,3.1336346 --> 00:32:40,753.1336346 Fine, as long as we have clarity on what's being uploaded and why. 329 00:32:40,963.1336346 --> 00:32:44,743.1336346 I'm really just more worried from a company security standpoint than anything. 330 00:32:44,983.1336346 --> 00:32:54,883.1336346 I think for now this is easy enough to leave in the hands of the respective teams and then let them decide how to hit the company level goals. 331 00:32:55,873.1336346 --> 00:32:56,163.1336346 Yeah. 332 00:32:56,723.1336346 --> 00:33:01,583.1336346 The other thing that struck me was something that Ken said, I think that teams aren't even. 333 00:33:02,258.1336346 --> 00:33:03,818.1336346 Aware of what's possible. 334 00:33:03,818.1336346 --> 00:33:16,928.1336346 Like it's hard for people to be like, think expansive enough about is AI applicable to this or not because they are getting their feet wet with like maybe having bad ChatGPT experiences or just using it for content. 335 00:33:17,108.1336346 --> 00:33:23,888.1336346 How are you overcoming that? What's the next hurdle that you're overcoming with your team or organization? He's just weekly team meetings. 336 00:33:23,888.1336346 --> 00:33:28,398.1336346 We call it marketing I love at Garner, which is just something we liked. 337 00:33:28,528.1336346 --> 00:33:33,778.1336346 it round Robins to literally every single person on the team, CMO marketing manager, every single person. 338 00:33:33,778.1336346 --> 00:33:51,268.1336346 we have the schedule laid out, marketing you loved, with a really heavy focus on AI tools, it could be the campaign sparks an idea, this AI tool, how we used it, how we trained this prompt, how to yield this output, But again, the specificity, and I think a lot of the disappointment in anything is really the execution. 339 00:33:51,268.1336346 --> 00:33:55,108.1336346 it's really easy to say, oh, it doesn't work because X, but. 340 00:33:55,873.1336346 --> 00:34:02,118.1336346 If you were to onboard another member of your team, and say, Hey, write me a good email campaign. 341 00:34:02,658.1336346 --> 00:34:03,888.1336346 They don't know who the audience is. 342 00:34:03,948.1336346 --> 00:34:05,778.1336346 They don't know what's converted well in the past. 343 00:34:05,778.1336346 --> 00:34:06,948.1336346 They don't know what you're going for. 344 00:34:07,108.1336346 --> 00:34:19,103.1336346 and so I think a lot of the disappointment is unrealistic expectations that we would not have anything else because the potential is so high, I really come down to this with everything in marketing you set your strategy first. 345 00:34:19,658.1336346 --> 00:34:22,838.1336346 Then your tactics and you evaluate in the opposite direction. 346 00:34:23,18.1336346 --> 00:34:25,928.1336346 You evaluate the tactics and the execution and go that way. 347 00:34:25,928.1336346 --> 00:34:32,108.1336346 It'd be like saying email marketing doesn't work for us because we sent one email to a list we bought and nobody converted. 348 00:34:32,158.1336346 --> 00:34:35,968.1336346 It's tactics evaluated before we make any decision or strategy. 349 00:34:35,968.1336346 --> 00:34:47,98.1336346 But before we would try these tactics, there is this clear hypothesis we're testing for this clear end, so the experiments ladder to them because it's not an experiment to just do a bunch of stuff and say what happened? It'd be like. 350 00:34:47,818.1336346 --> 00:34:50,248.1336346 I dunno, I'm gonna experiment with anything. 351 00:34:50,308.1336346 --> 00:34:53,218.1336346 And it didn't work because we didn't put the constraints around it. 352 00:34:53,488.1336346 --> 00:34:58,678.1336346 So I just think this clarity of the execution, the precision with the examples is super important. 353 00:34:58,768.1336346 --> 00:35:06,639.1336346 And if one of the concerns we have is the team doesn't know how to use AI effectively, lots of really great ways to integrate that. 354 00:35:06,823.1336346 --> 00:35:09,638.1336346 And if the expectation is that we want them to be using ai. 355 00:35:10,343.1336346 --> 00:35:11,423.1336346 Weekly team meeting set. 356 00:35:11,423.1336346 --> 00:35:13,403.1336346 it could be your marketing team or your team meeting. 357 00:35:13,643.1336346 --> 00:35:15,953.1336346 It could be the whole company if there's something important. 358 00:35:16,253.1336346 --> 00:35:20,33.1336346 And then sharing out wins, letting people have a easy channel to surface it. 359 00:35:20,243.1336346 --> 00:35:27,323.1336346 We have a Google form that populates a spreadsheet that dings of, Hey, I wanna present because I have this cool customer story or whatever. 360 00:35:27,813.1336346 --> 00:35:34,588.1336346 So it can come from all different levels of the company because the customer facing folks may have really cool insights about how they used AI to. 361 00:35:35,268.1336346 --> 00:35:53,778.1336346 Surface some pain point for a customer, that executives will have no idea executives will have access to a lot of resources and intentional frameworks that their team needs making this really recurring, iterative, frequent, makes this psychological safety high to say, Hey, I tried this. 362 00:35:54,48.1336346 --> 00:35:54,888.1336346 This is what came out. 363 00:35:54,888.1336346 --> 00:36:00,158.1336346 Obviously this lame blog post sounds like it was written by AI and wasn't. 364 00:36:00,953.1336346 --> 00:36:02,993.1336346 Uniquely valuable or worth publishing. 365 00:36:03,53.1336346 --> 00:36:07,583.1336346 What did we do to refine it? What did we learn about that? It's just really, making that a frequent part. 366 00:36:07,733.1336346 --> 00:36:11,813.1336346 Just like we frequently share customer insights more than competitive intel. 367 00:36:12,113.1336346 --> 00:36:20,463.1336346 The frequency with which you share specific examples reinforces the expectation, gives people ideas for how to use it. 368 00:36:21,273.1336346 --> 00:36:44,623.1336346 One more thought I recently read a quote, I'm sorry, I'm going to, I don't know who to attribute it to, but some CEO was interviewed in some newsletter I read, and he said, he asked himself every month, how am I complicit in the behavior? I say, I don't want, I think that's really important because our teams are looking at what we're doing Obsession with competitors over customers, for example. 369 00:36:44,843.1336346 --> 00:36:52,463.1336346 we have this random, noisy, competitive Intel Slack channel that's constantly bombarding people with every press release of any competitor. 370 00:36:52,463.1336346 --> 00:36:55,383.1336346 Even if we don't lose deals to them, we're reinforcing behavior. 371 00:36:55,383.1336346 --> 00:37:07,703.1336346 We don't want, as leaders with AI outputs, with our team's use of it, asking ourselves, how do we foster the behavior we want, and how are we complicit in encouraging behavior? We say we don't want I like that. 372 00:37:08,178.1336346 --> 00:37:08,388.1336346 Yeah. 373 00:37:08,723.1336346 --> 00:37:29,798.1336346 One thing I wanted to ask you because I know many of our listeners struggle with this in their day to day, how do you build organizational trust in the insights delivered from AI so that they actually drive strategic decisions instead of just collecting digital dust? Yeah, same as any buy-in. 374 00:37:29,798.1336346 --> 00:37:34,118.1336346 I really think for any product marketing insight, the transparency into the process. 375 00:37:34,478.1336346 --> 00:37:37,598.1336346 And the bi-directional involvement in creating it. 376 00:37:37,658.1336346 --> 00:37:41,318.1336346 we don't even know necessarily how some of these inputs are being leveraged. 377 00:37:41,528.1336346 --> 00:37:49,628.1336346 No one who creates any AI algorithm could tell you how it's making some of these decisions, and so trust the outputs. 378 00:37:50,508.1336346 --> 00:37:52,878.1336346 Are reliable, really starts with those inputs too. 379 00:37:53,268.1336346 --> 00:38:01,278.1336346 And I truly mean getting buy-in early to what inputs are important and then validating with people you really listen to. 380 00:38:01,278.1336346 --> 00:38:04,228.1336346 Does this sound right? This is Jeff Bezos quote that. 381 00:38:04,228.1336346 --> 00:38:06,88.1336346 I love that if anecdotes and data. 382 00:38:06,493.1336346 --> 00:38:07,423.1336346 Don't match. 383 00:38:07,543.1336346 --> 00:38:16,983.1336346 The data are wrong, there's so much power in the myriad inputs that humans are digesting any day that if something doesn't sit right, we get some AI insight. 384 00:38:18,528.1336346 --> 00:38:27,838.1336346 Let's test why maybe it's yielding something, we need to test more, but if people don't trust it because it's so at odds with their experience, let's find out why. 385 00:38:27,938.1336346 --> 00:38:33,268.1336346 in that process, we can either reveal insight that would say, great, this is a new reason to trust it. 386 00:38:33,718.1336346 --> 00:38:36,688.1336346 But we also could be testing if something really isn't fitting. 387 00:38:37,393.1336346 --> 00:38:41,683.1336346 Why is that? a transparency into the process, and bringing people into it. 388 00:38:42,43.1336346 --> 00:38:43,783.1336346 It's like the exact same process. 389 00:38:43,783.1336346 --> 00:38:53,283.1336346 The thing I flash back to is being six years into product marketing and thinking that everyone wanted a perfect go-to market strategy that I created, and I'd say, Hey, demand gen, you're welcome. 390 00:38:53,283.1336346 --> 00:38:56,43.1336346 Here's the go-to-market tracker, and here's what you're responsible for. 391 00:38:56,223.1336346 --> 00:39:00,333.1336346 Here's the positioning, here's the messaging, here's the personas, here's the thought starter, here's the plan. 392 00:39:00,483.1336346 --> 00:39:02,73.1336346 All you need to do is execute it. 393 00:39:02,478.1336346 --> 00:39:04,248.1336346 Of course, right? They didn't say that. 394 00:39:04,248.1336346 --> 00:39:05,88.1336346 They're like, no way. 395 00:39:05,148.1336346 --> 00:39:07,368.1336346 I don't like that strategy. 396 00:39:07,368.1336346 --> 00:39:09,408.1336346 This is not gonna fit well with this timing, et cetera. 397 00:39:09,618.1336346 --> 00:39:18,648.1336346 So it's same exact thing, right? If we just deliver something that we think is perfectly baked and no one had any visibility into the process, they're not going to trust it. 398 00:39:18,948.1336346 --> 00:39:21,848.1336346 But if it's early on, this is the problem statement, we're. 399 00:39:22,703.1336346 --> 00:39:25,493.1336346 Looking to resolve because we've all agreed this is important. 400 00:39:25,523.1336346 --> 00:39:31,103.1336346 Here are a couple of hypotheses, two to four that we as a cross-functional group, co-created. 401 00:39:31,493.1336346 --> 00:39:34,583.1336346 And then here's the experiments we jointly agreed to run. 402 00:39:34,743.1336346 --> 00:39:35,823.1336346 here are the results. 403 00:39:36,33.1336346 --> 00:39:37,743.1336346 Here's what it means about these hypotheses. 404 00:39:37,833.1336346 --> 00:39:41,433.1336346 Here's how we're collectively going to implement them to solve this problem. 405 00:39:41,703.1336346 --> 00:39:48,963.1336346 And that's just a really smooth, fluid conversation that actually takes less time, even though it's more recurring. 406 00:39:48,963.1336346 --> 00:39:50,313.1336346 So kickoff meeting. 407 00:39:51,48.1336346 --> 00:39:57,258.1336346 Monthly, biweekly, weekly, whatever the cadence that makes sense for the process with a cross-functional representative. 408 00:39:57,618.1336346 --> 00:40:00,168.1336346 it just means you have this recurring feedback loop of testing. 409 00:40:00,168.1336346 --> 00:40:03,678.1336346 Does this actually jive or make sense? We ran this experiment today. 410 00:40:03,708.1336346 --> 00:40:07,783.1336346 If all of product is so shocked by it, they refuse to act on it. 411 00:40:08,373.1336346 --> 00:40:08,883.1336346 Great. 412 00:40:09,33.1336346 --> 00:40:26,223.1336346 We had a small iteration window to figure that out and then go test more or bring them into those insights by the time we are ready to present something to the entire company or act on it or make a recommendation for big strategic pivot, everyone's been involved and they know what's going into it and they've made it better. 413 00:40:26,223.1336346 --> 00:40:29,643.1336346 They've actually improved it with the skepticism coming up early. 414 00:40:32,368.1336346 --> 00:40:36,418.1336346 I think it's the trust piece, like you're building trust to bringing them in early on. 415 00:40:36,518.1336346 --> 00:40:56,578.1336346 in terms of moving outside of the marketing silo, how do you find the right people that you're nominating from these different groups to provide the feedback? I'm curious, what is forming these groups look like? The internal groups of our team, the cross-functional groups It usually becomes clear from the functional head of that function and what needs to be involved in it. 416 00:40:56,628.1336346 --> 00:41:05,928.1336346 typically, if we're using, a go-to market launch, or picking a product roadmap, it's usually clear based on the resources it'll take to implement it. 417 00:41:06,693.1336346 --> 00:41:11,833.1336346 Who should have a nominee usually just default to the head of that unit. 418 00:41:11,833.1336346 --> 00:41:17,623.1336346 So the CRO may nominate their VP of sales or delegate it to a director, whomever. 419 00:41:17,993.1336346 --> 00:41:22,73.1336346 I think it's culturally specific and comes down to what it takes to implement it effectively. 420 00:41:22,73.1336346 --> 00:41:39,443.1336346 So who do you need to be involved to implement execute and act on this? Then who needs to be in which cadence emerges? And sometimes there's alternating loops We called a go to market team meeting, which only included marketing team other than the head of product marketing once a month. 421 00:41:39,743.1336346 --> 00:41:43,913.1336346 But product and product marketing were involved biweekly. 422 00:41:44,403.1336346 --> 00:41:48,623.1336346 It comes up organically, and I would say experiment with this one too. 423 00:41:48,833.1336346 --> 00:41:56,423.1336346 It'll be some problem you're trying to solve and you think of this as a war room team of iterating on it routinely, and that'll shift. 424 00:41:56,423.1336346 --> 00:41:58,523.1336346 Maybe you need more account managers now that you've. 425 00:41:58,903.1336346 --> 00:42:04,283.1336346 Started testing with these alpha users and then you don't based on the agenda, based on the goal. 426 00:42:04,343.1336346 --> 00:42:27,143.1336346 It shifts slightly, but it's usually clear if people are gonna get value out of something and what their role would be Once you've defined what it takes to implement it so I'm curious how you're using machine learning to evolve ICP or buying committees in ways that you previously couldn't given your product marketing background. 427 00:42:27,818.1336346 --> 00:42:35,188.1336346 Yeah, I'd say the segmentation I'm a lot more confident in and B, is along parameters that I didn't necessarily identify. 428 00:42:35,578.1336346 --> 00:42:36,298.1336346 So I would say. 429 00:42:36,873.1336346 --> 00:42:41,103.1336346 for much my entire career, I've had a defined process. 430 00:42:41,553.1336346 --> 00:42:51,463.1336346 Come in, look at some quantitative data, draw some hypotheses, do some qualitative interviews, find some patterns that just in my brain, it was like, oh, here's a pattern of this user. 431 00:42:51,463.1336346 --> 00:43:09,53.1336346 Because of it, it's often just not that granular I'd say now the parameters are a lot more confident in them because they're actually able to tie in a lot of these disparate data sources I would look at Amplitude and try to reconcile it with the touchpoint they used to convert and try to replicate that. 432 00:43:09,413.1336346 --> 00:43:20,123.1336346 But there was just a lot of silos in my thinking I wanna replicate this buying behavior, but then I had to go into, I wanna replicate this product behavior and just a lot of really manual. 433 00:43:20,753.1336346 --> 00:43:26,363.1336346 Processing and a lot of detail was getting missed because I didn't have a great way to string that together. 434 00:43:26,463.1336346 --> 00:43:27,958.1336346 So I'd say the inputs can be. 435 00:43:29,523.1336346 --> 00:43:30,543.1336346 A lot more streamlined. 436 00:43:30,543.1336346 --> 00:43:39,118.1336346 Now we can follow these with a lot more granularity We have a great business analytics team that pulls these together and helps identify this. 437 00:43:39,178.1336346 --> 00:43:46,258.1336346 the granularity of the way we're segmenting, I'm also just a lot more confident that it's actually driving the business behavior. 438 00:43:46,308.1336346 --> 00:43:50,328.1336346 before I was looking at how this group responds to this message. 439 00:43:50,973.1336346 --> 00:43:56,893.1336346 Especially in B2B, the buying behavior and the usage behavior important for retention may not be the same person. 440 00:43:56,893.1336346 --> 00:44:04,933.1336346 It was really hard for me to keep straight if we are selling HR software, for example, to ahead of people, but we really need their. 441 00:44:05,653.1336346 --> 00:44:07,918.1336346 HR managers to be the ones implementing 'em. 442 00:44:07,933.1336346 --> 00:44:10,933.1336346 And then we had this total comp, and this is a talent person. 443 00:44:11,203.1336346 --> 00:44:13,993.1336346 It was just really hard for me to mentally keep this. 444 00:44:14,293.1336346 --> 00:44:22,513.1336346 And so I'd say there was a lot more avatar like behavior that I would say, Hey look, we have these four personas and these are real people and I can give you real quotes from them. 445 00:44:23,73.1336346 --> 00:44:25,323.1336346 It definitely got the job done. 446 00:44:25,323.1336346 --> 00:44:28,383.1336346 I'd say 70% of the insights, it was certainly actionable. 447 00:44:28,383.1336346 --> 00:44:36,503.1336346 It was certainly directional, but it's now a lot more granular these are the kinds of really granular buying trigger behaviors, this kind of. 448 00:44:37,38.1336346 --> 00:44:48,18.1336346 Cadence and then with the product adoption, especially on the retention, the usage side, because there's just so much more competition being able to identify this is actually what time to value looks like, and it's shorter now. 449 00:44:48,18.1336346 --> 00:44:51,48.1336346 It's like minutes, maybe hours. 450 00:44:51,48.1336346 --> 00:44:55,128.1336346 I used to think, yeah, we can chill for 30 days and like showtime to value then. 451 00:44:55,368.1336346 --> 00:45:01,708.1336346 But just really being able to identify this is an at risk account for reasons we couldn't identify by just looking at the data. 452 00:45:02,48.1336346 --> 00:45:04,508.1336346 AI can start flagging that now, which is awesome. 453 00:45:05,108.1336346 --> 00:45:13,8.1336346 And I'll summarize the granularity of the segmentation and how confident I am in iterating and acting on it more quickly. 454 00:45:15,763.1336346 --> 00:45:40,98.1336346 What do you see or how do you see the role of product marketing changing in the next five years? I really think that the training inputs for our teams and our tools is going to be increasingly, proactively a strategic motion that we need somebody who has a strong command of the. 455 00:45:40,593.1336346 --> 00:45:44,523.1336346 Business strategy and as it integrates with the revenue and the product size of the houses. 456 00:45:44,943.1336346 --> 00:45:58,463.1336346 So when we think about onboarding people, the granularity of our segmentation and the specificity of things like voice and tone brand guidelines, we really need that to train tools And we need that to train people effectively. 457 00:45:58,793.1336346 --> 00:46:01,763.1336346 So I see product marketing as this center. 458 00:46:03,68.1336346 --> 00:46:08,368.1336346 Center set of the right questions to ask of our tools and people to centralize these inputs. 459 00:46:08,398.1336346 --> 00:46:22,898.1336346 How are we making sure that we have a critical eye for the inputs we need and can keep distilling those in ways we can act on I see product marketing's role being increasingly cross-functional and increasingly this nexus of. 460 00:46:23,993.1336346 --> 00:46:30,113.1336346 Questions, the right questions, the strategic guidelines that are going to be informing whether we're hitting business objectives. 461 00:46:30,673.1336346 --> 00:46:35,913.1336346 ironically, I see that as harder to measure than where we are now. 462 00:46:36,73.1336346 --> 00:46:50,403.1336346 that could change with a tool I can't envision because Five years ago, I couldn't have predicted things that are possible now, the way we capture and distill insights in the company is going to be increasingly centralized through the skillset that product marketing has now. 463 00:46:50,703.1336346 --> 00:46:54,633.1336346 And that's something I see changing and evolving in the next five years. 464 00:46:55,763.1336346 --> 00:46:56,608.1336346 Very cool. 465 00:46:56,658.1336346 --> 00:47:02,868.1336346 I think that's part of the key, and you were talking a little earlier about critical thinking and that being also core. 466 00:47:02,868.1336346 --> 00:47:09,538.1336346 I think that's throughout marketing gonna be something that is gonna be such a skill that everybody has to, embrace and adopt. 467 00:47:09,838.1336346 --> 00:47:17,343.1336346 I'm curious you said, you couldn't imagine where we're at today, what you think an emerging AI capability will be really foundational to PMM. 468 00:47:18,248.1336346 --> 00:47:23,878.1336346 Within the next year or two years what would be the dream AI capability. 469 00:47:25,253.1336346 --> 00:47:27,473.1336346 It's a great question things are changing quickly. 470 00:47:27,833.1336346 --> 00:47:32,3.1336346 I think right now there's still this indefinable. 471 00:47:33,713.1336346 --> 00:47:43,463.1336346 Discernment that humans who have emotions can make when looking at AI outputs that I haven't seen models be very good at replicating just yet. 472 00:47:43,523.1336346 --> 00:47:45,413.1336346 Things that just don't hit right. 473 00:47:45,413.1336346 --> 00:47:47,663.1336346 They sound AI generated to us. 474 00:47:47,663.1336346 --> 00:47:56,333.1336346 Even if AI detection tools can't capture them, there's this factor that humans with real experiences and emotions can say that's just not compelling. 475 00:47:56,663.1336346 --> 00:47:58,493.1336346 It sounds overstated, it sounds fluffy. 476 00:47:58,493.1336346 --> 00:48:00,888.1336346 There's this human discernment when we see. 477 00:48:01,718.1336346 --> 00:48:06,488.1336346 Any campaign human generated or not that I haven't seen AI filter through. 478 00:48:06,668.1336346 --> 00:48:14,978.1336346 It can spit out like 90 variations and you can say, make it sound more engaging, But it takes a human to say, no, that's dumb, that's stupid. 479 00:48:14,978.1336346 --> 00:48:16,208.1336346 That sounds patronizing. 480 00:48:16,568.1336346 --> 00:48:22,528.1336346 I would love to see this refinement of what I think of as emotional human experience If and when that happens. 481 00:48:22,558.1336346 --> 00:48:24,88.1336346 when we were closer to a GI, yeah. 482 00:48:24,118.1336346 --> 00:48:30,838.1336346 But there's just right now this discernment of, I would never put that in front of a human being. 483 00:48:30,838.1336346 --> 00:48:32,68.1336346 I would never launch that. 484 00:48:32,318.1336346 --> 00:48:39,818.1336346 AI doesn't seem to have that decision making but it's really good at taking all these disparate inputs with different formats and distilling some patterns. 485 00:48:40,58.1336346 --> 00:48:47,778.1336346 It's good at text generation, but it is that combination right now that still really takes a no way is that going out. 486 00:48:47,868.1336346 --> 00:48:49,518.1336346 I wonder if we can get past that. 487 00:48:49,618.1336346 --> 00:48:57,688.1336346 can it actually predict how a human would experience reading something? I don't know. 488 00:48:57,688.1336346 --> 00:48:59,598.1336346 And maybe, honestly, when that happens, we're just. 489 00:49:00,543.1336346 --> 00:49:03,573.1336346 Closer to some singularity we don't want. 490 00:49:03,583.1336346 --> 00:49:05,638.1336346 I was gonna say the singularity might be coming. 491 00:49:06,3.1336346 --> 00:49:06,903.1336346 Right? Topic. 492 00:49:07,18.1336346 --> 00:49:24,613.1336346 Hey, can you tell us one AI experiment you shut down and how you knew it wasn't the right path? I'd say I do a lot of deferring if something is really quick to run and we're confident that we wouldn't act on it has to be important enough to do it. 493 00:49:24,663.1336346 --> 00:49:30,413.1336346 So I tend to be intentional about is this the right use of resources, is a distraction to get value out of it. 494 00:49:30,803.1336346 --> 00:49:34,913.1336346 And so I really think it does come back to this just isn't a priority. 495 00:49:34,913.1336346 --> 00:49:36,893.1336346 We don't have enough of a pain point in. 496 00:49:37,643.1336346 --> 00:49:41,333.1336346 Optimizing this part of the funnel that has a relatively solid conversion rate. 497 00:49:41,693.1336346 --> 00:49:44,888.1336346 So I do a lot more deferring of, maybe later. 498 00:49:45,8.1336346 --> 00:49:46,118.1336346 It is just not a priority. 499 00:49:46,118.1336346 --> 00:49:49,538.1336346 Let's do these more promising, higher output. 500 00:49:49,973.1336346 --> 00:49:53,123.1336346 Things now and again, it just comes back to there. 501 00:49:53,123.1336346 --> 00:49:54,893.1336346 There's lots of problems that we could solve. 502 00:49:54,893.1336346 --> 00:49:56,33.1336346 There's lots of hypotheses. 503 00:49:56,33.1336346 --> 00:49:57,203.1336346 This is what's going into them. 504 00:49:57,413.1336346 --> 00:50:02,503.1336346 Then we just need to stack rank the experiments based on what the most important one is and revisit that quarterly. 505 00:50:02,503.1336346 --> 00:50:07,783.1336346 things that are not a priority now could be a priority in three quarters or something we can't even envision could be. 506 00:50:08,83.1336346 --> 00:50:15,93.1336346 So it's just a lot more, documenting the backlog of experiments that could be run for hypotheses. 507 00:50:15,453.1336346 --> 00:50:18,123.1336346 Of problem statements that we aren't yet solving. 508 00:50:18,543.1336346 --> 00:50:22,663.1336346 And if something's really quick a marketing manager is curious about running something. 509 00:50:22,663.1336346 --> 00:50:30,773.1336346 They can do it in a couple of hours the buy-in and energy of it is as important as the output the experimentation for themselves. 510 00:50:31,133.1336346 --> 00:50:36,173.1336346 So I think that a big part of my job is same way I saw copywriting nine years ago, is. 511 00:50:36,908.1336346 --> 00:50:37,388.1336346 Sure. 512 00:50:37,418.1336346 --> 00:50:39,758.1336346 I may think a subject line would be better with. 513 00:50:40,718.1336346 --> 00:50:50,468.1336346 Slightly different wording, but it's much more important that my email marketing manager feels the autonomy in the agency to execute quickly with quick creative iteration cycles. 514 00:50:50,828.1336346 --> 00:50:55,208.1336346 So I don't tend to intervene in a lot of the day-to-day experimentation. 515 00:50:55,208.1336346 --> 00:50:57,668.1336346 I wanna see the outputs and why we're prioritizing them. 516 00:50:57,998.1336346 --> 00:51:03,758.1336346 But if somebody wants to run something on their own time, it'll be quick and they stay energized and excited by how to use it. 517 00:51:04,28.1336346 --> 00:51:08,398.1336346 They're getting all of this implicit knowledge I tend to be less on the shut it down side. 518 00:51:08,498.1336346 --> 00:51:20,468.1336346 The iterative impact to our open rates or conversion rates is a lot smaller than the impact of an email marketing manager, who doesn't have the confidence that they can run things. 519 00:51:20,768.1336346 --> 00:51:22,293.1336346 So I tend to think that people. 520 00:51:22,738.1336346 --> 00:51:30,518.1336346 Feeling like they have the agency to do things and test and report back is much more important than, some hierarchical lens. 521 00:51:30,518.1336346 --> 00:52:02,238.1336346 this is where I actually think a lot of smaller companies have a great advantage and people who work at smaller companies right now is there's a lot of fear about proprietary data and risk centered conversations about using AI at bigger companies and a lot of shut it down mentality I get, but is a total bummer to somebody trying to develop in their career and figure out how to use these tools if they've not been allowed to until some higher up that they've never met on maybe a continent they've never been to, decides they can use it. 522 00:52:02,388.1336346 --> 00:52:10,448.1336346 And so I think we're very much in the, agency empowerment, moving quickly, being intentional and mindful of the data, we're uploading, the risks, et cetera. 523 00:52:10,718.1336346 --> 00:52:21,98.1336346 So there's constraints on that, but ultimately, if the risk is low and the time is low, yeah, let's just let people be confident using ai. 524 00:52:21,823.1336346 --> 00:52:22,113.1336346 Yeah. 525 00:52:22,748.1336346 --> 00:52:23,378.1336346 I like it. 526 00:52:24,158.1336346 --> 00:52:34,208.1336346 could give one piece of advice to a go to market or product marketing leader, Navigating ai, what would it be? Because I can only have one. 527 00:52:34,238.1336346 --> 00:52:37,268.1336346 I'll give you the one and then I have a backup of a kind of two-parter. 528 00:52:37,728.1336346 --> 00:52:40,548.1336346 more than ever, know your miss is really important. 529 00:52:40,788.1336346 --> 00:52:41,538.1336346 We all have. 530 00:52:42,808.1336346 --> 00:52:50,998.1336346 weaknesses, predispositions and mine is over preparing, over researching, over data ing, which the miss is going slower. 531 00:52:51,28.1336346 --> 00:52:59,213.1336346 And so if you know your miss is I tend to over prepare, which tends to make me hesitant to launch things too quickly or share them too broadly. 532 00:52:59,683.1336346 --> 00:53:06,648.1336346 Then when things feel rushed or haphazard, or they haven't been vetted thoroughly enough, it's probably on. 533 00:53:07,353.1336346 --> 00:53:09,3.1336346 Most people's, okay, let's do it. 534 00:53:09,3.1336346 --> 00:53:09,693.1336346 It's been thorough. 535 00:53:10,73.1336346 --> 00:53:15,593.1336346 knowing your miss is really important because everything you're looking at is filtered through That lens. 536 00:53:16,223.1336346 --> 00:53:19,883.1336346 If you don't know your miss, ask five people finish this sentence. 537 00:53:19,883.1336346 --> 00:53:21,983.1336346 If I didn't know you better, I would think blank. 538 00:53:22,503.1336346 --> 00:53:24,233.1336346 I'll reveal a pattern quickly. 539 00:53:24,403.1336346 --> 00:53:26,713.1336346 So that, that really know your miss is the one piece of advice. 540 00:53:26,743.1336346 --> 00:53:33,344.1336346 But secondarily, I think a lot of the greatest insight is going to come much more democratically that it. 541 00:53:34,3.1336346 --> 00:53:40,383.1336346 None of us have experience with these tools, so letting your team feel empowered to surface new ways. 542 00:53:40,473.1336346 --> 00:53:45,753.1336346 I'm learning things all the time from my teams about new tools I've never heard of that seem really promising. 543 00:53:46,23.1336346 --> 00:53:51,393.1336346 Sure, if you wanna use Jasper and you wanna use hoppy copy and I wanna use clot, whatever, just let it go. 544 00:53:51,753.1336346 --> 00:54:01,533.1336346 I think that a lot of really great ideas, we can't predict where they'll come from, and now more than ever, democratizing a way to surface them is so important for the team's growth in learning. 545 00:54:04,223.1336346 --> 00:54:05,393.1336346 I completely agree. 546 00:54:06,143.1336346 --> 00:54:13,343.1336346 So as we wrap up here at Futurecraft, we're all about getting the great minds we talk to, to share some quick bits of information. 547 00:54:13,553.1336346 --> 00:54:15,143.1336346 So you're ready for a bit of a lightning round. 548 00:54:15,588.1336346 --> 00:54:16,93.1336346 Let's do it. 549 00:54:16,833.1336346 --> 00:54:42,58.1336346 What's your favorite AI tool right now? It's just ChatGPT what's the most over-hyped AI buzzword? We all want this all in one blank, right? How are you keeping up with ai? Really learning from my team making sure that it's part of the iteration process of testing and validating things and servicing new ways we're using it. 550 00:54:42,58.1336346 --> 00:54:44,878.1336346 Just making that a recurring learning motion. 551 00:54:46,123.1336346 --> 00:54:46,453.1336346 Thanks. 552 00:54:47,23.1336346 --> 00:54:52,373.1336346 And what book, podcast newsletter has shaped your thinking on AI or go to. 553 00:54:53,273.1336346 --> 00:54:55,823.1336346 Human-Centered Marketing by Ashley Faus. 554 00:54:55,823.1336346 --> 00:54:57,173.1336346 If you haven't read it, you should. 555 00:54:57,173.1336346 --> 00:54:58,43.1336346 It just came out. 556 00:54:58,383.1336346 --> 00:55:00,93.1336346 Head of Lifecycle Marketing at Atlassian. 557 00:55:00,173.1336346 --> 00:55:05,783.1336346 so comprehensive, so thorough, the real examples and real templates and frameworks she uses. 558 00:55:06,23.1336346 --> 00:55:06,923.1336346 Fantastic. 559 00:55:06,983.1336346 --> 00:55:07,703.1336346 A plus. 560 00:55:07,973.1336346 --> 00:55:09,113.1336346 Can't recommend it enough. 561 00:55:10,23.1336346 --> 00:55:10,313.1336346 Okay. 562 00:55:10,383.1336346 --> 00:55:12,243.1336346 She's coming on the podcast later. 563 00:55:12,578.1336346 --> 00:55:13,418.1336346 Tell her I say hi. 564 00:55:15,93.1336346 --> 00:55:15,633.1336346 That'll be great. 565 00:55:17,508.1336346 --> 00:55:21,198.1336346 So Rebecca, thank you so much for joining us as a fellow product marketer. 566 00:55:21,198.1336346 --> 00:55:30,288.1336346 I loved this conversation and I love how you're thinking about the future of go to market and product marketing in this AI wild world we're in. 567 00:55:30,498.1336346 --> 00:55:34,278.1336346 And I think our listeners are gonna take a lot of things away from this conversation. 568 00:55:34,278.1336346 --> 00:55:36,363.1336346 thank you so much Thanks for having me. 569 00:55:36,413.1336346 --> 00:55:36,878.1336346 Thanks so much. 570 00:55:37,673.1336346 --> 00:55:47,515.1702502 Thank you, Uh, uh, uh, uh, uh, I feel like I learned a lot. 571 00:55:47,515.1702502 --> 00:55:49,735.1702502 It was a great conversation. 572 00:55:49,765.1702502 --> 00:55:55,275.1702502 I stimulated so many thoughts about where the future of go-to market's going. 573 00:55:55,335.1702502 --> 00:56:00,565.1702502 But I think the thing that really stuck out to me is the evolution of the role of product marketing. 574 00:56:01,15.1702502 --> 00:56:05,635.1702502 In some organizations, product marketing is the strategic lever. 575 00:56:05,665.1702502 --> 00:56:07,855.1702502 It really is the heartbeat of go to market. 576 00:56:08,410.1702502 --> 00:56:10,240.1702502 But in many organizations it's not. 577 00:56:10,270.1702502 --> 00:56:17,650.1702502 It's creates the pitch decks, it creates the position of messaging, which is important, but AI is going to free product marketers to think. 578 00:56:18,260.1702502 --> 00:56:20,150.1702502 Act much more strategically. 579 00:56:20,240.1702502 --> 00:56:43,490.1702502 And I think we're gonna see an evolution of product marketers to be more like brand managers in consumer packaged goods, who really are responsible for all the touchpoint that a customer engages with a with a brand, with a company, especially in B2B software right now, and responsible for delivering on the p and l supporting sales in a different way than they have before. 580 00:56:43,490.1702502 --> 00:56:45,230.1702502 And I think it's gonna be super exciting. 581 00:56:45,500.1702502 --> 00:56:53,490.1702502 Rebecca seems to know where it's going and I'm curious to see how she's driving that evolution But what about you? What did you learn? one of the things. 582 00:56:53,840.1702502 --> 00:57:06,680.1702502 That really spoke to me and brought it up in the interview is I tend to rely on the power users, the folks that always have interesting things to say and talking to those customers who I have a really good relationship with. 583 00:57:06,980.1702502 --> 00:57:10,790.1702502 And I think that does make you a little more biased to, or maybe overweight. 584 00:57:11,515.1702502 --> 00:57:14,95.1702502 What you see their perspective being. 585 00:57:14,515.1702502 --> 00:57:27,245.1702502 And I've really loved what Rebecca had to say about being able to take more of the insights of the middle layer, maybe the, that aren't power users yet, and understand what their challenges and problems are and what they really love about the solution. 586 00:57:27,245.1702502 --> 00:57:28,205.1702502 And leaning into that. 587 00:57:28,215.1702502 --> 00:57:33,55.1702502 being able to test with a larger group of customers can be really powerful. 588 00:57:33,55.1702502 --> 00:57:37,525.1702502 So I really took that insight away and really enjoyed the conversation as well. 589 00:57:38,325.1702502 --> 00:57:40,725.1702502 Yeah, another great day on Future Craft. 590 00:57:41,5.1702502 --> 00:57:42,835.1702502 That's it for us today. 591 00:57:43,75.1702502 --> 00:57:44,755.1702502 Thank you everyone for listening. 592 00:57:44,875.1702502 --> 00:57:53,55.1702502 And if you have a minute and want to get some good karma points, give us a follow rate us on Spotify or Apple please. 593 00:57:53,245.1702502 --> 00:57:59,35.1702502 That helps us get the word out or share it with a friend and reach out to us if you have ideas or comments on things we can do. 594 00:57:59,540.1702502 --> 00:58:00,380.1702502 That you wanna see. 595 00:58:00,380.1702502 --> 00:58:04,640.1702502 We've got much more to show you for season two, but thanks everyone for listening. 596 00:58:05,465.1702502 --> 00:58:08,885.1702502 Yeah, let's keep crafting the future of go to market together. 597 00:58:09,185.1702502 --> 00:58:09,815.1702502 Thanks. 598 00:58:10,225.1702502 --> 00:58:10,715.1702502 Bye-bye.

All Episodes

Episode Transcript

Popular Podcasts

NFL Daily with Gregg Rosenthal

On Purpose with Jay Shetty

Dateline NBC

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

.css-14f5ked{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;}Acceptable Mistakes & Ruthless Prioritization: How Top PMMs Are Winning in AI GTM

Episode Transcript

Popular Podcasts

.css-r6mb8g{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:1;overflow:hidden;}NFL Daily with Gregg Rosenthal

On Purpose with Jay Shetty

Dateline NBC

All Episodes

Acceptable Mistakes & Ruthless Prioritization: How Top PMMs Are Winning in AI GTM

NFL Daily with Gregg Rosenthal