AI in B2B marketing: insights from experts at AWS, Amazon, Cast AI and ModMed

Speaker A: Today we are going to talk about very, very interesting topic, how AI is changing B2B marketing.

Speaker B: We also had an AI idea thon so anybody in the marketing department could

Speaker C: submit an idea as a product marketing M manager. Uh, my main goal is to change perception about my product.

Speaker D: We can give our users one to one unique experience. It's essentially a website created for them

Speaker E: because of AI or because of quick seed. What has happened is I don't want to go out and look at some other benchmark.

Speaker A: Hi everybody. Welcome to the B2B Marketing Leaders Podcast. And today we are going to talk about very, very interesting topic on um, how AI is changing B2B marketing. And today with me there are, um, some bright and amazing experts from tech companies. So please introduce yourself. Let's start with in a circle with Sabari. Yeah, please start. Hey.

Speaker C: Hi Olga. Um, I'm Sabari. Uh, I work as a product marketing manager for Kubernetes at AWS. Um, been in the B2B tech space for eight years now. Um, started as a performance marketing manager and eventually found my way into product marketing. So that's me.

Speaker A: That's amazing. Thank you. Maddie.

Speaker C: Hi.

Speaker B: Um, my name is Maddie and I'm currently a digital marketing specialist at ModMed. I've been in marketing for about five years. I previously started sort of on the social media side of things and then I worked at an advertising agency and now I work um, for a company called Modmed and it's a B2B company that sells EHR software to doctors and we work with private practices across 11 different specialties.

Speaker A: Amazing. Thank you.

Speaker B: Yeah, Alex everyone.

Speaker D: Great to be here. So my name is Alexei or Alex. I'm director of demand generation company called Cast AI And I'm smiling because we are optimizing Kubernetes. So saying hi to Sabari and AWS. I've been in marketing for about like 15 years I think. X similar web access, semrush. Um, and currently leading the management team here at Cust.

Speaker A: Yeah, great, happy to meet you. Yeah, thank you. And Parth.

Speaker E: Hey everyone. My name is Parth, uh, and I've been into the P2P uh space for like about five years. I started off from a marketing agency so I was working with multiple different clients for P2P. Uh, then I transitioned into in house. I was working for JFrog which is an artifactory uh, solutions company. A uh, B2B SaaS company and I was working there as a growth marketing manager for like about three years. And then later on now I transitioned into Amazon. So I work for Amazon Global Communications, uh, over here, and I work as a marketing manager, so have been in this space for the last five years or so.

Speaker A: Amazing. So, such bright and interesting people here. So thank you for coming. And my first question to you would be, are you techno optimists or technopessimists? What do you think?

Speaker D: Can I be techno realistic? Like in between?

Speaker A: Yeah, you can.

Speaker D: Okay, so I would vote for this one. Like something in the middle. But I think in general we're living now through the most meaningful technology shift of our life. I think in general is like electricity or, I don't know, whatever you think of it. So yeah, it's something that would change our life in general. And if you do not adapt to this, then the world will make you at a. Like, you will have to eventually.

Speaker A: M. You're optimistic. A little bit, I believe.

Speaker C: I kind of like that answer being like, realistic because, um, I 100% agree, uh, AI is going to change a lot. It's uh, um, helping all of us, uh, be power users of a lot of different things that we probably never ever imagined using. Whether it's coding tools or something, um, like, you know, as marketers we've never used. But, um, at the same point, um, the reason I say realist is because we need to ensure like, you know, we take care of certain guardrails. There's like, still, um, issues with, um, how AI hallucinates with a lot of different things. Um, so, you know, with looking at that, like, I still don't see a lot of like, you know, empathetic content being put out by, by AI. Um, you know, not to forget the human psychology that comes, uh, with, you know, what we do in our space. Right? Like marketing. M. So with that being said, I 100% agree. I'm optimistic, but more of a techno realistic, like galaxy, so. 100%.

Speaker A: Yeah, me too. I believe. Yeah.

Speaker E: I mean, I, I would, I would also kind of align with the same, uh, idea of being more realistic because like, you know, I would say that optimism is where, when we use AI as a copilot rather than completely on it, because as of today there are a lot of uncontrolled deployments when it comes to privacy, accuracy and also related to a lot of brand risk. So, you know, Amazon has a saying, saying that, you know, customer obsession is like one of the major, uh, key principles, uh, for anybody at Amazon. And when we put customer at front, front, AI cannot understand the customer really well at this point in time. So I think so being more Realistic in this particular scenario is kind of very important.

Speaker A: Mhm. So don't trust 100%. Yeah, yeah, that's true. And my next question to you is about B2B marketing. How AI is currently used in your company's marketing processes.

Speaker C: There are different part. Uh, although he's at Amazon, I'm at aws. We basically I think fall into the same uh, marketing organizations. Um, our domains are different. Um like our day to day operations slightly differ. Um and mostly at ah Amazon we use a lot of internal tools. We don't have to rely on anything external. Um, but I see day to day I work with a lot of peer teams, like 16 to 17 different peer teams. Uh, some teams are working on uh, analytics. Uh, AI has been used in analyzing data. I personally use it in uh, a lot of my day to day work like segmenting um, data, uh, writing reports. Um, it's like analyzing ah, customer related feedback. It's just making things easier. But most of these tools are like internal uh, AWS tools that we're using. So if I am okay to promote it's, it's like Amazon's Quick See Quick suite. That's literally what I use at this point.

Speaker E: I mean I, I would say that I have a very similar answer to it because I mean um, I, I would say this transition has happened from 2025. Is Amazon only promoting internal tools? They are strictly against external tools and they're actually trying to build internal tools in such a fashion that uh, anybody working at Amazon has every relevant information about the data inside of Amazon, which kind of makes it a more secured space. Even if you are trying to experiment, try to do ideation, trying to do hypothesis or analysis, it kind of becomes more uh, easy for you because you can pick any data and kind of feed it into the AI and kind of get insights or answers to any of your questions rather than me picking the data and putting into some other AI tool like ChatGPT or Perplexity. Um, it kind of is a threat for us, you know, in, in terms of internal users because we are feeding our data to some other company which we don't want to do. And I think so that kind of becomes very uh, easy and important for you know, us as uh, people working at Amazon.

Speaker A: Yeah, and um, risks again. Yeah, and risks and uh, safety. Yeah, thank you for mentioning that. Um, so, so uh, Alex or Maddie, would you like to share as well how AI is used?

Speaker D: Yeah, sure. So I agree with the, with the security component of it. And we also have, we have, I think Claude Enterprise and it's kind of locked so you cannot go uh, share information outside of the company. But in general I was thinking of how we are using AI now because there are so many angles and I would group this into like maybe three big buckets. One, everything related to research and insights like call analysis, feedback, like anything, like all of the research and psych work. Second, content creation was called drag. So now we can hyper personalize any outbound, uh, emails or whatnot, knowing all of these contexts on the fly literally which saves like tons of hours and days will work. Um, and third, I would just name it like chief of Staff like everything else. So all of the internal processes we get in a submit a task in Asana, then it goes there, curate, AI takes it and converts to something. So all of this stuff uh, now just becomes way faster. So these three main groups.

Speaker A: Yeah, I think you covered most of the main scenarios, AI related scenarios. Thank you. And Mehdi, would you like to add something?

Speaker B: Yeah, sure. Um, I mean I have a slightly different answer. So currently at ModMed Marketing, I mean we primarily are still kind of in the phase of relying on Channel Chat, GPT and Gemini, you know, for things like productivity and research. But we have yet to kind of find those tools that we found, you know, useful for executing tasks. Um, however, we're onboarding a copywriting tool to help us scale copywriting for uh, marketing. So that is one area where we started to invest. But it's been a slow rollout. I think the governance around AI and the policy internally has sort of taken a very uh, measured approach in how we allow employees to use AI at our company. And I think a lot of it has to do with the privacy and security concerns.

Speaker A: Yeah, that's true. It's not so easy and I believe that you've mentioned several tools you are using in the company, but maybe there are some more tools uh, you would like to mention and maybe recommend or you are not going to recommend them to others. So uh, in your AI stack and your um, marketing team. So would you like to um, mention other tools?

Speaker B: Yeah. So for example right now we're onboarding a tool called Jasper AI. I don't know if any of you have used it or if you have any reviews on that, but um, we're basically using that to scale copywriting and content creation in a way that is like aligned with our brand values. Um, also Gemini, we have that integrated into our Google workspace, so Gmail Drive, Google Meet, which definitely helps improve efficiency and helping us summarize emails and notes and conversations. Um, we also use Drift, which is an AI powered chatbot on our website and that allows people that come there to book a demo. And then we also have like agency partnerships that have unlocked AI power tools for us. Um, so for example we work with an agency that has ChatGPT with a SpyFu integration which helps us do keyword research and also pulls ah, real data around volume difficulty, keyword intent. So it's definitely a combination of AI insight and then the actual metrics that matter to us and leveraging those partnerships that we have to help us do.

Speaker A: That sounds great. And agencies, they are offering their AI agents to you? Yeah. Mhm.

Speaker B: Yeah.

Speaker A: Yeah. Great. So this is a question to everybody. Would uh, you like to mention some other uh, tools in your marketing stack? I mean AI tools he used.

Speaker D: I would say honestly that AI is not integrated in every tool. So everything that you're using before, I don't know, HubSpot, Salesforce, you name it, now has anything AI. So even Zach, like I think yesterday I was building Zap, like in Zapier something very quickly again internally in which you have stuff and it was like hey Zapier, could you please do this and that and what yesterday took me for that three hours now, uh, literally five minutes. So these um, things. But on top of what Matt said, um, like Zapier and Ipan, we have actually a pretty cool thing called Webless. Webless, it's in startup from San Francisco I think they based. Um, we have it integrated on the website and it has all of the docs, all of the blogs, like all of the content we gave it. Uh, and it works as a sort of a chatbot, but it's not a chatbot in the, how to say in the traditional way. Like you're just uh, asking a question and you know that this is AI, but, and you know that this AI will answer based off uh, all of the content that company shared, our company shared with this WhatsApp. You can see we can give our uh, users one to one unique experience. It's essentially a website created for them. They're like, hey, I ah, using aws. Can we use cast AI with aws?

Speaker C: Boom.

Speaker D: Um, and they don't need to call us, they don't need to interact with anyone else. They just can have this right in front of them with a cta. Yes. Like we use aws. Book a call, like start free right now. That's a great thing. Really love it.

Speaker A: Yeah, great level of personalization. Yeah. Sabari, you wanted to mention something Yeah,

Speaker C: I mean, I mean it's great. I'm amazed at the number of tools, but honestly, I like Path will agree, like a lot of these things, M for us is internally enabled just by using like Amazon's Quick suite. But for me, I think the main part is uh, not these tools, but you know, as B2B marketers, like how are we using these tools? Right? Like for an example, um, um, one project that I worked on was um, literally um, uh, writing 33 scripts, um, in like a very short time frame. And these scripts were like LinkedIn long form, LinkedIn short form, YouTube, YouTube scripts. Um, it was for uh, Twitter, for Instagram. Right? Like very short, like very short time frame. Aggressive timelines for delivery. I was working with external, uh, content creators and you know, there are like number of parameters. There's SEO, there is ah, platform SEO, there's keyword SEO, then there is like uh, your messaging. So how do you integrate all of these things into um, all these 33 assets? It was very simple for me, um, internally the tool quicksuite, like I said, uh, allows us to build these custom agents. Just think of it as your uh, small language model. And I felt like, hey, this is my SEO checklist. This is my uh, platform algorithm checklist. This is my um, messaging checklist.

Speaker B: Right?

Speaker C: Like all these assets ensure, uh, they're wetted against everything. Give me a copy and then eventually I could add the finishing, you know, the human touches to it. But it accelerated my work by chance. Um, so I think like these use cases are very powerful for me as a marketer. That would be my addition. Uh, and I'm sorry, like I can't name tools because uh, we literally rely on like this one giant internal tool for everything.

Speaker D: Yeah, literally the same. We have exactly the same thing we have. It's kind of a writing assistant, a custom model that we've created and we gave it all of the SEO checklists, positioning, messaging. I see like everything. And then like, hey, I need to craft this. And based off all of this context, the tool gives you the content that you can just work with and then slightly, I mean adjust and publish wherever needed. Agree.

Speaker A: Yeah. Thank you. And part one to three.

Speaker E: Yeah, I would say that, you know, like we inside of Amazon, we only use uh, Quick Suit as one of our AI tool. But I work with external agencies and external vendors who definitely use a lot more AI tools that than we do, uh, for ad copywriting, for SEO, content writing, for YouTube, script writing, so on, M and so forth. So one good tool there is an Agency that we work with in eu, they use a tool called AS Translate AI. So that tool is very good when it comes to uh, doing live translation of your script into multiple different languages. So I work for uh, doing uh, uh so we have uh, fulfillment centers across 53 global locations wherein we are actually uh, helping Amazon uplift its brand reputation. Uh, so over there we want to translate them into like five, six, seven different languages. And initially it used to kind of take us a lot of time to do it because uh, for Japanese we need somebody who is expert in you know, from, from somebody in Japan or from somebody in you know, France for French, uh, something like that. But now because of this tool, uh, it literally like uh, took our time down to like, you know, maybe like five, ten minutes. And then I have like four or five different uh, scripts ready with different language translations done. But of course that is something that we don't use. We have an external agency who does it for us. But it's like a very good tool that they use and you know they actually help us out in our day to day.

Speaker A: Yeah, sounds amazing. And saves a lot of time and budget I believe. Yeah. And do you use AI for for example um, video content or visual content creation?

Speaker E: Uh, right now? I mean I actually did an experiment a couple of months back, uh, ah, working with um, uh, Shutterstock to actually create any content wherein we can do video, um, AI integration. But Amazon has very strict policies on not having anything on video, uh, which is AI. So they want real humans to actually be on the video so that you know, we are not kind of uh, creating an image in front of our customers that you know that it is not realistic. So that is the reason AI is like ah, Amazon has very strict policies around not using AI on video advertising. So that is the reason we, we just dropped that idea of you know, doing that.

Speaker A: Mhm.

Speaker C: Yeah, I 100% agree. And also I am personally a bit wary of like so seeing so much AI generated stuff that I've seen stop believing what's true and what's not. But I think um, where I, you know, like again for a lot of the video scripts that I work on, where AI literally helps us is um, you know like faster iterations of scripts or like you know, asking um, questions multiple times that hey, I don't like this angle. Give me another angle. Right. Like these are areas where I think um, in terms of video script generation it helps us. But uh, again for images, um, we're not using much. I still like the real. I would still like to see a real human behind whatever has been spoken. Um, again, that's my take.

Speaker D: Also, video editing is a huge thing. Every month, I would say we do a podcast and only kind of a webinar, you can call it. And normally it took me a couple hours to just remove all of the filler words and all this stuff. Now you just click a button and descript using this for you. Relatively cheap. Literally two minutes and it's done.

Speaker A: Can you recommend a tool?

Speaker D: Yeah, descript. We're using descript. It's a great thing. Yeah, super, super cool. We love it. Plus all of the interviews, it's super easy to transcript them all. And going back to the SEO part, you have a webinar, you quickly remove, uh, all of this stuff, make it shorter. Um, then you have the transcription. You upload this on YouTube, publish it on your page as a video, plus the transcript below. So it works for LLM to crawl it eventually. Then, hey, writing assistant, could you please create the recap? So, yeah, absolutely.

Speaker C: Love it.

Speaker A: Yeah, sounds great. Maybe we'll use it for this episode. We'll see. Yeah. And Maddie, would you like to add something on, uh, visual generation? Visual content generation?

Speaker B: Yeah, I think a couple things came up for me too. Like one, the authenticity piece of it. I definitely think as well, we've hesitated to kind of go in that direction because we want to, you know, just be, be real and like, be human. Like, at the end of the day, it is about patients and doctors and the relationship and the trust and the care between those two things. So unless it's like an animated video of some type, I don't imagine we would ever want to use AI to produce content for video specifically. And then the other aspect of it is I've heard a lot about these tools too, that, um, you know, can cut out the filler words and, you know, edit things and chop them up for, you know, YouTube shorts or TikToks or, you know, short form content in general. So I think that's super exciting and that's something that I'm definitely pushing internally to, you know, kind of get in the game of short form content.

Speaker A: Yeah, yeah, it's important now. I think everyone watches this kind of format. Yeah. Great. And, uh, let's talk about, um, a little bit of training of your marketing teams, um, um, to work with AI. Uh, do you have these kind of processes in your marketing team that, uh, someone trains the teams and, uh, or teaches them how to use AI? Um, so please tell more about that.

Speaker E: I mean, I would say that um in Amazon there are two different ways on how you can get trained on AI. One is um uh in each team uh what we have is we have somebody called as um the influencer of AI who helps within the team, uh who has better knowledge to kind of do peer to peer review and kind of give more knowledge out to them. And what we do as a team is uh every week before we start off the team meeting we make sure that we share AI tips and tricks with the team so that you know on a daily basis uh everybody's working in silos but everybody kind of comes together and kind of talks more about how they used AI in their day to day work in the past week or past month however it is. And uh a very good benefit of uh uh working at Amazon is like they have uh something called as AIML University wherein you can actually book sessions and you can go and talk to industry experts who might not be there in your own team but they are working in gen AI teams or they are working in some other teams who are like hands on, on AI LLM M uh working models and they can actually help you understand how you can better your work uh using AI on a day to day basis. And then I think so that is something that everybody's kind of you know taking a lot of advantage of. Uh we have sessions I think so every month there are like four, five different slots that you can go, go talk to them and learn more about how you can use AI uh on, on your day to day. Maybe it is an internal tool like quicksuite. You want to ask more information, you are doing things in a certain way and you want to understand how I can better my work or there is some other external tool that you want to grasp knowledge of. You can definitely go ahead and you know talk to people uh who are hands on working on all of these processes.

Speaker C: Mhm.

Speaker A: Great. Many many great opportunities to learn.

Speaker C: Yeah. And then like to add to what part said um like you know we, we have these enablement sessions on like uh how to do prompt engineering. Well like you know what are like government governance and like data related issues. Um how to use like uh AI in your day to day task. So uh we have these enablement sessions that happen um on a monthly basis. Uh you can just join plugin, uh get out uh personally as a, as a B2B like you know as a marketer in general. Right. Like it's, it's important to understand a lot of these um governance policies uh around AI um especially you know like Parth might agree. We work with a lot of um, like ah Fintech and these regulated industries. Right. Like um. So I personally try to uh upskill by reading ah like all these governance data related policies for AI. So there's a lot to read, a lot of material internally. I also rely on like um, external like uh Stanford has a lot of courses so like you can just take uh up a free class and now

Speaker A: it's more and more personal data loss. So yeah it's important to learn more about that. Okay Matthew, you wanted to share as well?

Speaker B: Yeah, um, so yeah I think training has been pretty intentional. We did a 12 days of AI initiative. It's kind of a spin on like the 12 days of Christmas. Um but each day it kind of focused on a different AI related topic. And this just was hosted in Slack. We have a channel that has all the employees included and it just really was a place where each day you could learn a little bit about different, different tools, terminologies, places you could get certifications, hands on interactive activities that you could participate in. And the goal was just basically to make it feel more accessible rather than intimidating and just how to use it responsibly as well. Um, we also have a book club called the Data Literacy Book Club. So on a bi weekly basis I like to attend and just be a fly on the wall and listen in on what people are talking about, the latest AI news and then also what we're building internally. We're currently in the process of rolling out our own AI scribe which is like medical transcription software so that doctors don't have to spend hours after hours, you know like spending time on notes. Um, so we're working on that and it's just really interesting to sit in and listen on how that's being built. Um, we also had an AI idea thon so anybody in the marketing department could submit an idea that was focused on internal operational efficiency. So for example like one of the ideas was um, you know, automating lead notifications for marketing to sales and enriching that data with better context. Um, and then we also had people go to different conferences. So I attended um, Inbound by HubSpot and then I learned a lot about hey, how is marketing being impacted by AI? How do we up level our go to market strategy, what's important. And then I was able to bring all those insights back and present it to the department. So definitely a lot of learning going on. And um, yeah I'm happy that it's kind of been since day one that the Company has made it a real mission to get people participating and actively learning.

Speaker A: Yeah, idea generation. And I like that uh 12 days of Christmas idea of learning different AI, uh tools and uh, tips and tricks. So um, would you like to add something else? I mean all of you or we can jump to the next question.

Speaker D: Yeah, I think I already told everything. So on our end we don't have such a, such a thought out enablement program for we're kind of a small company, about 200 people and the company is heavily engineering so they all know AI way better than us. So it's just same as medium, just being kind of a fly on the wall listening to what they're doing. Uh, so something like that.

Speaker A: Yeah, it's also great. And um, then let's talk about the dark side of AI. What kind of maybe concerns, risks or challenges do you see when it comes to adopting AI across the company? So of course we've already mentioned a lot, the privacy question and uh, this um, kind of question but maybe there is something else.

Speaker C: 1 so as a product marketing manager, uh, my main goal is to change perception about my product in the industry. M which means the kind of messaging, copy language that goes out for whether it's a feature launch or a uh, feature adoption related campaign. It's very important that the words are precise, um, they consider human, uh sentiments, uh, have empathy. So one of my biggest worry and this is something like you know, it keeps, I've kind of like you know my cards are literally up since this whole um, you know the, this whole um, AI system um blast has happened is um, these vendors relying a lot on like you know, external vendors, content creators relying a lot on these tools. Um, we work with them like you know for an influencer campaign. But you know if a lot of this content is very AI heavy it kind of um, goes against the purpose of me changing perception. Um, in those cases my cards are up. I kind of literally re script everything. Um, so somewhere it's a challenge for me, uh, relying heavily on external people using a lot of AI tools, um, without really considering um, the strategy part of it like the real intent behind what we're doing and what are we trying to uh. So that's my challenge.

Speaker D: I would agree. But concern that I have is that now AI is so hyped uh up like everyone AI is everywhere and on one hand it's a great thing, on the other hand there is a lot of noise that I million tools and tools doing the same thing and you don't really know which tool to start adopting because tomorrow there will be another tool and then another tool and you don't really know again, how do you, how to, how to collect all of this together and where to start. Um, so cutting through this noise, I think the most important thing, having a strategy and understanding that, okay, I'm doing, at the end of the day, AI is just a tool. It's not something that will bring you money per se. There is no money button. So as a marketer, you still need to find your customer, understand who your customer is and all of this stuff. So you need to use AI wisely, not just AI for the sake of AI. And by the way, there's going to be another part of it because what I saw, not in our company, because again, we're super technical, but I saw a couple of examples. Leadership is very hyped up about AI, especially investors. And then like, hey, we need to have AI. Let's do AI. Let's adopt AI. Let's do like there's an AI person in the company that's doing that, which is kind of weird. Um, but if there is a strategy, I think it's a brilliant thing.

Speaker E: I mean, kind of echoing what everybody said and you know, it's the same thing like, you know, when, like for me, when I'm working on marketing for, uh, compliance and brand safety. So it kind of becomes very important for us to, you know, have the right language, the right tone, the right claims. Because, uh, if I have an ad copy or if I have a content which I just rely on AI and just, just kind of want it to produce, it will not comply to the brand safety and compliance that I want for a particular brand. And I think so that is one of the major risks, uh, when it comes to global communications. We need to be very sure about what is the kind of language do we have when we're communicating to our customers. Number, uh, two is, uh, you know, even when we are going back and doing post campaign launch or post data analysis, there is a lot of accuracy issues because, uh, whatever data I feed to a particular AI, uh, it has, I don't know, like it thinks that it has the leverage or the convenience of giving me anything just to make me feel better. So which is kind of something that you need to be very aware of. When you're looking at the data. You need to be pretty sure about what data did you feed and whether it is giving you the right response back or no. Uh, and also measurement drift. There can be a lot of hard to know truths or facts about how do you actually measure all of your whether it can be conversion rate, whether it can be click through rate, whether it can be how many people actually interacted with a certain product or a landing page of yours and how many conversions you are driving from it. So all of this is something that you need to have at the back of your mind before you kind of feed the data to AI and kind of wants it to help you to do something better. Uh, and finally, just in terms of shadow AIs that a lot of people generally use, because in my personal life I use a different tool versus in company I might use a different tool and that can happen with everybody. And uh, because of this there is a lot of confusion in terms of which AI tool is better and which is not. And I am probably more convinced with some XYZ tool. And I always want to rely on the data that the other tool gives where in the ones that I'm using on a day to day basis maybe I just might feel that it is not that reliable and I want to jump back and forth. So I think. So that is also something that can be a uh, challenge or a risk or a concern that you know, definitely has to be looked into.

Speaker A: Mhm. Yeah. And uh, if you need to calculate something, AI is not good in it. Yeah, it's better to maybe write a Python script with AI and calculate something than just rely on it. Yeah.

Speaker B: Good timing. Associated too with like the over reliance on AI. I think another side of it is like people's internal fear around their job and their job stability. You know, I think a lot of people feel like, hey, like is this just going to replace me? And I feel undervalued for my work and even at times like possibly offended, like if you were to hand a writer something that was completely generated by AI and just ask them to approve it, um, that, that can be really difficult. And so I think a lot of the conversation at my company as well is how do we position AI more as a partner that makes people better at their jobs and not something that is to replace them or undermine their skill set and their expertise.

Speaker A: Yeah, yes. Sometimes they can be intimidated a little bit by AI. Makes sense. Um, have you been in these kind of situations when you think thought about using AI and you decided no, no, we are not going to.

Speaker C: Yeah, I mean, uh, I think I had a very recent uh, experience. So I was working on a sensitive launch, a uh, sensitive feature launch. Um, and by sensitive it was something that would change pricing perception. Um, and you know, by perception Again you know, a lot of what I do is like putting messages out, promotions, uh, for uh, you know, getting my feature adopted.

Speaker A: Right.

Speaker C: So in that case we realized like um, AI will give factually correct information, copy um, but it doesn't again relate. It doesn't again consider uh, what the customer is going to perceive uh this message as. Right. Like it's um, it might upset them uh, related to the price point. So in such cases we consider not using AI for like promoting, creating promotional materials. Um again this is just one recent example I can think of like you know pricing is a very big human, I would say emotion. Um, people are very um, skeptical when it comes to like anything related to money. So um, that is one example where we decided not to rely on AI for uh, anything for promoting M this.

Speaker A: Thank you for sharing. Yeah, this case makes sense.

Speaker E: I mean I would say uh, that anything requiring uh customer level data, um, or proprietary information of a particular customer, uh, it is like unapproved, uh, or you know, not. It is better not to use any AI because customer information is very sensitive and you know, you need to be very responsible enough when you are handling customer level data. Um second situation when it comes to uh, you know, situation ah, like uh, where there is high compliance risk wherein the content that you are creating or the content that you're presenting or there are some sensitive topics that you are addressing to a particular customer or your particular audience. Uh over there it is very important that Amazon doesn't allow or doesn't want to be. There has to be an intensive use of AI over there. Um, for me how I understand this is a useful default is if you can't safely explain uh, or audit it then it is better not to use AI ah in that particular scenario or situation. It is rather to rely on your own knowledge or your own self. Ask your peers, ah, talk to them, brainstorm with them. Rather than putting everything on AI and just trying to find uh, answers to your questions.

Speaker A: Mhm. And for some specific situations it was for example when you created a case study or something else with the client data.

Speaker E: Yes.

Speaker A: M. Okay, thank you, thank you for sharing. Yeah, makes a lot of sense.

Speaker C: Um, yeah, I think I wanted to add something to the case study. It's very important because you know sometimes AI will pull uh information about a customer based on I don't know, like stored information. So uh, we always go back to a customer we are working with. So you know, their numbers are like perfect. You know like numbers can really break or make stock markets. So um, we have to Be very careful, uh, especially with vendors, customers, partners. Right. Like, how do they want to be portrayed? Um, how do they want, uh, to report their numbers? Like, not everybody's comfortable, um, sharing numbers.

Speaker A: Right.

Speaker C: So it's a very good example. You can't rely on AI completely for, um.

Speaker A: Yeah, makes sense.

Speaker D: Maybe add that in general making like, kind of echoing same, same. Making any sensitive decisions based off AI's work. It's a tricky thing. It's better not to do that, uh, because again, AI is a tool. So garbage in, garbage out. And sometimes I can give you an example. Like recently we connected our gong to AI to Python, uh, and we analyzed 11,000 of sales calls. And based on these calls, we see some, um, picture, but we know for sure that it's not true. So just we see this picture because our salespeople shaping the conversation this way, while we know that in fact it should be different. And we have other signals not from AI. So we did this with AI, but we did not use this eventually. So this is important to take into account.

Speaker A: So because of different prompts, it changes, um, the truth a little bit. Or it was.

Speaker D: No, it's not because of different prompts. It's more of like we have a base of 11,000 calls. And then, yes, we're asking AI to just summarize these calls.

Speaker E: Right?

Speaker D: So we have different segments. I mean, uh, close one poc, whatever, any segments, or like negative versus positive sentiment and all these things. So we asked AI to summarize this. But AI is summarizing the actual content of these calls. And we know that these calls were shaped by the salespeople salesperson when he or she is having this calling it some way. So yes, we know that this is a summary. The summary is correct. But we are not sure that this is the summary we want to like the Overall summary of 11,000 goals. We need to take into account, um,

Speaker E: in the product marketing area.

Speaker D: So at least we need to take this into account.

Speaker A: Yeah. Interesting. Thank you. Thank you for sharing. Maddie, have you been in these situations?

Speaker B: I almost feel like sometimes it's funny working at a larger company, it's. The answer is almost no until it's yes. Um, because there is so many different tools out there. And I think especially in a highly regulated industry like healthcare, we have to be incredibly careful. I mean, when you guys are talking about, you know, customer information or patient data, like, obviously there's hipaa compliance and SOC2 and all these different things. So it does really kind of hamstring us in terms of what we can use AI for in the first place. Um, and also I think in addition to this, even though it seemed like AI could really help on the sales and customer experience side of things, I think that you know, having a white glove approach, especially with a high ticket um, solution that we're selling, I think it still requires a lot of humans to be involved at every step and to have a lot of QA and critical thinking, thinking. So um, yeah, I think there's definitely a lot of areas that we have not fully integrated AI with. And right now it's just sort of this like slow rollout, governance first approach, very measured. So that's kind of the direction.

Speaker A: Yeah, makes sense. Thank you. And um, everyone says that AI saves time. Yeah. Right. But how much exactly? In your personal experience, what do you think? How much time does it really save? Um, say for example during the day or a week or so.

Speaker C: Yeah, it's hard to quantify, quantify, but it definitely has uh, made me faster. Um, like I mentioned earlier, um, I can iterate faster if I don't like something, um, I can generate variants of it faster. I don't have to spend a lot of time in two copies, uh, creating three copies or full copies. Now I can generate ten variants then, then decide, oh, this is the waiting time like call. Right, like, so it's definitely made me faster, more um, efficient I would say. Um, like I literally start with AI for my first draft, right. Like I fix my grammar with AI. Don't have to now manually like read through so many sentences.

Speaker B: Right.

Speaker C: Like I, I do a lot of quick data analysis through AI. Um, you know when I'm like sifting through like, like alexei mentioned like 11,000 customer related uh, line items in an excel 10 minutes and I have my segregation, I know the patterns, I can figure out commonalities. Um, if I were to say a very recent example, um, for again a product launch, I was trying to do some competitive um, benchmarking analysis. So typically I remember before AI it would be like analyzing social media platforms, analyzing competitors web pages separately, um, then analyzing like feed the customer feedback separately. Like now I remember all I had to do. And again like you know, it's just because Amazon has this one Amazon Quick suite tool. Like just one tool, it pulls data from everywhere. All I had to do is ask question. Competitive benchmarking. I had data and like 30, 40 minutes. So it's saving me a lot of time.

Speaker A: Um, saves hours. Yeah, great example.

Speaker E: Just to add on to the same scenario when you're like you know, uh, running Any social media or paid media campaign. And when you want to do competitor analysis, usually previously, what used to happen is you go to a number of different uh, websites, kind of look at competitor benchmarks, kind of see what is the kind of right CPC or CTR or CBR benchmark for a particular industry and then you kind of base it according to that. But now because of AI or because of Quick Seed, what has happened is I don't want to go out and look at some other benchmark. I have my own company benchmark and I, I just have to ask a question saying that if this is my campaign, if this is the intent of the campaign and this is the kind of idea customer profile or this is the intent audience that I'm going to market this towards, uh, what should be the ideal benchmarks for all of these conversion metrics so that I have it in back of my mind before I launch the campaign and it beautifully does it in like no time. So I would say in terms of more data accuracy, data accuracy is only possible when you have such a level of sophistication wherein you are feeding only one particular data. Now what happens with general tools is that general tools have general information or general data. So it is very good for general day to day activities and saving time on general day to day activities. But you want to do one particular or specific task, it is much better to isolate the data, look at only one data, keep feeding it continuously so that it gives you very relevant and very useful information. I think that is where we do save a lot of time with all of these Amazon.

Speaker C: Yeah, and I think now that we have spoken so much, I think one of the tool I would actually recommend um, marketers to try is Amazon Quick Suite. Um, because it has definitely been a game changer like part mentioned, right? Like I no more have to like go to different tools, different sources, everything in one place. Uh, saves me a lot of time.

Speaker D: While you guys were speaking, I just wanted to chatgpt that I'm using pretty heavily every day ask hey, could you estimate how much time you saved me last year? And Judge is saying that on average we could estimate it saved me 1 to 1.5 hours a day, which is 250 to 350 hours a year. So yeah, a day, a week I think roughly.

Speaker A: Interesting. I need to ask mine.

Speaker E: I mean I would say that I wouldn't even be surprised. It's just because the amount of you um, know, low hanging fruits that I just pass on to an AI to just get it done with, I Mean it has become. Because I used to waste a lot of time just to write an email, just to write, you know, create a draft or you know, just respond to somebody's uh, you know, very data heavy question. Now what I just do is I just, you know, uh, all these low hanging fruits, just pass it on to AI and you know, have more dedicated work done from my end. So I mean I wouldn't be surprised if AI is saving like one day in a week for me. Uh, for sure, absolutely.

Speaker D: But I think on the other hand you see, like let's imagine Slack, there's a conversation on Slack and I think all of us now doing that. Like there's this question, you just go to ChatGPT or whatever you're using, like, hey, could you please just craft an answer so it devalues a bit the answer itself. If in the past it was like super, this was a human answer, now you see that this is a chatgpt or something and you're like, okay, you're answering with the ChatGPT, so I will ask you, I'll see.

Speaker A: Yeah, maybe you wanted to add.

Speaker B: Yeah, yeah, no, I think it's very similar to other people's answers. Like it could be up to one day a week probably. I mean it's just crazy. I think too, I'm um, susceptible to over relying on it at this point because I will literally have a brain dump of ideas and be like, wait, I need to like say this to an executive. Can you edit this and make it sound better? So, um, you know, just like little things like that. Um, but yeah, definitely I would say that it probably saves me anywhere from a few minutes to a few hours. Just really depends what I'm working on. So yeah, it depends how deep I need to go. And um, of course like fact checking the AI too. So, um, but even with that, definitely a lot of time. And I think the biggest example could just be keyword research, honestly, like coming up with new keyword ideas. I mean, I can't imagine trying to do that manually now in a way.

Speaker A: Um, yeah, yeah, before it could take hours. Yeah, no, it's fast. Yeah, that's great. And a tricky question, uh, does AI make you more competent in your work?

Speaker B: I would say yes and no. Um, I think that I would say yes on the positive side, like understanding if I want to do research, understanding that information more quickly, having access to that knowledge that I may not otherwise have as easily and it definitely makes me more effective. But at the same time like I thought about this the other Night. And I'm like. It almost feels like critical thinking can become optional if you're not careful. Like taking the AI outputs at face value, especially knowing that hallucinations can still happen, I think is dangerous. So I think it, you know, if you stay intentional about validation, accuracy and using your own human judgment, I think it can make you exponentially more competent, um, and quickly, you know, quickly be able to do things. But I think it almost makes critical thinking optional depending on what you're doing. So I would be careful a little bit.

Speaker A: Yeah, yeah.

Speaker C: I think agree completely with Madison. But then I think um, it's making me competent in a very different way like you know, using AI. So uh, my angle is now that I save a lot of time, uh, with my day to day stuff I can upscale in my product. Right. Like tech is changing every day. Alexey knows like Kubernetes, like it's hard to keep up with the way uh, the whole cloud native, uh, the field I'm in is changing like so now I spend a lot of time upskilling about the product side of things. Right. Because AI is taking care of a lot of these like day to day stuff that I was working with. So it's making me competent, uh, or I would say, um, helping me um, you know, upscale in other areas, uh, as well. So it's a plus point for me.

Speaker A: I also like how AI can explain some technical stuff, some very complicated technical questions. Yeah, it can answer it very easily and uh, you can understand something that was too hard to, to figure it out before. Yeah. Um, so part or Alexei, would you like to add something?

Speaker E: Uh, I mean I would say that um, for me in my current role, uh, AI has increased the range of the amount of work I can get done in a particular given time. Like um, the amount of time I can do multiple different tasks, uh, in a given day has kind of increased because of course there is saving time. There is one task doesn't take like four, three, four hours, rather it takes like maybe 30 minutes. Right. So over there is definitely where.

Speaker A: I'm sorry, yeah, I just wanted to interrupt you a little bit because yeah, of course it saves time, but does make you are more competent in some way.

Speaker E: So I mean it does save time and makes me competent enough to kind of get things done in a much quicker fashion. Uh, you know, if I have to be more consistent and of course consistently, uh, consistency kind of becomes uh, more uh, important when you have the right kind of prompts that you can rinse and repeat. So you have to Take that effort to kind of have your own prompt playbook wherein if you are doing certain tasks in a certain given week on a repeated fashion so it will make you competent enough if you have some playbook that you go back and refer to every single time when you're doing that task. Uh, but having said that, one thing is AI can raise the floor. Like you know, it can elevate you, but it cannot completely define what good means. Because at the end of the day what is good at your job or what is good at you delivering results, uh, is something only you can determine. So I think so there is where we need to have a very clear definition or a clear gap in terms of how much elevation are we getting through use of AI versus how can we define good by our own work? Uh, how much time can I invest in what work to make myself more competent in the current work that I'm doing or in the current role that I'm doing.

Speaker A: Yeah, thank you.

Speaker D: And Alexey, yeah I would agree with all the guys said so yes, AI is doing us more competent. But on the other hand, just to say something, on the other hand, um, I think AI like everyone now can produce good looking content as Madison said and all of us are good in this. Like you're doing any report or whatever for stakeholders and like, okay, there's just one bullet point. I need more hr, GBD or I don't know, you're interviewing someone and you see that this test assignment was done with AI. Uh, you just can see it. So this thing and a bit changes the balance between excellent, between the rock stars in marketing and just kind of average people without being fs. But yeah, I think that's important just to keep in mind.

Speaker A: Thank you. And the last question for today is about the future. Uh, how do you think will our marketing B2B marketing teams change uh, with AI? So let's just think about um, how our marketing uh, will look like in the future. What do you think?

Speaker C: Um, I think definitely there is change. We, you know, like, like now in, in our conversations if you see like people have stopped talking about SEO, it's all AEO right now. Answer engine optimization, as, as uh, as marketers and you know like uh, although I'm a PMM product marketing manager, I still have the mater and b, um, now I have to adopt, adapt to these new tools because we are no more competing for page views. Right. Like people don't even go to your web pages anymore. Now I am going to compete for like how is my companies or my Products information going to be placed in these uh, LLMs, right? Like um, so focusing on intent based content like ensuring you provide specific answers, very specific details in your blogs, in any documents. Because if you don't do it ah, AI is going to pick some opinionated pieces from you know, somewhere, some like you know that you think wouldn't work for your product. Right. So um, upskilling again in this how to deal with this answer Engine optimization or generative engine optimization era becomes uh, important for future uh and I don't know, in five years there might be something else. Uh, but at this point that is what I see. That's where B, uh2B marketing is more like more intent based up be uh very specific, give detailed information because people no more go to your web pages, they just ask a question and they get answer right? Wrong. Depends on how you have uh, configured it. So that's one aspect uh that I'm seeing in terms of future for B2B marketing.

Speaker D: I'm going to be generalized this to say that um, even the past we're doing websites for people, now we're doing websites for LLMs. And the whole entire buyer journey is moving from Google Websites like from your side to LLM. Um, like if in the past we were creating top funnel blogs, mid funnel blogs and bottom funnel blogs, now the whole journey happens within ChatGPT. Hey, I have this problem, what can I do? Then ChatGPT gives you the list of answers, what kind of tool can I use for that? And then the list of tools. So you should be there. Um, that's why. Yes, it's changing a lot actually. We just need to adapt and make us be listed there somehow. There are some techniques and let's try to research and then upskill there. Yeah.

Speaker C: And then I also as, as a marketer I think basics of marketing, right? Like understanding human psychology, understanding your icp. Um, there are jobs to be done. Like their problems, their pain points, all these things still don't change, right? Like uh, you still have to understand this very deeply about a consumer you're selling to or a user you're selling to. And I don't see AI doing a very good job. Right. Like as a marketer I every marketer uses different cycle like different frameworks. I use my own frameworks um, when I like you know, promote something right. But like all those things still remain the same. Um, I don't think AI is there where it can like literally like replicate how you think about the problem, how you want to um, promote your service. Um, so like that, that understanding those core for marketing folks is, um, still going to be there. I don't think AI will replace that.

Speaker D: Yeah. At the end of the day, you're still selling to humans.

Speaker A: Yes, for now. For now.

Speaker D: Yeah. Because then it will go to AI, like, hey, could you please buy this thing for me? And then AI will buy all of the tools that you need down the road.

Speaker A: Yeah, yeah. Maybe in the future AI will make decisions on and purchase. Yeah, we'll see. So, um, part or medi, would you like to.

Speaker E: Yeah, I mean, I would say one of the big shifts that I, um, I do see coming in in B2B as a marketer, uh, is number one is what is the differentiator. Because nowadays everybody is using AI, right? Uh, so it would be very important for us to produce a lot of content which has original point of view, which, uh, has very strong evidence, which kind of gives unique examples and clear positioning. And I think so that will become more important is because everybody now, even in this panel as well, we know to differentiate which is an AI answer and which is a human answer. Right. So our eyes are kind of, Our eyeballs are kind of catching the very significant difference between what is an AI and what is not. So I think so original point of view will definitely become, uh, something which is very important, especially when it comes to B2B. Because, uh, what I feel is in B2C, there can be a lot more integration even on the customer side as well, wherein even the customer is an AI and the product is an AI and who is promoting is also an AI. But when it comes to B2B enterprise, this kind of becoming will be much more like, there will be a lot of more free friction, uh, to actually kind of completely convert everything into an AI world in the B2V space. Because at the end of the day there would be a lot of fluctuation when it comes to price points, uh, on a B2B scale. Uh, so that is one thing that I feel would be very, uh, important or the learning curve has to be passed on. Especially when we look at, uh, ABM messaging, it will kind of be like how I can test different, uh, ABM messaging for different industries. That is something that would become very easy to personalize based on each account base. Uh, so whether I have a particular client, how do I message that particular client, how do I make sure that that, you know, account has briefs done in a more tailored fashion, how can I tailor my landing pages in a much more efficient way for that particular industry? Like, let's say, for example it's a medical industry or a tech industry or automotive industry. I can tailor and create multiple different funnel stage journeys that would become more important or that would become more easier, uh, you know, moving forward in B2B. But at the end of the day what I feel is one measurement, second is uh, having original point of view and third is personalization. All of these things kind of will, you know, be the decision factors for anybody who is converting into a potential customer. And marketers do need to really think about, you know, how we can kind of pivot ourselves or make ourselves uh, competent enough to, you know, create, uh, create journeys around this.

Speaker A: Thank you. Yeah, um, and human work will become luxury. Yeah M Maddie?

Speaker B: Yeah, I mean, I don't know how to answer this question. I mean I think what was said about user generated content and authenticity really stands out to me. You know, especially if LLMs are scraping sites like Reddit, for example. I think that it's really important um, to you know, be human in a world where the, the Internet is now going to be increasingly oversaturated with AI generated content. Um, so, you know, being able to differentiate yourself and um, you know, really appeal to authenticity. Um, and then yeah, LLM visibility is everything. So you know, leveraging different strategies to appear in, you know, the, the output whenever someone asks the question. Like in our business, if someone's searching for AI scribes for dermatology, like we pretty much show up now. And I think that is a lot to do with the fundamental marketing work we put in on the SEO and GEO front, earned media and different avenues like that to make sure that we are visible in LLMs. Um, so yeah, I think that's kind of the name of the game at this point. Um, but yeah, and then I would also say the point on personalization, I think that's one that we need to work on at our company as well, is you know, just really using AI to make everything as personalized as possible for that user and their intent and even you know, their job title, the company they come from, like using all these data points to personalize their experience.

Speaker A: Thank m you so much. I especially loved the uh, your idea about being human, being authentic. Yeah, it's very important in the era of AI. And thank you so much, uh, to all of our amazing experts, amazing speakers for sharing your expertise. It's very practical and it's bright and I um, think that can um, inspire marketers and both we could just use something. So it's very practical, very useful. So thank you so much for your time and for your experience, expertise. Thank you.

Speaker B: Thank you so much.

Speaker E: Thank you, everyone.

Speaker C: Thank you.