Out-of-the-Box AI Tools, Content Scaling, and the Foundations Every Marketer Needs Before Adopting AI
Guest: Mike Kaput, Chief Content Officer, Marketing AI Institute
Hosts: Seth Earley, CEO at Earley Information Science
Chris Featherstone, Sr. Director of AI/Data Product/Program Management at Salesforce
Published on: October 1, 2021
In this episode, Seth Earley and Chris Featherstone speak with Mike Kaput, Chief Content Officer at the Marketing AI Institute - a company that provides online education for a worldwide audience of marketers and produces the annual Marketing AI Conference (MAICON). Mike's path to AI is an unusual one: political science major, studied abroad in Cairo, moved back to Egypt during the economic crisis to work as a journalist and editor for multiple magazines, returned to the US for freelance writing and marketing consulting, then joined PR 20/20 as the firm began exploring how AI could transform content marketing operations. In this conversation he explains what "out of the box" AI actually means for the average marketer, why documenting your processes before touching any AI tool is non-negotiable, how to identify where to start using a simple spreadsheet exercise, what the Marketing AI Institute's State of Marketing AI report revealed about why adoption is lagging, and why human judgment in the AI loop is not a temporary limitation but a permanent strategic requirement.
Key Takeaways:
- "Out of the box" AI for marketers means tools you can sign up for with a credit card, get started without a technical lead, and begin generating value quickly - but even these tools require your marketing strategy and customer journey to be documented and centralized before any AI implementation can succeed.
- The right way to start with AI is not to ask "how can we use AI?" but to write down everything you do in a day, week, month, and quarter - then identify the top three tasks that take the most time, cost the most money, and that everyone dreads - those are your first AI pilot candidates.
- You cannot automate a mess or automate what you don't understand: AI tools identify specific intervention points within documented processes, they do not replace the strategic thinking that defines what those processes are trying to achieve in the first place.
- The biggest barrier to AI adoption among marketers is not fear - only 15% of respondents in the State of Marketing AI report cited fear as a barrier - but lack of education and training, cited by more than 70%, combined with no clear path from awareness to implementation.
- Out-of-the-box AI standardizes what all your competitors can also do; the real competitive advantage comes from freeing up human time to do the things machines cannot - storytelling, strategy, creative judgment, and the distinctly human insights that differentiate one brand from another.
- AI bias and explainability are underappreciated risks for marketers: the Apple Card example shows how an algorithmic decision can create a massive brand crisis within 30 minutes, with no one able to explain why the algorithm behaved as it did - every AI adoption decision should include a worst-case scenario analysis and a crisis communication strategy.
- A human will always need to be in the AI loop - not primarily as a safety net to prevent the system from running amok, but as an active partner who extracts value the machine cannot generate on its own and who understands how to work alongside AI as a core professional skill.
Insightful Quotes:
"If you scratch an itch for AI you pretty quickly just run into a typical marketing strategy. A lot of the things people say - 'can't we just use AI to solve this problem?' - the answer is probably yes, but it needs so much foundational work in terms of your processes. I would not be trying to seriously adopt any AI tool without your basic marketing strategy and processes documented and centralized - what does the customer journey look like, what do your processes and approvals look like, even if it's not perfect." - Mike Kaput
"These tools are advancing to capabilities we probably can barely even dream of today, but unless we're talking about some very far future sci-fi scenario, a human will always be required in the loop and always should be in the loop. That is going to be a critical skill of employment in the future - understanding how to work hand in hand with machines. It's also going to be a key piece of strategy: mapping out where your humans are in the loop, what they do, what their responsibilities are, so you get the most out of the systems." - Mike Kaput
"Standardization is great for efficiency but differentiation is great for competitive advantage. Right now there are brands straight up creating content at scale in an automated way to dominate search results - if you can imagine it, it's being done, probably by one of your competitors. But further down the line we need to free ourselves up to do those human things that are the differentiators. Telling the right story to your market is something you need to be devoting a lot more time to - and AI adoption is what should enable you to do more of it." - Mike Kaput
Tune in to hear Mike Kaput explain how the Associated Press automating earnings report writing was the original spark that got him thinking seriously about AI, why Marketing AI Institute took on over a million dollars in seed funding during COVID, what questions to ask any AI vendor about where their training data comes from and what the machine is actually doing versus the human, and why the key insight from surveying 400 marketing leaders was that most of them know AI is important but have no idea how to get from where they are today to actually using it.
Contact Mike:
Mike@pr2020.com
https://www.linkedin.com/in/mikekaput/
Links:
Marketing AI Institute
MAICON 2022
State of Marketing AI Report
The AI Powered Enterprise
Watch on YouTube
Podcast Transcript: Using AI to Enhance Marketing Operations - Out-of-the-Box Tools, Process Foundations, and the Human in the Loop
Transcript introduction
This transcript captures a conversation between Seth Earley, Chris Featherstone, and Mike Kaput about how marketers can use accessible, out-of-the-box AI tools to scale content operations today - covering the foundational process work that must come first, practical frameworks for identifying where to start, the ethics and bias risks every brand needs to plan for, and why human judgment remains not just necessary but strategically central to getting value from AI.
Transcript
Seth Earley: Good morning, good afternoon, good evening - whatever your time zone is today. Welcome to our Earley AI Podcast. What we're going to be doing is talking to thought leaders and practitioners about what's possible in artificial intelligence, as well as what is practical in this space. Really talking about what's emerging, what we're going to be expecting in terms of enterprise AI, and how to get from here to there - where "there" is going to be artificial intelligence embedded in all aspects of our personal and professional lives. I'm here with my co-host Chris Featherstone.
Chris Featherstone: Seth, always a pleasure to be with you. These topics are super fun to discuss. One thing I do want to add: all opinions are my own and Seth's own and don't affiliate with anything outside of this. But it's fun to be with these folks, and I think the key here is that a lot of people believe AI is this mysticism that only a select group of people can implement - and that's not the case. We're going to demystify these things and get to the heart and soul of how AI is now table stakes in the application space, and how low-code or no-code scenarios are now critical to enable business stakeholders to actually implement these things.
Seth Earley: Today we're going to be talking about how AI tools are increasingly embedded in various applications that don't require a lot of coding - they don't require a machine learning expert, a data scientist, or a data engineer. They're really out-of-the-box types of applications. Our guest is a man who has a real passion for making AI approachable, actionable, and accessible. He does this as the Chief Content Officer at the Marketing AI Institute. The institute provides online education for a worldwide audience of marketers and produces a yearly conference called the Marketing AI Conference, MAICON - the 2021 event just wrapped up and the next will be in August 2022. With that, I'd like to welcome Mike Kaput to the show.
Mike Kaput: Seth, Chris - thanks so much for having me. I'm excited to be here.
Chris Featherstone: Mike, why don't you give us a simple overview of your day-to-day and background?
Mike Kaput: My day-to-day is really concerned with building an audience for the Marketing AI Institute. We're trying to reach as many marketing professionals as possible through written, audio, and visual content. On any given day I'm either creating content, promoting content, or figuring out how to make our content strategy work better. And as part of that, taking all the traffic we get from our site and turning them into an engaged audience - there's a community element to that, and a serious revenue operations element as they move further down the funnel and become paying members of our online education products or event attendees. So on any given day I'm trying to increase traffic, increase leads, and hopefully convert leads into happy customers.
Seth Earley: I should mention how Mike and I met. When I was doing research for my book, The AI-Powered Enterprise, I was looking at different events and venues, and I came across this Marketing AI Conference. I had low expectations - I thought these are going to be marketers, this is going to be arm-waving, this is going to be marketing buzzwords. I was very surprised to see the depth, detail, and quality of the presentations. A number of my case studies for the book came directly out of that conference. It's a really high-quality event, highly recommended.
Chris Featherstone: Mike, I understand you started as a journalist in Egypt - give us some details on your background, because it's fascinating.
Mike Kaput: I started as a political science major in college, thought I probably wanted to do something in government. I ended up studying abroad in Cairo and then moving back there during the depths of the economic crisis - studying Arabic in Egypt seemed like a good idea if I was going to go the government route. Once I got there my plans changed pretty seriously. I fell into working for a couple of magazines doing both reporting and editing and just really loved it - I'd always loved writing, reading, and learning about the world, so it was a really good fit. Worked in that for a few years, then ended up moving back to the States doing freelance writing and marketing consulting, still mostly focused on creating traditional content.
As anyone who's read the news recently knows, the world of journalism has been turned upside down over the last 10 to 15 years. I saw that, saw colleagues and friends getting laid off and overworked and underpaid, and thought there may be some other ways to use these skills. That's how I ended up working for a marketing agency called PR 20/20 - which is still where I work today. The CEO and founder Paul Roetzer has always been focused on marketing automation and where technology is going in the marketing industry. We were HubSpot's first ever agency partner. I got my start cutting my teeth on using content marketing and marketing automation to grow client businesses.
As part of that, Paul and I both got really interested in artificial intelligence - for Paul it was many years back, for me it was probably about six years ago that we really started seriously talking about it. At that time we'd just heard about AI, read a couple of articles, maybe a book or two, and started thinking: maybe this could have a real impact on our business and our clients' businesses. There seemed to be a lot of confusion in the market, a lot of buzzwords, and a lack of connecting the dots between this interesting technology that is often overhyped but actually very powerful, and actually using it for business results. That's where we gravitating toward. We started building a very large audience for what we were writing and talking about. A couple years ago, the Marketing AI Institute spun off into its own company, we launched our conference in 2019, launched AI Academy for Marketers in 2020, and also took on about a million dollars in seed money late last year because we realized - between COVID, between the trends we were seeing, between what we had observed over the last 15 years in the marketing industry - we made a big bet that the time has come.
Chris Featherstone: What are your top AI use cases that you're driving right now?
Mike Kaput: I should caveat that we're always growing and learning. At any given time I'm using anywhere from half a dozen to a dozen different AI tools and platforms, kicking the tires on them. We have a few tools we always recommend and are actively using in depth, and many others we're just trying out. From our perspective, what we're really doing a lot with when it comes to AI - if I had to nest all these use cases under one thing - it's scaling content marketing and scaling content programs. One of the biggest things that got us really interested in AI in the first place was hearing stories about how certain AI systems supposedly could write news articles. Associated Press had used some type of tool to automate writing earnings reports for publicly traded companies - that kind of led us down this rabbit hole.
At our agency at the time we'd been hiring a lot of writers with journalism backgrounds because we were doing so much content marketing for clients. The problem is - as anyone who has created content knows - it takes a really long time and it doesn't always work. So our big thing is looking for AI use cases in the content space: tools that can help us identify what type of content we should be creating, help us brainstorm what should be included in pieces of content to rank in search or appeal to the intent behind what people are looking for, and tools to actually create parts of the content itself. Language models have dramatically improved in probably the last year - that's one of the really interesting signals that we're finally coming into an era where AI is going to start really changing how people do their jobs.
There's a heavy element of content creation, content strategy, and personalization in what we do with AI today.
Seth Earley: A lot of people think of AI initiatives as requiring a lot of heavy lifting, a lot of preparation, a lot of work on data sources. The premise here was there's stuff we can do that's out of the box or baked into other technologies that doesn't require all that heavy lifting. How would you define that flavor of AI technology - and what's truly out of the box versus what do you have to do to make it work?
Mike Kaput: That's a great question and it's a huge challenge. There are so many solutions that may work really well but are not suitable for a smaller business or a business without the right foundation. The good news for us is that we are a small startup, so we're forced to find things that work out of the box on a relatively small budget. When I think about out of the box, what I really mean is: can your average marketer get started without having to consult any type of super technical lead? Some of these tools literally function like your average SaaS tool - you can sign up for a free trial, pull out a credit card, and start using them. I've literally done this: even if someone's not in the office or I can't find someone with a company credit card, I just whip out a card and start paying for a month. That's really exciting to me - I don't think we were able to do that even three years ago.
Seth Earley: That's valuable for experimentation. But you still need supporting processes. You still need to understand your customer engagement strategy, still need to do some work around what content is strategically important. You need to think about what's different in the marketer's job when it comes to some of these out-of-the-box tools.
Mike Kaput: I joke that if you scratch an itch for AI you pretty quickly just run into a typical marketing strategy. Because a lot of the things people say - "can't we just use AI to solve this problem?" - the answer is probably yes, but it needs so much foundational work in terms of your processes. One of the things we learned the hard way, because we've swung and missed on a couple of pilots that really set us back in time and money, is that I would not be trying to seriously adopt any AI tool without your basic marketing strategy and processes documented. What does the customer journey look like? What do your processes and approvals look like? It needs to be centralized in one place, well documented - even if it's not perfect. Then you can say: where does AI plug in and play into these processes?
Seth Earley: I have a phrase I like: you can't automate a mess, and you can't automate what you don't understand. The AI tool is not going to figure it out for you. Humans need to understand the strategy and the process, and then you're looking for interventions in steps of that process - as opposed to some grand galactic AI that's going to solve all your problems.
The upside is you get to experiment and try and fail at low cost and low risk. The downside is when you start doing all these independent things it can become a little shadow IT. When you get these experiments and try to bring them together or centralize them, what do you need to be aware of in terms of standards and policies?
Mike Kaput: That's always the danger. The more you adopt certain technologies the more complex things get - technology creates a new problem all over again. My read on the marketing AI market specifically is that it is still early enough that fragmentation is unfortunately going to be a cost of doing business for the time being. What's critical is sitting down before you even look at AI and understanding your use cases at a deep level. What are the business problems we're trying to solve, and what do they look like in their smallest, most granular form?
Our problem is "scaling content marketing" but that's enormous - it really boils down into about 30 or 40 different things we need to do better. You may start extremely small and say: it's a real pain that every time we write a blog post, we have to do keyword research, we don't have visibility into the intent of the search, we don't have anything smarter than generalized search data to pick topics from. That's the core problem. That's the specific use case. Then you go ask: is there AI that can help me with content topics, SEO research, keyword research? You have to be really careful upfront and say: here's the handful of very specific things I want to achieve. I don't know if AI can achieve them, but I'm going to go look.
Seth Earley: That's a great way to think about it. Chris, you also wanted to talk about ethics - where does that play a role here?
Chris Featherstone: Mike, where does ethics play a role in all this?
Mike Kaput: That's probably a trillion-dollar question on a long enough timeline. I think a lot of people have that question but not enough are talking about it in a robust way. Everyone knows it's important - there are ethical concerns, bias concerns - but few people I've seen have really been diving deep into these issues. And for brands it's going to be critical.
Here's a practical example: Apple released a credit card and got put on blast by an entrepreneur with millions of Twitter followers because him and his wife - who had essentially the same financial picture - both applied, and he got approved for ten times the credit limit she did. He was furious and went public about it in real time. Within 30 minutes Apple had a massive PR issue on their hands. And the worst part is, as they tried to navigate this in real time, they couldn't answer why the algorithm had done it - because Goldman Sachs had built it, Goldman Sachs wasn't available to talk about it, and suddenly you have this explainability issue where the algorithm is having a real effect on someone's life, the brand is losing equity in real time, and nobody knows who's actually responsible. Apple later said: sorry, we can't really help you, we didn't make it.
That alone is one simple example of all the considerations around ethics, bias, and explainability that nobody is really building a strategy around. There should be a crisis communication strategy for when your AI tools go wrong.
Seth Earley: What questions should marketers be asking to safeguard their brand?
Mike Kaput: The first question I'd be asking any vendor is: where is the data coming from, and how is the machine making its predictions? They won't be able to tell you every nuance, but knowing where the inherent biases are in the data matters. And when we say bias, we don't only mean discrimination against a group of people - that's the blatant example. It can also just be a biased dataset that's optimized for the wrong outcome. Like any survey: if someone says we surveyed 4,000 business professionals but they all worked in the same industry, you'd say that's not widely applicable. You have to watch out for both. I also think it makes sense to get creative and brainstorm the worst-case scenario. What does the machine do versus what does the human do? That lets you isolate how much control you have. For the marketing tools we're talking about today, nothing is going to run off and talk to your customers on its own without your approval - but when you do get into more sophisticated recommendations and personalization, it can get interesting.
Chris Featherstone: Where should someone start if they're thinking about adding AI to a solid existing marketing workflow?
Mike Kaput: It sounds simple but it's important: sit down with a spreadsheet and write down everything you do in a day, a week, a month, and a quarter. Literally: "write blog posts - every week." "Performance reports - monthly." Do it for yourself first, then have your team spend ten minutes doing it too. What you end up with pretty quickly is a list of all the activities that take up your time.
Time-saving is not the only opportunity with AI and probably not the biggest one, but it's a good starting one because it's easy to understand. I would go through that list and ask: which of these do I dread? Which are the ones I really don't like? For me it's putting together performance reports that have stuff I already know - they take forever, they're really repetitive, they still rely on data. That's actually a really good starting use case for AI. You're not taking away activities that are giving you value or that you like doing. You're just looking for a smarter way to handle something you hate, and you come out net positive. Start with the top three things that take the most time, the most energy, the most money, and that everyone dreads - and then ask: is there a way to do them smarter? That's when you have a really good place to start looking into a small AI pilot.
Seth Earley: That's a really great way to think about it. AI is going to reduce a lot of the work that people don't want to do anyway. When we first started using steam shovels at the turn of the 20th century, people worried about what all those ditch diggers would do. Well, people don't like digging ditches - and once you automate that, you can build cities and skyscrapers and highways.
There's a question of standardization versus differentiation. Out-of-the-box tools raise the bar for everyone - if all your competitors can also do this, how much competitive advantage is it really providing? When does something need to go a step further and differentiate rather than just standardize?
Mike Kaput: Right now there are so many marketers who aren't using any of these tools yet, and a small number of forward-thinking brands that are and are reaping massive competitive advantages. From a competitive advantage standpoint today, I don't think you can afford not to be using some of these tools if you're able to. But as adoption increases there will be diminishing returns. The technology will get better and will lift all boats.
Further down the line, we need to free ourselves up to do the human things that are the real differentiators. Storytelling is so important - it always has been. Telling the right story to your market is something you probably need to be devoting a lot more time to than you do today, and AI adoption is hopefully what enables you to do more of it. Language models have gotten to where they are now surprisingly fast - most of that progress happened in the last 12 to 18 months. That totally changes the game. There are brands straight up creating content at scale in an automated way just to dominate search results. If you can imagine it's being done, it's probably being done by one of your competitors.
Seth Earley: You have a research report - the State of Marketing AI report - that talks about barriers to entry. Can you tell us about that?
Mike Kaput: Marketing AI Institute partnered with Drift - one of the bigger AI-powered conversational marketing companies, recently valued at unicorn status - to better understand the gap. Literally the question was: if marketers should be using AI, why aren't more of them? We had a lot of hypotheses. We thought maybe people aren't there yet, maybe they're afraid. So we decided to do a survey and write a report. You can find it at stateofmarketingai.com - free for download.
We surveyed more than 400 marketing leaders across a range of functions, big enterprises and small companies. The biggest finding: when it comes to barriers to adoption, fear is not really a huge factor. Only about 15% of people said fear was actually a barrier. What was more interesting is that over 70% said education - a lack of education and training - was the biggest barrier. They knew AI was important. Other questions made clear marketers expected even five years from now that many of their tasks would be intelligently automated or augmented with AI tools. They just didn't know how to get there. They didn't know what tools to prioritize, what use cases to focus on, how to actually pilot these things. And furthermore, very, very few companies were doing any training for their employees on AI. Marketers are mostly at this beginning stage, ready to move on - they just don't know how to get from point A to point B.
Chris Featherstone: How important is having a human in the loop?
Mike Kaput: I'll be honest - I don't think we're ever going to not have that. These tools are going to advance to capabilities we can barely dream of today, but unless we're talking some very far future sci-fi scenario, a human will always be required in the loop and always should be in the loop. That is going to be a critical skill of employment in the future - understanding how to work hand in hand with machines. It's also going to be a key piece of strategy: mapping out where your humans are in the loop, what they do, what their responsibilities are, so you get the most out of your AI systems. It's more about having a human in the loop to get the most out of the tools and to do the things the machine just can't do - not primarily as a safety net to prevent the system from running amok.
Seth Earley: Mike, this has been tremendous and very valuable. Our next episode will feature Massood Zarrabian from BA Insight, a search enhancement technology provider we've worked with for many years - we're going to talk about how AI and machine learning have evolved in enterprise search, what's new, what's different, what's easier. Thank you Mike, thank you Chris, thank you Sharon for all your work behind the scenes keeping the trains running on time. Thanks to everyone who joined us and who's listening online.
Mike Kaput: Thank you guys so much - this was excellent.
Chris Featherstone: Mike, where can people find more information about you?
Mike Kaput: Find me on LinkedIn - that's a great way to reach out. You can also email me at Mike@pr2020.com. And go to marketingaiinstitute.com - we have honestly over 1,000 articles on the subject at this point. Please avail yourself of those resources.
Seth Earley: There will also be show notes with those links and other resources. Again, tremendous session. Thank you so much - we will talk to you next time.