How Descript Uses AI to Sharpen Marketing Messaging (Full Transcript)

[00:02:21] Speaker 1: Um, um, um, um, emulate the taste of, like, like, non-chicken nuggets. So, these nuggets are made from chicken, but they're made to emulate the taste of non-chicken nuggets. Dope.

[00:02:36] Speaker 2: Oh, that's much better. Edit all the blather out of your videos, because my time is very precious.

[00:02:50] Speaker 1: Oh, that's fire. Make it less teal and more cerulean? Sure. Replace your background with something more fun? The cold void of outer space. Let's boost that sound quality. Emulate the taste of non-chicken nuggets. Emulate the taste of non-chicken nuggets. Dope. It needs more style.

[00:03:04] Speaker 3: It needs more clips, more gifs, more more.

[00:03:12] Speaker 2: Well, I might have made it too gnarly.

[00:03:19] Speaker 4: Hello. Welcome, everyone. Thank you so much for joining our webinar today. Super excited to have you all today. We have our guest and resident product marketing manager at Descript, Matthew, joining us. How are you doing today, Matthew?

[00:03:34] Speaker 5: I'm doing good. How are you?

[00:03:36] Speaker 4: Good. Excited to do this webinar that's all around marketing messaging. Before we dive in, and folks are coming in, would love to hear from the chat where people are coming in from. Also, because we're talking about how to use AI today, would love to know what everybody is using AI for in their workflow and what their comfortability level is, too. While folks are coming in, Matthew, do you mind sharing a little bit about what you do and what your thoughts are when it comes to using AI?

[00:04:09] Speaker 5: Sure. Yeah. Like you said, I'm a product marketing manager here at Descript. Generally, that means helping figure out campaigns, positioning, the way we bring new features to market. For today's discussion, I'm actually going to go back to a feature that we launched and a campaign that we rolled out last fall. Besides that, I also wear a couple of hats here, put on events and help us go to trade shows and organize all of the activity. Plus, a little bit more less creative things like research and talking with customers and users to just generally understand how people are using it and where we can continue to make the product better and the ways to show up in the world as Descript.

[00:05:04] Speaker 4: Amazing. Yeah, I'm looking in the chat and it looks like lots of folks from all over. We have folks from New Zealand, from Idaho, and using AI for all sorts of things. And we have people that are very comfortable using it too and using it even using entrepreneurship in high school. So I feel like there's so much you can do with AI, which is super exciting. And for today, we're going to show how we actually use it at Descript in our marketing. So let's jump right in. Okay. So for general webinar housekeeping, this video, this event is being recorded. It's live. If you guys have questions, feel free to throw it in the chat. We'll bring it up during the demo or if not, we'll bring it up during our Q&A. We also have Gabe in our chat supporting us. So shout out to Gabe. And if you need more help specifically around Descript features or so, you can go to help.descript.com. But today, again, we're talking more about marketing messaging and how we use AI. So please give us your questions as we go along. Cool. So for today, we're going to talk a little bit about the framework around how to actually use a strategy for messaging with AI. And then Matthew is actually going to show us how to apply that framework, like he mentioned with a previous launch that we had done here at Descript. And then we'll open up for Q&A. Cool. So this is a webinar for marketers, content teams who want to create product videos to do more without adding more production time and those that are creating messaging as well in their videos marketing. You'll walk away today with a framework of how to land your message based on research process that usually can take days compressed into one working session. Can't repeat this enough using AI. So right now, when you are looking to create a message, there's a big research part of the process. And that usually falls into these areas where you're gathering intel from competitors. You may be looking at their website, their ads, the content on their socials, just seeing how they're positioning and what the messages they're saying and also how much they're spending on it, too. And you may be also looking into opinions and reviews from your customers on like Reddit sites or review sites just to kind of get a gauge of what people are thinking. And also, you may be looking at press and reports around how the market's being framed and also looking into your own customers voice of like what your own customers are saying, like what their problems are and trying to understand the language that they're using. So all of this research often takes a lot of time. And you're gathering a lot of information. And all of that is done. But you still need to make it into a message. You need to make it into part of your campaign that you're running and just make it applicable to your target audience. And that can take a really long time. And so either you do it really slow or you're doing it rushed. If you do it wrong, it may take a really long time. If you're doing it slow or you're doing it rushed, if you do it wrong, it may not land. Or if you are not thinking of the right message, it will not connect on that deeper level for your customers as well. So you want them to be recognized in the message. And doing this process can really help. And especially if you're doing it in a way that speeds up the process and makes everything a lot quicker. So the framework to kind of reposition how you approach this is around first you're listening. You want to gather the intel from your competitors, market chatter, and how people are receiving it. And then you want to take that and you want to learn from it to synthesize the message that you want to create and then launch it to either your go-to-market plan or the channels that you want to roll it out to. And so for today, we're going to show you how we actually do this using a workflow that is using AI. So I'm going to pass it over to Matthew.

[00:09:17] Speaker 5: Yeah, I'll start talking here. This framework is a very typical and long-used marketing framework. And so early in my career, before the age of AI, I did a lot of this manually, right? Maybe you have notebooks or files stored full of history of what you've seen other people doing, whether it be competitors or what users are saying, what people are posting about the content. You use that to, you know, it's part science and a little bit of art to try and, like, come up with what the right message might be, where it's a nice overlap between what people, where there might be, like, an unmet need or what people are looking for in terms of, like, a future becoming better. And then you match it to, like, how your product could be used in their real life. But with AI, we can now compress a lot of this and actually not just get the research phase done a bit quicker, but also have a conversation, explore some loose thoughts, some different angles. And so, you know, before I start showing you kind of, like, how I do it and the tool I'm going to use here is Claude, just because I'm quite familiar with it, but you could kind of use any one of the main ones for this. I just want to set the stage as to, because I'm kind of taking us back to last fall. So we launched On the Lord, the AI co-editor in Descript last summer. And the team was hard at work at adding more features and capabilities to Descript, but also in a way that makes On the Lord work even better. And so one of the big capabilities that we had worked on was called media understanding. What that means, that's just a technical term around how the AI works with your media, to basically give On the Lord the ability to actually see what's in your footage, not just read the transcript. As you know, we're really great on the text-based editing. The transcript comes in. But having that extra capability to understand the visuals, understand the content, and work with it intelligently is something that we wanted to roll out pretty quickly after. But the marketing challenge was, so we had this capability, but I didn't really know what the message was. What's the story? What's the best way to actually take it to market, to tell people about it? Is it B-roll? Is it about time savings? That's like a benefit-led statement. There's a bunch of different directions we could have gone. And so in a short period of time, I couldn't spend all that time exploring all the different angles. I needed to get to a confident stance as to, okay, I think this is a recommendation. I think this is the way to take it to a market. And so I applied this framework quicker and compressed the time. So, yeah, I'm ready to show what I did because I've essentially created a bit of a replay of what this compressed experience was like.

[00:12:27] Speaker 4: Awesome. Let's get into it.

[00:12:29] Speaker 5: Cool. So first things first, I wanted to... I gave a big prompt that essentially says, look, I have an idea. Here's roughly what the product is going to do. But first, I need to research the competitive landscape. So I'm asking it to look at some competitors, look at what people are saying on G2, on Reddit, how people are talking about AI video features because, again, this is last summer. So mostly people are talking about, you know, like, I don't even know, like, image gen or they're talking about avatars. But what about features for editing? And then what the gaps might be. And so I essentially ran this query, and I'm going to speed it up, like, 15 times because it kept... It had to look at a bunch of stuff. It was using web search. There was a bunch of stored information I had on my computer, competitive analysis. So it's just going through everything and kind of gathering all this context. And this is the way that I like to work with Claude, where I don't have a clear answer yet, but I need you to bring up everything so I can kind of sort through and figure out what I want to talk about, figure out how I want to position this. And so that first step is what we just go through, which is compressing multiple days of research and reading into a couple minutes. And so what Claude found was that the market kind of splits into three camps, right? There's generators, there's clippers, and then there's enhancers. And so there's some potential positioning gaps that we could launch this new feature, this new capability into. But as I saw this response, I remembered something, something got triggered in my mind, which is our PM had told us one of the capabilities, this thing called semantic stock selection, wasn't technically ready for when we wanted to launch it. And so I gave a new prompt to essentially say, hey, you know, one more thing. This is not ready, but something like visual content matching is, and also a smart B-roll suggestions is something that works pretty well. So how, you know, what do you think about this? I wanted to get like a bit of a playback with all the context of the research in mind. And so I sent that prompt off. And this essentially finishes that initial listen phase, right? So now once we had gotten that back, Claude said something kind of interesting. Instead of me just jumping right to, okay, we've got the research, let's go in and start writing up my plan, my document. I wanted to have a conversation, right? You know, you don't want to treat it like someone who's giving you the answers, but rather like a smart associate who did the research, and now you're going to work through the strategy together. And so there's three moments that I'm going to show you next that kind of brought that experience of a conversation, a thinking partner to life. And so the first one, you'll see it essentially looked at, you know, it's essentially saying, look, this constraint is actually helpful because that gives you a positioning angle of AI that works with your footage, which is kind of cool because, of course, we're really good in – our product, Descript, is really good when you have recorded media that you're working from. And so it's hard to make some suggestions around what to lead with versus what to hold back. And so, you know, I'm essentially reviewing this. I'm thinking about it. And so let me just fast forward a bit because I want to get to the next moment where I give it another prompt. So now I say, okay, we know visual content matching is our strongest capability. Let's move to that next phase. I'm thinking what I want to do is actually break it up into a couple – a few different stories, and I'm going to call them story beats. And part of the reason is with this campaign, I wanted to start putting ads on LinkedIn and YouTube and Instagram for that matter. So I wanted to have not just one story to tell, but over the course of a month maybe, like, have different creative in market. So I was looking for a positioning that allows me to tell multiple stories. And so here was my original thought. Okay, beat one, we'll do visual content matching. Beat two, we'll do smart B-roll suggestions. Beat three, we'll do semantic search across project files. So I gave it to Claude. What do you think? Push back if you think there's a better approach. And this was the first interesting moment that happened. Claude pushed back hard, said it itself right here. I'm going to push back. You're doing three capabilities, three beats is the wrong frame, and here's why. And so it's essentially saying I'm organizing. I'm just trying to tell the market what the product does, but not framing it in a way where people would care. Like, what does it do for me, right? And so we have a, you know, it gives me all kind of discussion. It ties it back to the research that we did at the start of the session. Then it starts to suggest a bit further down an alternative approach. What if I structured the entire campaign around one workflow? Because this is, you know, it's basically saying, look, don't keep talking about all these different things. Don't just try and talk about the capability. Actually turn it into something that would be relevant. And so it proposes a use case for a product marketer making a demo video, a content marketer repurposing webinar footage, and a brand marketer assembling a video. So I thought that was sort of interesting. It's the same capability, visual content matching, but three different personas, much simpler to message, easier to produce, hopefully a clearer story. And so this is that first moment where, like, a structural insight came through a discussion between myself and Claude. Yeah, go ahead.

[00:19:20] Speaker 4: Do you feel like saying push back really helped bring this into light, into the conversation?

[00:19:31] Speaker 5: Not because it said push back, but because what it suggested immediately I recognized as a clearer approach. Because when I said, hey, I'm thinking about three beats, I was in an expansive mode. I was trying to cover all my bases, you know, based on like what I learned about the feature from the PM. And it essentially said, this one is super clear, especially with the information you told me and with what we found out about the position and gaps in the market. Right? So it pushed me to do something which a marketer should always do, and that is find the benefit, find a scenario that's recognizable. So when you talk about the feature, it makes more sense to the audience. So now I said, okay, I like that approach. One core workflow, three marketer scenarios. Yeah, much cleaner. And so I said, hey, can you help me crystallize that into a core value proposition? Something concise that captures this insight. So I send that off. And what was really cool is it comes back, starts to document it, because this is something I really like when I'm working with cloud specifically, is create some documents, even if they're a version one, then a version two, as we continue to make terms or narrow in, I like it to have something that it can refer back to, especially if it's like a research report or in this situation, or in this particular case, a draft of the brief. And so at this point, it said something that I didn't even realize was a positioning at first. And that was marketers are the subject matter experts of their message. And so it kind of brought this up and I immediately glommed onto it, this idea of like, okay, I want to position this as respecting the user's expertise. Like you're the expert of your message. And when I think about Underlord, the co-editor that we just launched, we talked about it being a co-editor. It's there to help you. You control the message and the vision. It's there to do the drudgery, to help make it come to life. You can talk to Underlord and it will do what you need. And so Claude picked up on that and crystallized it. Like, you know the message, we'll make it look professional. And that ended up becoming the core value proposition for the entire campaign. It didn't come from a brainstorm or a meeting with the creative team or me working through some process in a template. It came from this conversation. So you see here now, I said, okay, you know your message. We'll make it look professional. I like that. Let's stress test it. Who would disagree with this positioning? What would our competitors say in response? Where is this vulnerable, right? I just wanted to kind of check my priors. And, you know, when you're working with a chatbot, it can tend to reaffirm what you say you like. So in this particular situation, I'm taking it with a little bit of grain of salt, but I also have a high confidence it's going to be useful because of the research from the very first phase, that listen phase, where I'm gathering all the context that I have. So I know it's not just making it up based on, you know, what's already in its training, but it's going to refer to this content.

[00:23:10] Speaker 2: Right?

[00:23:10] Speaker 5: So I stress tested it. Who would disagree with it? And it didn't pull punches. It actually told me like what Synthesia would say, which is in your position, you won't need footage at all. You could just generate what you need and you can get exactly what you want from media generation. Haygen could claim they work with existing footage for video translation. Adobe, what's interesting is Adobe launched something back in, earlier in the summer, and then they've since been building it out. But, you know, essentially it's given me other ways to think about it, including internal skeptic. Like what my own sales team might say as I'm preparing them and trying to explain to them, here's what this is, here's how you can sell it to enterprises. You know, maybe what someone else like myself, what's really nice to have, that's what, when it says a sophisticated marketer, that's something I have built into my system where it says, give me another person's experience. If someone else was trying to come up with this plan, what would they say? And I really like that because, you know, it feels like I'm parallelizing my brain. So each counter clarified where our angle holds and where there's some strength.

[00:24:30] Speaker 2: Right?

[00:24:31] Speaker 5: And so after all of that, you can see it lands with the strongest version of the value prop. After stress testing, you're the expert on your message. Underlord is the expert on your footage. So that's a positioning frame that I can take to the creative team in order to start coming up with creative ideas to actually figure out what should we put on the website, what should the videos look like, that sort of thing. So I said, okay, you know, I think what we can do now is this position makes sense. Let's take everything from these earlier phases and actually create the GTM plan. And I give it what I need, kind of like, here's the template you need to fill. Here's the template I always use. Here's the sections that are relevant for when now I'm going to work collaboratively with my team. That's stakeholders on the product side, as well as the creative team to bring this to life. So, you know, it's the usual stuff, core message, describe the three beats, the scenarios that we want to display, what channels we're going to put this content into, the timeline, success metrics. You know, I give it some other stuff from before, not shown here, but like we had metrics already built in both from a marketing perspective, as well as a product perspective. So I gave it all that information separately. And then I said, make it a real plan I could hand to my team. And then I also like to remind it to use our brand voice principles, just so it doesn't give me, you know, like jargon. And so as this is loaded, I just want to point out something. What happened here is the AI didn't write the positioning. I also didn't write it just straight out of my brain. What it did is it compressed the research. So when I sat down to think strategically about how to bring this capability to life, I was working with the evidence that it gathered. The core message came through the conversation, the conversation. And then, of course, working with Claude here helped recognize and stress test what might be the right bet to take. Because, again, when it comes to marketing and launching campaigns, there's still that bit of art on top of the science, and you're taking a bet as to how you should take something to market. So now it gives me a summary of what the plan would look like. It looks pretty good to me. I generally review it, like just review the doc, or do what I'm saying now, which is push it to Notion, and then I'll review and make some edits, and I might do a couple more rounds back and forth, make sure it really fits with what I'm thinking. And then, great, then I have a Notion page, and that's where we start working with the team. That's where I do the kickoff. That's where I speak to the creative director and say, okay, what do you think about these concepts? How can we take it to life? And so that's how we take everything from the listen and the learn phases and feed it into the launch plan. And we came up with a structure of three beats, all built from a single insight, that there is that one capability that's really strong and actually is relevant to all of the target audiences' use cases. And I can actually show you what this GTM plan looks like. I'll just swipe over. When it pushed it to Notion, this is what it looks like. Gives me a nice executive summary, right, setting the context for what this capability is, a little one-level, sorry, one-sentence description, why we're doing this approach, and then, of course, the normal stuff, positioning and core message, supporting claims, what to say, what not to say, because, again, positioning is about, like, really getting into that spot where your message will be clear and resonate. I described the beats. What I like to do is thought starters, right? I set the scenario, and then I'll actually give, like, a starter for the ad. I'm going to show you what the real stuff looks like in a second, but I just wanted to show, like, this isn't the actual script that we turn into the content. This is what I put into the team, and then we start concepting. We start seeing storyboards. They start making some example version of the video, and then we refine it down until we're like, okay, this is great. And so, you know, it essentially hits this concept, but it modifies from there. It's not just me thinking about it. So we had three main beats, product demo, customer story, and brand video.

[00:29:24] Speaker 4: So going back to the beats that you were just showing, within each beat, are those different scenarios, workflows, are those a different video and the messagings that go into it, or how does the video – how do we pull a video from the messaging from this now?

[00:29:46] Speaker 5: Yeah, so what we did is, like I was saying, with the channel strategy here, I wanted to have multiple videos going into, like, Instagram ads and YouTube and whatnot. And so these are three different scenarios. And then on top of that, what we actually did is we had our – Ramdi, who does a lot of our videos on YouTube, he actually demoed these as well. And so what I'm going to swipe over to right quick is just showing you, this is the landing page we made to talk about this feature, and this is where the ads drove to. And on this page, you'll see we had the product demos, the customer stories, and the campaign video. And so I'll just let this one play so you can hear it.

[00:30:36] Speaker 6: Turning killer product shots into a killer demo video doesn't have to take years off your life. Meet Underlord, your AI sidekick that turns your photos and videos into a show-and-tell masterclass.

[00:30:49] Speaker 5: Nope. Sorry about my buffering.

[00:30:54] Speaker 6: Your AI sidekick that turns your photos and videos into a show-and-tell masterclass, all without you even lifting a finger.

[00:31:00] Speaker 5: Oh, that's weird. I'm sorry, guys. I'll try the other one. Oh, there it is. I will try this one because this one's been fine.

[00:31:10] Speaker 6: And bam. Sift through all your customer interview footage to find the good soundbites ever again. Just have Descript's AI do it for you. Let me show you how. So let's say you're working on a case study video. You got your narrative, some screen recording, some B-roll. All you need is some soundbites from the customer to give it credibility. So just upload all your customer interview footage to your project, then pay a visit to your pal Underlord down here. If you already got a video going and just want to add the soundbites, enter a prompt like this. Or if you're starting from scratch, you can just tell Underlord to make you a whole video just using quotes like this. See, Underlord can not only read the words in your script, but it can actually watch your video. So it knows the context and other stuff that goes into choosing the best quotes. So you can just kick back and wait a few seconds while Underlord pours over all that footage. You know, the part that would have taken you like half the day. Let's be honest, the whole day. And bam. Next thing you know, you have your quotes placed in your video for maximum impact or a narrative made out of the best quotes from your interview. Then you'll probably want to adjust the timing, add or tinker with the transitions, make a few other minor edits. But even so you'll be ready to export a professional looking video in a fraction of the time it would have taken you on your own. At least I think I never really understood the whole fraction.

[00:32:21] Speaker 5: So that's an example. And what I'll just show you really quickly is like one of the videos, the ads that we made in support of that. And so that's an example of a short 15 second ad that talks about finding customer quotes. And then the video I just showed you of Ramdi talking about how you'd actually do it was on the landing page as well as on our YouTube. And so those match up with, that part matches up with B2, customer story, right? So folks who are cutting webinars or cutting case study content into, and we're showing here that you don't need to have to scrub through and find it all. Under can see, and it can also see the transcript. So I can just drop the content in. I can find it for you and then put it together.

[00:33:29] Speaker 4: Nice. So basically essentially just reiterating what you did is like, you created this message and now you had all these different kinds of assets, video assets to create like either for the ad or on the landing page, essentially. Exactly. Cool. How long did this whole process take compared to like, if you were to do it manually?

[00:33:49] Speaker 2: Yeah.

[00:33:50] Speaker 5: So, you know, this whole thing, I would have said, I think that's like probably like an hour and a half in, you know, like a focus working session. Obviously once it goes to the notion page and we start working with other folks, like then it's, you know, I had like a kickoff meeting. We had a good discussion. That's a half an hour then actually working and refining the videos themselves because, you know, I'm not shipping this doc. This doc is just what we work from, but this is not the story that actually goes to the internet. But once we got past this part and we moved into execution mode, I would say this like initial planning and collaboration might've been two or three meetings across like a week. And then we were off and running, making these videos, building the webpage, Randy doing his thing. And, you know, he's a wizard with Descript. So he probably cut and edited this in like a day or two and he made, you know, three different ones for me, which was fantastic.

[00:34:56] Speaker 4: Exactly. That was actually going to bring me to my next point around like, if you're already using Descript to kind of speed up your process, now you have another tool to add under your belt around like the planning part before you actually get into creating part, because this is going to speed up your process so much more. We did have a question earlier in the chat that I wanted to bring up around how does the average person actually approach this or use this? Like, what does the setup look like? Or like, what should they think of as a like start from scratch?

[00:35:25] Speaker 5: Yeah, the tips I always, it's a great question. And the tips I always give is the more context you can give one of these chatbots, whether it's Claude or Gemini or chat GPT, whichever one you have access to or prefer, the more context you can give is the more value you're going to get out of it. That's the first thing. And so like, you know, I give it my cloud, I will give it access to a folder where I have as much documentation as possible. So it knows everything I know, right? So from the product spec to, you know, any kind of research and competitive intelligence or like website screenshots for that matter, or, you know, like links and documents full of reviews and that sort of content, I give it all of that as well as, you know, the direction so you can go search the web for more information to make it relevant. That's the first one. The second part is try and do that thing that I said in that learn and, you know, phase, which is have a conversation, try and have it explore all the angles, have it push back. And so usually, you know, some tips that work with any chatbot that you use is instead of giving it like, you know, some people like to talk about, give it a specific role. I just say, what would a few different experts say on this and have it kind of, you know, explore it from a few different angles, because what that does is it allows it not just to try and get you an answer quickly, but activate more parts of the model itself, more of that kind of corpus of intelligence that they have to explore things from different perspectives. I find that particularly valuable, especially when you're in that early figuring it out strategy phase.

[00:37:29] Speaker 4: And then as you are building this out, are there any things to keep in mind as you are continuing to work in this workflow?

[00:37:38] Speaker 5: Yeah. I mean, there's tons, but ultimately it's like, if you have a really valuable conversation that starts to meaningfully change the project direction, bring that back. Cause that's content, right. And you can continue to explore and build on it. The other thing I would say is, you know, I put like a little extra bonus bit in here where we can fast forward. Yeah. We have some information about like our target persona. So we, you know, when we think about marketing to or promoting Descript to marketers, we have a person we call Malik, right? And so this is like a description is a senior marketing manager at a midsize tech company, et cetera, et cetera. And so what I've done, I have like some information kind of saved in my docs here that give a lot of it's based on real research that we've done about folks in this kind of scenario. And so what I can do is actually ask it, give me some of that feedback. It's like an initial synthetic user testing as it were. Right. And allow it's another way that you can make use of AI to help you gut check your work. So, you know, I say, Hey, put the Malik hat on, give me some instant reactions, almost like it was a focus group. And so it gives me stuff to think about. And that usually helps later on as we're in the creative refinement phase.

[00:39:11] Speaker 4: Got it. Cool. We'd love to open this up to more questions. If anybody in the chat has further questions. I also had another question around, like, as you are having these conversations, do you keep a log of the prompts that you're using? So like for future campaigns, do you like reference those or anything of that sort?

[00:39:32] Speaker 5: Yeah, I do. There is a way to like, save those as like skills or like little things that I can copy and paste. But over time, what I find to be the most valuable is just to not just ask a question that says, help me make this plan. But actually you could see how the prompts I was pasting in were in this video were like whole paragraphs. That's part of the give it more context, like exploratory. It works for me where I just say everything I can think of. I even, you know, do the voice thing because sometimes typing I will edit myself down. Whereas if I just do like a dictation, I might say a bit more. I might make some connections as I'm thinking out loud. And so I find that that's as valuable than perfect prompts that I just keep using over and over.

[00:40:32] Speaker 4: That's a really good point. Yeah. I feel like, would you say it's almost kind of like a brain dump of a prompt to start with?

[00:40:40] Speaker 5: Exactly.

[00:40:41] Speaker 4: Are there things you kind of keep in mind as you are doing like that brain dump of like areas that you want to make sure you cover or so, like certain context areas, like focus areas?

[00:40:57] Speaker 2: Yeah.

[00:40:58] Speaker 5: For that, it's just, you just want to feel like you've really thought of, you know, taking all the information that you have at hand into account and then building from there. Another thing is you're never going to have perfect information, right? Like you don't know by the time you launch your campaign that a competitor is going to come out with something that's way better. You don't know there's things you just can't plan for. So what you're really trying to do is make the best bet that you can in the timeframe that you're launching it. And so that's what I like to keep in mind that I'm not going for the ultimate truth. I'm going for what's the way that we think would be compelling again, given our brand, given the capability of the feature, given the life cycle of where we're at with launching it, that capability. That's I think what's most important when you're a marketer. Nice.

[00:41:58] Speaker 4: Cool. So I'm seeing a lot of questions around video workflow. We do have a video workflow webinar coming up next week where we will dive deeper into how to actually use Descript to like bring your messages to life, essentially. So Matthew just showed us his workflow around like refining that message to make your video stronger. So if you're interested in joining that next week webinar, please feel free to sign up. But before we wrap up, Matthew, is there any, any other last words of wisdoms or tips or tricks or anything people should keep in mind as they're thinking of messaging and using AI? Yeah.

[00:42:32] Speaker 5: I mean, I, I think it can be really helpful provided you give as much context as you can. Always give it more context. If you think you've told it enough, just maybe tell it some extra stuff. Then I find that these working with AI can be helpful. You can get the value from removing some of that drudgery, that heavy duty, like I'm going to the subreddits, I'm going to G2, I'm looking at app store reviews, that type of stuff. You take a ton of time, like Googling for days on end to learn stuff. Cutting that drudgery out is one of the biggest unlocks that AI has brought in my workflow. And I mean, if you think about it, that's kind of what Underlord in Descript does for you as well, right? Like instead of all that timeline scrubbing, you know, finding the ums and ahs, like it can just do it for you. And that doesn't change like your imprint on what it is that you're trying, the story that you're trying to tell. And so, you know, hopefully you guys could see with that recreation of how I worked with it to come up with the plan for this launch, that you could really cut out some of that drudgery stuff just by leveraging the power of the tool.

[00:43:50] Speaker 4: Exactly. Awesome. Well, thank you so much for sharing this with us. Again, this is going to be on replay. So if you missed it or so, you can always go back and dive deeper into what Matthew showed us. Just to wrap up for us today, if you wanted to learn more about how to actually build out your videos or so we have a bunch of videos and articles to support that. We have a Learn Descript videos and we have the blog on our site. We also have office hours that our team can show you in real time, anything that you're struggling with. And we have weekly support as well here on YouTube on Tuesdays. And as I mentioned earlier, our next webinar next week is going to be about using AI to create the videos. And so we'd love your feedback as well. So please feel free to scan the QR code up in the corner. And if you're a part of a team and you're ready to start using with your whole team, essentially, feel free to scan that other QR card to get you into an enterprise trial as well. And I think that is it. So we'd love to see you all next time, but thank you so much for all for joining.

[00:45:03] Speaker 2: Thank you. Yeah.