How Marketing Teams Get 30% More Impact Using AI Agents with Tejas Manohar

From my perspective, what I see from having worked with thousands [music] of marketing teams at large scale companies is that they're all super slanted. Candidly, they're all just super busy, super overwhelmed. And the same is true for a lot of the partners and [music] consultants, agencies that they work with. I think there's also a lot of frankly inefficiencies in the program due to how busy and siloed everyone is, right? there's a person who's operating the Google channel or the Facebook channel that's super busy and again can't look for those cross channel opportunities. >> Hi, my name is Demetri Bonichi and I'm a content creator, agency owner and AI enthusiast. You're listening to the AI agents podcast brought to you by Jot Form and featuring our very own CEO and founder Idkin Tank. This is the show where artificial intelligence meets innovation, productivity and the tools

shaping the future of work. Enjoy the show. Hello and welcome back to another episode of the AI Agents podcast. In this episode, we have Tjas Manohar, the co-CEO of Hightouch, a really cool company that's building the AI platform for marketing teams. How you doing, TJ? >> Doing well. Thanks for having me on the show. Dimmitri, >> it's really good to to meet you. Um, really good to chat about stuff. I love marketing and I've been wondering how marketing and AI agents are kind of going to come together. So, what I'd love to know a little bit more about is uh to get everyone started is how did you get into AI, right? Pretty cool stuff. Obviously, everyone's talking AI nowadays, AI agents. How did you get into the industry in general? And um then tell us a little bit more about what uh Hightouch does.

Happy to. So, um I actually entered AI through the data industry. Uh so my background is in building highly scalable um data infrastructure and distributed systems. I used to work at a company in the Bay Area called Segment that was acquired by Twilio uh for a few billion dollars back in 2020. And I was like a you know backend engineer there before starting the company at Hightouch. >> Um and so yeah I don't come from a research background. one of my co-founders uh did did study machine learning in college and yeah I've absorbed some of the information from him and other people we've hired um but I really got fascinated by AI um really when uh when ChachiPT came out just like a lot of people and just realized like holy crap this is really cool and I think it's going to change uh a

lot of things but definitely the way that our customers which are marketing teams operate >> um and How have you specifically uh kind of taken, you know, marketing is a broad term, right? So, how have you kind of taken the world of marketing and um uh gotten AI agents to work in it? And what kinds of marketing would you say specifically do you think hightouch uh excels at the most? I'm just trying to get some broad strokes here and then we can go into the finer tune details. >> Yeah. So just to give everyone some context on Hightouch uh because we're you know relatively young company. We started the company a little bit over five years ago. Um I'll give some background. Hightouch is you know fast growing startup. We're about 300 people based here in San Francisco. Uh we serve companies like you know

PetSmart or the NBA um large airlines, hotel chains, uh companies like Domino's Pizza which we all know and what um we really started with was this idea that the way marketing teams try to use data today um when they want to personalize a campaign that's going out whether it's like an email or a push notification uh or a text message or or an ad that they're they're launching on Facebook or Google, it's just really cumbersome. Like a lot of marketers don't have the technical chops to, you know, use SQL or Python to like sift through all the data. Um, forget things like data science models, right? And so, um, but companies have a lot of data these days. And so I just noticed every company's putting a ton of data in the cloud in like Snowflake, data bricks, AWS, Google Cloud, etc. But that was

disconnected from the way marketing teams operated. And so we built a platform called hightouch uh which is really focused on sitting on top of the data warehouse and making all that data accessible to marketing teams. Um now fast forward a few years later chat GPT comes out and we realize that um not just when when it comes to bringing technology to marketing teams like BTOC digital marketing teams um why stop at data right there's going to be this new technology around generative AI that is going to change everything in how these marketing teams operate. Um, so that's when we decided to create uh what we now call high-touch agents or um uh an AI platform specifically for marketing teams. And you're right, marketing is very broad. It can be brochures, billboards. When I think about marketing, I'm talking about um digital marketing like ads, emails,

you know, banners on your website for large uh highscale BTOC companies. >> Okay. Yeah. So, um I I really like that. I actually have a background in um digital marketing to a to a certain extent. I was a paid media manager for a couple years before um getting into all this kind of stuff. So, it's very cool. Um I guess kind of speak to that group of people. I'm kind of curious what would be your like sales pitch for marketing agencies or sales pitch to even individual marketing managers and how it's going to change uh their lives. >> Yeah. So, this is the AI AI podcast after all, right? So, I'll dive into our our AI product specifically which is the >> Absolutely. Um so really what hight touch agents is uh is is two things. One we've taken the state-of-the-art um AI models um

so think about GPT anthropic Gemini and we've made um a platform for marketing teams to use with these models that's really purpose-built for marketers. So um you know when you ask it questions it knows about things like attribution. It knows how to grade ads. It knows the top two things when you are grading an ad is like efficiency and scale and stuff like that. two, um, we bring a bunch of unique context to these AI models that they wouldn't have if you just went to chat GPT, right? If you go to Chat GPT and you say, "I'm a marketer at Domino's. Give me some give me some like offer ideas I can run." It has no context of your customers, how often they're purchasing, what they're buying, whether they need discounts in the past, do different people respond to different things. But companies have a

lot of that one customer data and things like data warehouses. So naturally, since Hightouch is connected to the data warehouse of all of these companies, um, we bring that customer data as context to uh, in our AI platform first and foremost. So you can actually ask it questions about your customers and say, "Hey, who's likely to respond to a sitewide sale email that I send out?" Boom. Our AI knows that. The second piece of context that we bring to AI is um is really all your marketing today. So all the campaigns that you've run on every channel. So we connect to your email marketing tools, your tools that you're using for push notifications, your advertising platforms like Facebook and Google and pull down all the creative and campaigns and metrics and images and videos and ad spend numbers and all that kind of stuff. Um

so that you know when it's suggesting when AI is suggesting ideas if you ask it for that it actually knows what you've tried before and how they performed. And then lastly, there's a lot of private knowledge that marketing teams have right when they work at a company. Brand guideline documents like strategy documents, um the campaign goals for the quarter and that's the third piece of context that you we really bring in to our AI platform. And so once you have these things when hight touch agents, you have a platform where marketers can go in and they can actually get a much better experience than chat GPT when using it for marketing tasks, right? they can use >> um as a baseline. They can ask it for feedback on a creative and it actually starts looking at similar ads performance. They can ask it for um

ideas on a campaign and it thinks about what you've done in the past, your brand guidelines, what works for you, your customer data, etc. But the next thing that we're doing is actually building agentic workflows and sets of um little agents that can help marketing teams um do specific tasks. So examples are like ad reporting, looking for creative fatigue, which just means, you know, your ads been in the market before, it worked for a while, it's not really working as well anymore, reach too much of the target audience. >> Um uh and then similar on the life cycle marketing and email side, as well as um you know, engines to do things like generate emails like code HTML ready emails and all that kind of stuff. Um so yeah, first and foremost is this foundational platform, but then we're going to every discipline of the

marketing team and building agents specifically for them um that are customizable for each and every company as well. >> What does that mean by customization? Talk a little bit more about that. Yeah, I [snorts] would say the other big difference between when I look at the SMB and like mid-market AI opportunities and the enterprise ones is um there's a lot of companies out there just saying like we'll be your ad agency using an AI tool or we'll help you generate like creative run it on Facebook tested create more etc. If I was running like a super small business, I would probably be pretty interested in that or absolutely be interested in like features from Facebook that use generative AI and >> you know, generate the ads for me. But at an enterprise scale, >> um, you know, you look at all the tasks that go

into launching a marketing campaign, there's like 50 things. >> They use an agency for this, an agency for that, one team does this, one team does that. And people >> I've seen the QA lists alone. Those will want to make those will kill you. >> Yeah. Those alone can have 50 things. And these um tasks, whether it's like legal legal, you know, uh approval information or brand guidelines or um the strategy document that we write that has to get signed off by stakeholders, the experiment design of any marketing campaign, making sure it meets a certain opportunity side. Like all of these things um are very custom to a company. And so you can't really just go to an enterprise and be like, "We're going to be your agency and just do all your marketing." Um, and so basically customizability is super important. Like we can

build a base agent that can do things like build a weekly ad report um report on all your ads and the performance, but then you might want to tweak that. Or maybe you're comp you're a company like um I'll just use a hypothetical example like Home Depot or 7-Eleven that also runs an advertising agency for brands that you work with to advertise on your website and you want to give give the brands that are advertising on your website reports on how their ads are performing. Right? There's all these nuances that are specific to each business and so the customizability element is really important. >> You know I you mentioned something earlier about reporting. Okay. Uh I think reporting is is was something that I remember that was really really big deal. Um prior uh I really liked the I guess um manner in which uh

when I when I had a good report that I was putting putting out to clients. It really really felt like it it made the experience better. You mentioned reporting there. Um, I did feel like honestly though a lot of the dashboards that you could make in like something like Google Data Studio or I used Datarama for a while, which I would not recommend anyone use. Um, but uh I'm curious kind of from your perspective, what is the state of uh marketing analytics in general with paid ads and and how do you kind of help out with that and reporting? because reporting used to be the worst time suck that I felt like was practically just I don't know it felt like a waste of time to be honest. Clients didn't read those things but um tell me tell me a little bit more that that

>> yeah [snorts] today a lot of reporting when it comes to marketing analytics looks like a huge slideshow of like five tables bunch of data with a bunch of ad names on it that don't make any sense. There's just tons of underscores and um it's a mess, right? It expects that looking at a screen of data gives you insights that you need. And so we've had BI tools and analytics tools forever now, right? Tableau, PowerBI, Looker, whatever it is, but they're all focused on presenting data when ultimately marketers or business people, they don't want data. They want insights and they want actual suggestions of what to do. And so um it wasn't my first intuition when chatpt came out a few years ago but later after the reasoning models uh came out and after spending more time with the solution um I I realized that

like AI was going to fundamentally change all of data analytics because we finally have a technology that can actually interpret the data versus just present it. And so what we're seeing is that today um you know a marketer might on a Monday um spend hours and hours and hours just going to different systems pulling down tables of data putting it into different Google doc templates for their to report to their team to different audiences or different spreadsheets and that might just take the day and then then they get to start answering questions like why did this metric go down why did this metric go up but a lot of times it's like seasonality or it's noise or >> oh boy, >> you know, it's going to go back next week or other times there's the the 10% of time or 5% of the time there's

actually an insight there which is like you need to make new creative, your creative is not working anymore. Um or you need your competitors started bidding on this uh same keyword you're bidding on on Google and it's getting more expensive and you should move your budget. But 95% of the time there's a lot of noise and just a lot of data to stare at. And so now what we're seeing companies do with hightouch agents is actually um create a custom agent in our app that does their reporting in the way they want and it can actually um we've taught the agents to know how to interpret ad metrics, how to tell if like patterns are, you know, noise or signal and what to check. For example, check year-over-year trends, check current week versus four-week average. um like look for at least this percent delta like

stuff like that. And so now instead of actually digging through um just a bunch of data and everyone in the room in a marketing planning meeting or weekly business review just acts like I get this, I get this, but you know, no one's really able to process all the data on the screen. Now they can actually read actionable insights of like what they should actually care about and spend more time thinking about next steps and and true interpretation. You know, I think it's it's funny to me like the I I love how you mentioned I mean, you feel like you take like a day. I remember I had back in the day literally Mondays were reporting days and Tuesdays to some extent and it was it was so stupid like in retrospect it was so it was so silly. Um I I I really think

that this >> done instead what would you have done instead? >> Oh dude. Okay, I'm getting PTSD. I there's a reason I don't work in that industry anymore. Um, but no, so the uh what I would do, right, so I would go into data or whatever it was, make sure that there was a week over week or bi-weekly or month over month kind of comparison mattered on the day of the week and I would literally go through and essentially kind of templated check out, okay, what was the performance one week versus the difference? What was the comparison change? try to convolutely say, "Oh, well, there's a conversion uptick here. There's a conversion rate uptick there. Why is that happening?" And I'd use a method that was like fact understanding and action or something to that effect where I would try to write out what was

occurring. And a lot of it, frankly, was pattern recognition and conjecture like in order to get through it sometimes because you're in busy season. And it's like, okay, I know the client's not going to read this stuff. Um, unless it's like really insightful. And the problem is, I feel like you end up getting put into this weird cycle that it isn't insightful. And if they do read it, they end up asking you a very specific question, and then you have to go into that specific question. Um, and spend a lot of your day on it. And I do have one interesting uh comment here and question. What would you say about the ability for a client or the the marketing manager to answer the questions of clients about reporting? Like, do you guys have the ability to like like an LLM would like access the

data and and get the answers to the questions in a quicker manner? Because I feel like that would have been such a hack for me. >> Exactly. I mean, that is the whole platform. So imagine it as catch GPT >> that knows how to interpret marketing. So that means reporting, marketing analytics, attribution, customer segmentation, all that kind of stuff with the full context of everything you need. So all the campaign data, no more downloading CSVs from all these platforms, your data warehouse, which has your purchase data and transactions and everything, your brand information, your strategy docs, your quarterly goals. And now you don't have to spend all the time pulling the data down. You can also get a first pass of interpretation and do all those cuts of the data that you had to do manually in data before and then actually start to get

suggestions of what you might want to do next. Yeah, that might not be exactly what you do. Um, but it's a great starting point and it seems one day per per five day week of a typical B2C marketer >> 2020 one 2022 2023 Dimmitri is kicking himself right now. He's like what could I have just graduated a little later? This is ridiculous. What is Okay, so I love that for you. That's awesome. And how how do you kind of articulate to um I mean it's very obvious to me but how how do you kind of go about um spreading the word on this and saying to uh clients like or potential clients like hey this is a big bottleneck. this is a problem, right? Like what you guys are doing and and how do you articulate the value you provide and um I guess what

would you be your go-to way to frame like this this is the new way to do things because I I know personally speaking from experience again a lot of companies can be set in their ways uh in any industry but in that industry specifically I felt like there was always like interest in internal innovation more than interest in like taking on vendors and stuff so speak to that a little bit. Yeah, I think a lot of companies are very scarred from marketing technology solutions that make big promises and don't deliver. And I think we've seen this in this the customer data space and personalization space, which is where our core product is. Um, we've seen it from, you know, different, you know, email tools you might buy that say they're going to make your whole email program better and stuff like that. But at the

end of the day, like you've got to tell it exactly what to do. Um, so it's you that's making your email program better. We've seen it from marketing analytics vendors um that again pre AAI it wasn't really possible to have solutions that truly interpret the data properly. Um and so that's why there is a lot of churn from average you know marketing technology tools. People switch it around all the time. There's a lot of failed promises. We're really happy at high dish that we have, you know, exceptional retention metrics. um you know like 90 mid 90 plus percent of our customers renew every year and um that is something that's like super super important to us. And so the the main thing that we've found is just because of all the scars um companies have from trying tools that didn't really deliver, you've just got

to make it easy to try. >> And so yeah, >> sorry I had a followup based on that. I apologize. It kind of fits into the the line of questioning. Um I would say my main concern if I was a company actually wouldn't even be that honestly I think the reporting makes the most sense, right? It's it's it's something that I could tangly how do you really >> um across all of the different components of something >> make creative consistently iterative? Right? So, we have the copy um we have the um actual like technical media assets. Uh and then his follow-up question is when it comes to media assets, like what kind of ones do you make? Do you make like general images, HTML 5s? Like what's kind of the level of uh capability there? It's just kind of curious. >> Yeah. So right now

um if you think about all the steps in a marketing process um like end to end that a marketing team does there's like research planning campaigns actually writing the briefs creating the assets getting approval um you know designing experiments launching the everything measuring the results etc. We excel in my opinion at at the areas other than content creation. And so um you know we connect to all the data sources across the company. We help you plan and strategize what's the next content you should create. Um we analyze your customer data and suggest campaigns that you can run that would be really high ROI. We can look at ad performance and see what's what's working, what's not and help you create more winning ads. Um and again across every channel on the digital marketing site. So, not just ads, but emails, push notifications, website experiences, um

creating content or you know, we're working on uh we're building a content creation um kind of studio with AI, starting with some of the the simpler um assets like emails and really taking an approach of um having AI connected to all the repositories of existing content you have. So whether that's scraping all the stuff from your website, looking at all the past emails, looking at Dropbox folders of images that you give it and help giving AI the ability to reuse those assets and make slight modifications is really the approach we're taking overall. Um but yeah, you know, what we've realized is that there's so much other work that happens in marketing teams outside of just the creation of the asset. um like you know 90% of the time of marketers is spent in meetings and planning and Google docs and spreadsheets um writing briefs to

send to agencies that create the assets and there's a very small team that actually focuses on the asset creation and so our goal over time is to build you know cursor for the entire marketing process um but uh we really excel in all those like data and strategy kind of elements right now. >> Okay. Well, that's still very valuable, too. I think, you know, the it's there's a lot of different aspects and and uh I appreciate you kind of laying that that out for everybody because honestly, um I don't I don't want to miss any opportunities here of highlighting what you guys do well. So, another thing that I noticed that I think is always interesting um in regards to audience uh or marketing is audience, right? audience segmentation, audience uh targeting, right? Um, how does your AI really help in regards to segmentation? Because

I I know with with strategy, it kind of comes into a position sometimes where you're like, "Oh, I think I should target um this group of people, whether it be by age, gender, um buyer intent, all these different things." What are the different ways that you can break things down? And how does your uh AI help analyze the different metrics that exist in current campaigns or uh help you start from scratch, right? Because I think this is the biggest like blackbox portion of what to do with uh marketing. >> Yeah. Um what I see marketers having to do today is a lot of context switching, right? They go from one task of promoting uh you know their existing product line and then next week they have they're in charge of promoting a new product line that's being launched. The week after they're trying to promote

a credit card you know and it's context switching and context switching context switching and each time they go to a new context um you know there's there's a lot of thought that has to happen like it's not just what's the marketing message but who should I reach out to and that's the audience element that you brought up. Um so with the audiencing um we actually connect to all the customer data that companies have. And so if you're a retailer and you want to uh ask the AI a question like suggest me audiences for um promoting uh my you know lipstick product as an example. It'll actually go through look through your data warehouse um both build a bunch of it's it's given that we have LLMs now it actually like can do um like research and think about like what's semantically associated with lipstick maybe

it's people who are into more fashion products stuff like that build and test those hypotheses against your actual data set in your data warehouse so generate SQL queries etc against your data warehouse also run predictive models so one thing we built um as tools for Um, our AI platform is like actually the ability to run predictive models against the data in your data warehouse and see what are the characteristics in your data warehouse that indicate that a customer might be interested in a lipstick product and does that vary for different age ranges or past products people have bought etc etc etc. So um basically when you ask it a question like suggesting audiences that's one workflow that we've really designed the product around since our core product is all about personalization and it'll actually do similar work that analytics teams and data science teams do

to figure out what might be the right audience to uh run a campaign against to promote a new product for example. And then what's really cool is because it's interactive um and it's an AI solution that's always on versus an analytics team that you might only hear back from it a week later um because they're so busy with other tasks. You can actually just chat back with it to ask follow-up questions as well. So, for example, you could you could be like, "Hey, and tell me if I send an email blast to these customers, what what percentage of them will actually uh open it or or click a link in it?" or if I want to reach at least 100 thousand, how many times should I send it and are there certain days that are best, etc. And so you can start with sort of

interactive chat and just ask it for things like audience suggestions. But we've also built um a whole sort of module that's around planning out a campaign and and and helping write the strategy brief for a campaign that does audiencing as a part of it as well. >> Okay, interesting. Yeah, that gives me a pretty good explanation. And um you know speaking forward with that, obviously there's a lot of different types of customers that you work with, a lot of different industries that you work with. What would you say is some of the favorite um industries that you you guys work with? Obviously retail is always big with digital marketing. I see that on your website. Um what would you say is the the main ones that you're really excited about and some cool stories you have to share there? >> Yeah, it's really the the

top um BTOC direct to consumer industries. So um you know we're quite fragmented actually. We we we do a lot in retail financial services. Um a lot of the NEO financial service apps that you all probably use are are customers of Hightouch. Um uh also media companies, so you know streaming providers where you can watch TV and stuff like that. Um gas stations, you kind of name it. It's all over the place. hotel chains, [clears throat] airlines, >> and so overall, um, it's really any company that has a direct contact with their customer and is trying to use digital marketing as like a really big channel to drive them to to uh interact with their business. Um yeah, and >> in terms of customer stories, I think what's really interesting is uh our work with um with a financial services company in the US, pretty

large financial services company. And basically prior to to using our AI solution, you know, they were uh always bottlenecked on their data team as an example on the life cycle marketing and growth side where they're planning different campaigns their customers base. they were always bottlenecked on their data team to be able to answer questions around how many customers will actually respond to an email in this demographic or what customers would be most likely to buy my premium subscription uh or like look at all the marketing that I'm running to drive people to the premium description subscription which customer types are actually not engaging with that and what tactics in my the rest of my marketing do they engage with they just like tried a lot of data driven stuff to drive conversions because you think about financial services like any conversion to a new offering

is just worth so much, right? The LTB is so high and so um they're pretty bottlenecked on like data science and data analytics resources to answer all these questions and now like the marketers are actually logging in um like pretty much every day or every other day and to using the tool to strategize the next campaigns and experiments that they run. Um the other example that I think is really compelling is is that the advertising team there for example is trying to reduce and reduce their reliance on agencies. Um so previously they use an agency for a lot of their media management. Um, but now they're moving that in-house. And as they move it in-house, the typical model is that you have a person working on every single channel. And then like Google, someone dedicated to Google, someone dedicated to Facebook, someone dedicated to search

ads for a particular business unit or product that you're selling. Um, and that person's like always looking at the ad account and trying to figure out what are the trends, like which kind of creative art are working better. um where should I put more budget especially when you're operating on five 10 different marketing channels and social media kind of providers and now what we're seeing is uh two things one those people are actually able to spend more time like actually you know writing briefs coming up with the next creative etc run instead of analyzing the account every day because AI will just tell them when there's actually changes they need to care about and then two um there there was a whole world of cross channel learnings that were kind of being lost before of like yeah this thing in Facebook's actually really working and

you're not doing it in Google um or or in other word uh hey there's actually something in Google that really works that doesn't work in Facebook you've tried it a bunch of times so next time you're trying to advertise this product do it in Google instead um as a as a first suggestion of where you put budget and so it's it's both efficiency of just like less hours to do the same thing as well as finding more opportunities that you wouldn't have been able to otherwise. >> Nice. Um, you know, I got to say it's very it's very impactful uh what it seems like you guys are doing to the point where it's like, okay, I I have tangibly heard so many different ways that you're saving marketers time. Um, obviously I think from a standpoint of how it probably works on your in your

pricing, you're saving even money in some sense. And just to kind of touch on marketing in general, the industry in general, um, how do you feel like these types of products like yourself are going to start impacting um, jobs in the marketing industry, right? Uh, obviously, you know, in a positive standpoint, you're probably saving uh, I would say companies like all this precious like valuable time. Um there's a couple different camps people fall into like, oh, this is going to make it so that we're going to have less jobs. Uh that's one camp. Another camp is um I feel like this is going to make people's jobs more in depth and less uh nonsense work. Um give me your take on on where you think not only your company but other companies in the industry and agents are going to kind of impact things. Yeah,

from my perspective, what I see from having worked with thousands of, you know, marketing teams at large scale companies now or and at least talked to them, uh, is that, um, they're all super slammed. Like, candidly, they're all just super busy. >> Absolutely. They are super busy. Yeah. >> Super overwhelmed. Um, and the same is true for a lot of the partners and consultants and agencies that they work with. Um and and I think there's also a lot of frankly inefficiencies in the program due to how busy and siloed everyone is, right? There's a a a person who's operating the Google channel or the Facebook channel that's super busy and again can't look for those cross channel opportunities. um or is so busy looking at all these ads that they missed a really stable ad that's starting to not perform that if you made a

creative a new creative for faster uh would would help you hit your next quarter's number. And so honestly, I would say everyone's super busy from what I can tell in marketing teams at large companies. >> You saw my you saw my reaction. I did not like what was going on. >> Fair enough. And everyone is like there's a lot of opportunity and a lot of both wastage due to not having like the data at your fingertips and the insights that you need and just missed opportunities. And so yes, I believe like some of the more monotonous tasks will be removed from marketer plates, right? Like I think AI is going to be taking a first pass of reporting. If in a year from now or even sooner in my opinion that your your marketing team is like not using AI for reporting on metrics, it's

just kind of crazy, right? That's the first place we start because it's easy to prove and it works really well. And but at the same time, I actually think there's opportunity to not just take that work off the plate, but actually make the marketing programs like frankly 30% better at an at an average large enterprise that's spending wow >> tens of millions, hundreds of millions of dollars uh on advertising. And I'm not, you know, I'm also thinking about their all the rest of the marketing programs that they're doing on the digital side. And so honestly um I don't think there's going to be a huge labor market change in marketing soon. I think the teams will actually just start outputting a lot more um and spending less time on monotonous work. Um the area that I think is TBD is like creative as the creative

models get better and better and better and especially the agent the a creative agency side of the equation. Um but at the same time I think there's a lot of work to go on the creative models and case is again the hardest thing for AI to replicate and so um I could you know maybe the variations maybe stuff like that but I don't think people are going to like completely you know remove a creative team anytime soon frankly. So >> yeah. Okay. No, that's fair. I mean I mean it's it's a it's a really hard kind of like place to balance being like ah we're going to save you a lot of this time and money and it's like you know and where what are the implica everyone kind of tries to make their own logical entailments from that but >> I think the way

I think about it is like large companies want to drive more topline like >> um they want to drive more revenue and uh I see huge opportunities to actually do that um with the same budgets they have today or um or to to make the whole program more more efficient on the actual spend side. Like when you look at any marketing organization, it's 80% spend and 20% people um frankly like program spend right and so um in the BTOC side and so I think that I think that the opportunity for for companies is really to improve the efficiency of their spend and the impact from it. >> Yeah. No, absolutely. Um, are there any other kind of uh things or well actually you know the one last thing I like to ask always um as we kind of wrap things up is obviously you have

your own product really cool you got your opinions what what's going on in in general obvious what it came out what is your favorite personal AI tool that uh you're using just for yourself right um whether it be in work or in business that you're like yeah this saves me tons of hours love it it's the best >> the the best uh AI. I guess the the more um hidden hack I guess I mean unless you're a pro that I've learned um recently from some of our product and engineering team here is using cursor for general AI tasks. >> Keep hear this. I keep hearing this. And what do you do personally with it on a >> daily? Everyone of course like everyone uses cursor if they're if they're programming or if they're coding. Unfortunately, I'm not I'm not doing that all the time anymore.

Um, used to be my job. Would love to have cursor back then. Uh, but not doing that all the time anymore. Um, as the business has been scaling and but but cursor as well as cloud code actually have really good agentic planning. Um, so what that means is you a big task, right? Um, this is very important if you're using it to to to code whole features and stuff like that. and put in a big task and it'll it'll break it down. It'll manage a to-do list um as an agent, right? It'll actually save a to-do list, a task it needs to do to the file system and keep checking its progress and doing them. And so it just doesn't lose track of things. And so you can use it for really big tasks like look through, you know, deep research type tasks, right? like

look through all of these uh list of companies I have and find all the rules responsible for X Y and Z and caveat it to me when it doesn't feel like they may actually own that but it's the closest title you could find and >> etc etc etc etc >> um and that's the same way that they do the context management to be able to generate like large pieces of software right in the agent mode of cursor or club code and so I think it's a very powerful hack you just when you have complex tasks you want to delegate to an AI the gen one you know LLM um way of thinking about this was like let's go just always build an agent right and chain prompts together with a tool like Zapier or NAD or something like that. But now I'm and that while

that still makes a lot of sense obviously what I'm finding is like when you just have an interactive complex use case you don't need automation you can just put it into a software like cloud code or cursor that has really strong agentic planning and it'll break down the tasks and figure it out. Um and so um it's actually one of the reasons that uh we uh have actually like played with using claude code as the backend of our hightouch agents product um and replicating also some of the agent loop and stuff it does for other models as well. So I think it's >> wow that that that's you know I think that's really good insight. I I've heard such great things about um I've used cloud code for some mass um marketing planning before. um it's it's something people really need to look into. So

great um great commentary there. And uh last but not least, I would just say to everyone listening, thank you so much for listening and watching, but what you need to do right now is go to highttouch.com if you're looking for uh something to help out your marketing needs and prove it with their amazing AI tools. So that being said, uh thank you so much uh uh TIS to you for being on the show. Uh we appreciate it. Make sure to leave a like, comment, um even you Titos or Tatos if you want to um get more uh listens on the on the show. Uh leave us a review on Apple and Spotify. Um the old five stars will do just fine. And we'll see you in the next one. Best. >> Thank you so much.