← The Agile Brand with Greg Kihlström®: Expert Mode Marketing Technology, AI, & CX
The Agile Brand with Greg Kihlström®: Expert Mode Marketing Technology, AI, & CX
#748: Markup AI CEO Matt Blumberg on creating a winning AI-supported content strategy
00:23:47
Loading summary
Transcript51 lines
- [00:02]
Matt Blumberg
And Doug, there's nowhere I wouldn't go to help someone customize and save on car insurance with Liberty Mutual. Even if it means sitting front row at a comedy show.
- [00:12]
Greg Kilstrom
Hey everyone, check out this guy and his bird. What is this, your first date?
- [00:15]
Matt Blumberg
Oh no. We help people customize and save on car insurance with Liberty Mutual. Together. We're married. Me to a human, him to a bird. Yeah, the bird looks out of your league. Anyways, get a quote@libertymutual.com or with your local agent.
- [00:27]
Greg Kilstrom
Liberty Liberty Liberty Liberty. At Strayer University, we help students like you Go from Will I to why not? For over 130 years, we've been innovating higher education to make it more affordable, accessible and attainable so you can reach your goals. Go from thinking can I?
- [00:47]
Matt Blumberg
To Yes, I Can and keep striving.
- [00:50]
Greg Kilstrom
Visit strayer.edu to learn more.
- [00:52]
Matt Blumberg
Strayer University is certified to operate in
- [00:54]
Greg Kilstrom
Virginia by Chev and its many campuses, including at 212115 Street north in Arlington, Virginia. Welcome to Season seven of the Agile Brand, where we discuss the trends and topics marketing leaders need to know. Stay curious, stay agile and join the top enterprise brands and martech platforms as as we explore marketing technology, AI, E commerce, and whatever's next for the Omnichannel customer experience. Together, we'll discover what it takes to create an agile brand built for today and tomorrow and built for customers, employees and continued business growth. I'm your host Greg Kilstrom, advising Fortune 1000 brands on martech, AI and marketing operations. To make sure you always get the latest episodes, please hit subscribe on the app you listen to podcasts on and leave us a rating so others can find us as well. Now onto the show. How can large organizations harness the power of AI for content creation while simultaneously mitigating the risks that come with exponentially scaling content volume and velocity? Agility requires adaptation to new ways of working, as well as adapting governance and oversight processes to maintain brand integrity and manage risk in the new paradigm that AI has helped to create. It also requires the ability to quickly react and adjust content strategies in response to real time performance data and market shifts. This episode is brought to you by Markup AI. AI can create content, but Markup AI's content guardian agents perfect your content. Markup AI instantly scans, scores and rewrites any content to enforce your standards for brand voice, terminology and compliance, giving you the confidence you need to scale your use of AI. Learn more at www.superup.AI. today we're going to talk about navigating the Exciting, but often treacherous landscape of AI driven content creation. The sheer volume of content being generated presents unprecedented challenges for maintaining quality, consistency and brand integrity. To help me discuss this topic, I'd like to welcome Matt Blumberg, CEO at Markup AI. Matt, welcome to the show.
- [03:10]
Matt Blumberg
Thanks for having me, Greg. I'm a big fan of the show. I'm excited to be here. I love that you're a thought leader, but you have your roots as an entrepreneur yourself, so you know what it's like to actually build a brand, not just to tell people how to build brands.
- [03:22]
Greg Kilstrom
Yeah, yeah, I've seen it. Seen it from a few. A few angles, let's just say. But yeah, no, I appreciate that. And yeah, before we dive in, lots of. Lots of great things to talk about today, but why don't you give a little background on yourself and your role at Markup AI?
- [03:36]
Matt Blumberg
Yeah, absolutely. So I've been a tech entrepreneur for over 30 years. This is my fourth company. I started in the, like, early, early.com years, like 1995, building a website called moviefone.com which was inside of an existing public company, if anyone remembers the old 777 film moviefone. So I built moviefone.com from 95 to 99. We sold the company to AOL in 99. People who know me probably know me best for what I did after that, which is I started a company called return path in 99 and grew it for 20 years. It was a global enterprise SaaS company and email marketing. Started from scratch, built it to over 100 million in revenue, sold it in 2019. Then I did a startup in the recruiting space called Bolster, which was less successful but no less fun. And then I started at this job, CEO of what is now called Markup AI. We just rebranded the company in January. And along the way I've written a few books for entrepreneurs. A blog called startupceo.com, as I just mentioned to you, a done a couple podcasts as well. So my journey has always been centered around helping businesses communicate more effectively, build stronger brands, and was very excited to join Markup AI to lead its transformation. And that's been my. My job and my goal is to help companies really tackle one of the biggest challenges and opportunities that exists in 2025, which is ensuring that your use of AI can scale as rapidly as AI can handle, but that you as a company are doing it smartly and safely when it comes to your brand and the content that comes out of AI.
- [05:19]
Greg Kilstrom
That's great. Yeah. And that's really where I'd love to start, is really aligned with that and the strategy part. So again, I'm sure most, if not everyone listening to this has used AI to create some kind of content or answer some kind of question, in essence creating content. But the trick is doing it at scale in a strategic way. And so, you know, markup AI talks about content guardian agents. Let's start there. And can you elaborate a little bit on this concept and you know, how does it differ from traditional content review processes?
- [05:54]
Matt Blumberg
Yeah, so AI has real generative AI has really kind of flipped the script on content creation, where content creation used to be about armies of writers producing lower quantities of really high quality content. And what AI enables you to do is to produce, you know, mass volumes of content, you know, click, click, click, with fewer people, but inherently with lower quality. Right. Large language models are predictive by nature. They're not perfect, they're a little bit generic. All of us who've used them know like, oh, the answer comes out the same all the time. They all sound the same. So the concept of the Guardian agent is something that has been emerging as sort of on the front lines of the AI landscape right now. And the concept of the Guardian agent in general is an AI system that is purpose built to monitor and oversee other AI systems. So AI systems are great, but they can't correct themselves. Right. They know what they know, they predict what they predict. Any of us who've ever used ChatGPT know when it's wrong and you tell it it's wrong, what does it say? You're absolutely right. So they're very pliable. So they do good work, they don't do great work and they don't self correct. So a Guardian agent is an AI system that corrects an AI system. And they're going to be increasingly important in lots of realms, lots of walks of life as AI gets more pervasive. So there are going to be lots of kinds of Guardian agents. Security is an area where there's going to be a big one. And what markup AI does is we provide content guardian agents. So we have a purpose built AI system so that companies can help oversee their content creation process, whether it's content created by a human or by AI or both. And basically our engine looks over your content as you're producing it. And the first Guardian agent that we've launched is called the brand Guardian Agent. And what it looks for is spelling and grammar as kind of table stakes. But then it looks at your brand voice, it looks at Your style, it looks at your terminology, it looks at a bunch of deterministic criteria that you upload to our system when you create an account. And we basically use those things to put a wrapper or guardrails around the, the AI output so you, as an author will experience our product integrated into the systems you use to author content. So whether it's Word or docs or PowerPoint or slides or Figma or Adobe or Contentful or whatever, will be a writer's companion that will scan, it will score and benchmark, and then it'll let you one click rewrite a whole block of content to make sure that it is on your brand. It uses your terminology. Everything's spelled correctly. You know, the weird terms you have, the weird, you know, things that your product people say, oh, if it's in this context, use it this way. If it's in that context, use. All those rules are programmed in as like a wrapper around your AI output.
- [08:50]
Greg Kilstrom
Yeah.
- [08:50]
Matt Blumberg
And the, the other great thing we do is in addition to being an authoring assistant, we can also scan in batch mode, large repositories of content. So we have clients that have like 10 million pages of active content online. And what do you do when you retire a product family and you re, or you rename a product family or you acquire a company or your brand? People decide, like, hey, you know, our tone needs to be more inclusive now. Like, how do you fix 10 million pages? The answer is you don't. Unless you're using a system like ours where we can scan, score and rewrite 10 million pages at an instant for you.
- [09:30]
Greg Kilstrom
Well, and that's. It brings up the context of speed and scale. Right. You know, because any, anyone, you know, they can probably look at a single document and make it, you know, give it the brand voice and do all that stuff, you know, one, one thing at a time. But whether it's, you know, your example of, okay, we inherited 10,000 pages from this company that we acquired, and now we got to make it all in one brand voice, or the rate of creating new content. Right. You know, the speed that's required often to do that, you know, Omni Channel essentially these days. So, you know, I know you touched on this a bit, but, you know, how do you not only enable things like that, but make that sustainable? Because that's, it's not going to stop. You know, a company that acquires one other brand or company, they're not going to necessarily stop, and they're also not going to stop going to market across channels. So like what's the, the approach for sustainable content strategy?
- [10:23]
Matt Blumberg
Yeah, I mean that's, that's the combination of a large language model and a guardian agent is the, is the approach. Right. The, the tools and techniques are now infinitely scalable, they're really affordable. And you know, the price of a token is very low and the price of a guardian agent is based on the price of a token. So it's, and it's tokens that companies like us get wholesale. So you're paying even less than you would, you would pay if you were, you know, buying it directly from Anthropic or from, from OpenAI. And the sustainability comes from integration with your workflow. So yeah, I think everyone, everyone knows like it's a little hard when you have like multiple systems that are like sitting outside of each other figuring out how to stitch them together. Our guardian agents are accessible by API or through kind of low code, no code integrations like Zapier or N8N. So we exist wherever you want us to exist. And in that way it becomes a seamless part of what we call the content supply chain.
- [11:28]
Greg Kilstrom
And so I know you mentioned a few platforms when you're giving an example, but could you maybe walk us through what's an example of maybe how some of these integrations might work together through whether it's APIs appear, something like that, you know, what's maybe an end to end workflow that somebody could imagine?
- [11:46]
Matt Blumberg
Yeah, I mean, so there's, there's no set end to end workflow that any company has. Right. Most companies quite frankly have content coming out of like 82 different places and system. But I can give you a couple of examples. So we are, we have an integration that will exist essentially as like a sidebar in something like Word or Docs or Contentful or Adobe. So we're literally in the application itself. It all works off of the same API key. So when your company creates an account, or if you're a developer, you create an account, you get an API key, you plug that into any of those applications, or there's an application in that company's app store that you put the API key into and it's all there. And what you've done when you set up your account is you attach some things in our admin console to your API key, like your brand book, like your terminology dictionary, and then those are things that we give you the opportunity to manipulate sort of off to the side. But the same API key works with all of them. So in an authoring application will appear as A sidebar. And for things like batch checking, you know, we could also be a sidebar attached to a content repository. But because we're API first and we have an MCP server, the results can show up in lots of different places. So we have clients that have built the API into their own internal, you know, homegrown IT applications. And when the, when the API returns a JSON payload, they, they format it and they do something with it. So it can be sort of a bespoke thing. You know, we show up as a GitHub Action for People who are committing code and therefore committing pages, et cetera. So, you know, by being API first, we're sort of anywhere that clients want to work us, however they want to work us. But I'll give you another thing that'll kind of bring it to life. One of the things we do is we have a consistent scoring methodology for your content. Again, looking at your standards. Spelling and grammar aren't about your standards, but your terminology, your tone, et cetera. So we have a lot of clients that use the score and they'll say, oh, you know, if the score, if it scores over 85, you can automatically publish it or automatically advance it to the next stage. If it scores below whatever number you plug in, kick it to a human. Right. Human in the loop.
- [13:50]
Greg Kilstrom
Yeah.
- [13:51]
Matt Blumberg
So you can imagine how useful that is for batch content. But even for things like real time content where, you know, clients will plug us in in front of a, a chatbot wanting to make sure that, you know, they're the, you know, they've reduced the humans that are involved in customer communication, but they want to make sure chatbots don't send out stupid things either. And, you know, the ability to escalate to a human in real time or publish something through a chatbot in real time, super, super useful. So there are kind of a few ways you can imagine us fitting into a workflow.
- [14:21]
Greg Kilstrom
Yeah, yeah, absolutely. And, and the fact that it fits within tools that people are already in gets, Gets away from. Because, I mean, I think the, I know there's been a lot of enthusiasm around, like, hey, let me try this thing. But then, you know, you end up either copying and pasting results from one thing to another, or there's just disconnects. Right?
- [14:41]
Matt Blumberg
Yeah.
- [14:41]
Greg Kilstrom
I wonder, you know, how granular does this get? You know, because I, I'm also thinking in my mind, okay, you know, you've got a, you've got a large company, but different products may have different tones, or you may be B2B and B2C or you know, things like that. So like, you know, how specific can you get?
- [14:59]
Matt Blumberg
Yeah, you can get pretty specific with it. So you know, an enterprise license, you can upload an unlimited number of style guides. And, and a lot of our, we have some very large enterprises that we work with, some of the biggest ones in the world who in fact do have multiple style guides for multiple business units. You know, the same company will have like an office suite and a gaming platform and they don't speak to their customers the same way.
- [15:20]
Greg Kilstrom
Right.
- [15:21]
Matt Blumberg
So yeah, so the platform gives you a lot of granular control over which, which rule set you're applying to which pieces of content.
- [15:28]
Greg Kilstrom
Yeah, great. And so certainly the speed and scale has major benefits as well as just unified brand. There's all kinds of stats out there that point to that consistency being valuable. How do you recommend that organizations measure and not only measure roi, but just effectiveness over time of this, know, type of AI driven content strategy?
- [15:56]
Matt Blumberg
Yeah, I mean ROI is, is obviously, you know, the thing that makes the world go around in, in business. But, but I would sort of talk about, you know, kind of a couple of different things. So there's sort of classic ROI measurement. So what operational efficiency do you get from something? How many writers, lawyers, editors are you replacing or how much more productive are those people? How many hours does it take them to review a document without the tool set? Now how many can they review with the tool set? So operational efficiency is clearly important. Risk mitigation is another pretty important one that's sometimes a little harder to quantify the ROI because it's like, well, what, what bad thing didn't happen?
- [16:36]
Greg Kilstrom
Right, right.
- [16:36]
Matt Blumberg
You know, it doesn't take much for you. One of our clients, and it was business unit that we don't work with had, it was a UK banking client that had £125 million fine last year because the, the marketing they had didn't follow the consumer duty law regulations in the UK and it was at, I think it was something at a 12th grade reading level and it needed to be at a sixth grade reading level. Like that's a lot of money. So you know, the absence of that is important to measure in terms of risk mitigation. So those are sort of more classic ROI measurements that marketers, you know, would be used to. The other one that though, that I would point people to think about is something that one of my board members who's a, you know, an AI pioneer says all the time, which is that AI gives you an opportunity to do things that were previously unimaginable. And it's hard to put an ROI number on something that it's not. Like you used to do it one way and now you do it better, but you couldn't do it before and now you can. So maybe you can get yourself to an ROI on that, maybe you can't. The example I always give is like, oh, you, you know, you just changed a whole bunch of terminology and now you're just going to fix 10 million pages at the, with the click of a mouse. Like, I guess you can figure out what that would have cost you, but it's more likely you just never would have done it.
- [17:49]
Greg Kilstrom
Yeah.
- [17:50]
Matt Blumberg
So anyway, I think there are lots of ways to get at the value and the value proposition of guardian agents, whether they're ours or someone else's.
- [17:58]
Greg Kilstrom
Yeah. And I mean, I think to, to that end there's a, I think there's a limit to efficiency alone as the only, as the sole measure of success. So, you know, what you're talking about is let's elevate the role, not only the role of the people, but what the business is capable of doing. And that's. Yeah, admittedly, I wish there was a formula for that. Right. But yeah, exactly, but, but yeah, it's, it's a little harder to quantify, but that's where you get near exponential. I know is a strong word, but like, you get, you get some exponential gains from, from that unknown stuff when applied correctly. How do you think about, you know, markup, AI staying ahead of the curve as far as, you know, with, with everything. I mean, you know, you mentioned mcp, you know, nine months ago. I don't think many, if any, were talking about that. You know, this is, this is a fast changing space. Agentec, all this stuff is moving pretty quickly. Like, how do you look at staying ahead of the curve?
- [19:00]
Matt Blumberg
Well, we think that a company's brand voice and values and rules are going to increasingly be a company's differentiators in terms of content strategy. So content strategy, you know, for the last 10, 15 years has been a workhorse of B2B marketing, but it's been about like volume, basically. Like, you know, how much can you put out there for SEO juice and you know, driving the inbound funnel. But in a world where large language models can produce infinite amounts of content for free, the differentiator is going to become brand and the ability to have strong, clear, brand, agile brand, as you say. Right. So when I think about our strategy and what we're going to come to market with over the next few months, it is to make sure that we add new guardian agents to our suite that help brands be even more surgical about quality. So as I mentioned, our first agent, which is the brand guardian agent is very focused on like tone terminology, you know, brand voice, et cetera, spelling and grammar. There's going to be another one that comes out pretty soon about, about policy and risk. So making sure your content is following your policies. There'll be a regulatory guardian agent that we put out that makes sure that you're not running afoul of the laws that you're subject to in different jurisdictions. There's going to be an accuracy guardian agent that makes sure when you, when you know someone puts some content out, it doesn't disagree with other content that you've put out in the past. There's going to be a data leakage guardian agent. So our objective is to, is to really own the space of content Guardian agents and make sure that enterprises competitive edge around the quality of their content and brand is something that they can kind of one call to an API in the middle of their content supply chain just gives them peace of mind at quality.
- [20:50]
Greg Kilstrom
Yeah, yeah. And, and as you're navigating those things, what do you see as. Certainly there's a lot of existing challenges today with content strategy, but looking ahead a little bit, like what are some of the challenges that you think, you know, may enterprises may be facing in the, in the months to come?
- [21:08]
Matt Blumberg
I think the biggest challenge is going to be fragmentation. You know, it. Content production used to happen out of like a team with like a limited number of people and it now happens out of every corner of the organization. Everyone is using different large language models. Even if a company feels like they've standardized, you know, that your people are like doing something off to the side on their personal cloud account.
- [21:28]
Greg Kilstrom
Right.
- [21:28]
Matt Blumberg
And so it's just fragmentation, kind of proliferation of content sources, content creators, content models. And you know, and that's why this, I think this concept of governance and single source of truth, you know, control over the algorithm, all those things are what companies are going to have to really invest in the state, you know, kind of stay on top of that fragmentation.
- [21:50]
Greg Kilstrom
Yeah, yeah. And I mean it's so definitely just to go back to markup AI. I mean it definitely sounds like the, in the native context or however you want to look at it. Like being within the app that you're already in can certainly it's, it's, it's one less place that people have to go. And it's also a consistent thread throughout all of the tools. I mean, that, that should be huge for those orgs out there that are siloed. And to your point, the large ones, they've got one of everything, even if they think, even if they've spent time consolidating and everything like that as well. So.
- [22:23]
Matt Blumberg
Yeah, yeah, no, that's right. I mean, that's, that's certainly what, where we're, where we're trying to go, where we see the market going. We have Fortune 500 companies that, that have millions of pages of content and tens of thousands of creators that use our platform to kind of centralize that, that governance. And, and we also have, you know, tiny companies where there may only be one user, but they still want to make sure that even at lower volume, that they still have that brand edge and consistency.
- [22:46]
Greg Kilstrom
Yeah. Yeah. Well, Matt, thanks so much for joining today. I really appreciate you sharing your insights. One last question before we wrap up. What do you do to stay agile in your role, and how do you find a way to do it consistently?
- [22:58]
Matt Blumberg
Agile in my role. So I love learning and I love learning from other people. There's only so much you learn from talking to yourself and talking to your own team. So I, I spend as much time as I can, not just kind of reading and listening to podcasts, but actually kind of talking and networking with other CEOs and understanding what are you seeing, what's working, what's not working for you. And every time I come away from one of those conversations, I got at least two notes to go apply back at the office.
- [23:24]
Greg Kilstrom
Yeah. Love it. Well, again, I'd like to thank Matt Blumberg, CEO at Markup AI for joining the show. AI can create content, but Markup AI's content guardian agents perfect your content. Markup AI instantly scans, scores and rewrites any content to enforce your standards for brand voice, terminology and compliance, giving you the confidence you need to scale your use of AI. Learn more at www.superup.AI. thanks again for listening to the Agile Brand. If you enjoyed the show, please take a minute to subscribe and leave us a rating so that others can find the show as well. You can access more episodes of the show@theagile brand.com that's theagile brand.com and contact me. If you're interested in consulting or advisory services or are looking for a speaker for your next event, go to www.gregkillstrom.com that's G R E G K I H L S T r o m.com the Agile brand is Produced by Missing Link, a Latina owned, strategy driven, creatively fueled production co op. From ideation to creation, they craft human connections through intelligent, engaging and informative content. Until next time, stay curious and stay agile.
- [24:46]
Matt Blumberg
The agile brand. And Doug, there's nowhere I wouldn't go to help someone customize and save on on car insurance with Liberty Mutual. Even if it means sitting front row at a comedy show.
- [25:04]
Greg Kilstrom
Hey everyone. Check out this guy and his bird. What is this, your first date?
- [25:08]
Matt Blumberg
Oh, no. We help people customize and save on car insurance with Liberty Mutual together. We're married. Me to a human, him to a bird. Yeah, the bird looks out of your league. Anyways, get a quote@libertymutual.com or with your local agent.
- [25:20]
Greg Kilstrom
Liberty. Liberty. Liberty. Liberty. The new strawberry watermelon refresher is now at McDonald's.
- [25:27]
Matt Blumberg
It's made with strawberries and a whole
- [25:30]
Greg Kilstrom
lot of whimsical tea. It's one of many new drinks now at McDonald's.
- [25:35]
Matt Blumberg
Refreshers contain caffeine.