“For a lot of marketers, we’ve been getting by. We’ve managed to make some things work. But it’s time to have some help.” – Lisa Gately
Generative AI can unlock significant advantages for content teams, like faster content creation, better optimization, and improved creative abilities. However, there are also critical concerns that marketers should be ready to address—copyright issues, hallucinations, and unexpected inefficiencies, for starters.
In this #ContentChat recap, Erika Heald is joined by Lisa Gately, Forrester analyst (check out her content here), to discuss generative AI and content marketing, explaining how content teams can use genAI to its fullest (and most ethical) potential.
Q1: What is generative AI? How does it differ from the other types of AI that haVE been around for decades?
Generative AI tools are trained on structured and unstructured data to create content.
“Generative AI is a set of technologies that leverage structured and unstructured data so that you can generate new content. It could be in the form of text, image, audio, or video.” – Lisa Gately
GenAI enables people to engage with datasets using natural language prompts, which large language models can understand.
“What that does is different [than other types of AI] because you might have some natural language prompts and be able to interact with this data to do that generation. ” – Lisa Gately
Marketers have been using AI for decades, it just wasn’t as talked about or accessible.
“Marketers have been using AI for ages… [it was] helping us [do] things like clustering, segmenting, predicting things. A lot of stuff for content; content targeting, personalization, predictive scoring, you can go on and on about some of the uses we’ve had. AI has been embedded in a lot of the martech stack.” – Lisa Gately
“With generative AI, you now can get at some of the data that previously wasn’t as available to average marketers, and to be able to use that with language prompts, or with some vendors you are even able to do some generative matching with imagery and things like that.” – Lisa Gately
“Many of us had been using marketing technologies that had AI components without even realizing it. There would be recommendations or insights that would automagically turn up. There wasn’t a person behind there pushing out those prompts to you. Stuff was happening with AI behind the scenes, but it feels like because of the availability and accessibility of the GenAI tools like OpenAI’s ChatGPT people started really talking about it and experimenting with it. With the other kinds of marketing tools, you had to have an enterprise budget and to have signed on and done the training to understand how the AI was working.” – Erika Heald
Q2: What are the benefits of using generative AI for marketers?
Generative AI tools can help improve your time to market, productivity, and boost your creative possibilities.
“It’s early days, despite what everybody is saying. I see a lot of people starting with improving time to market and productivity, and especially in content, you hear people talking about creative possibility. That’s one side of it. There are so many activities that help you be able to riff on and streamline some things to get those activities done in the content lifecycle.” – Lisa Gately
Marketers are increasingly using generative AI to personalize content effectively at scale.
“We also see some things where more marketers now are talking about not just doing things faster or more efficiently, they want to see some effectiveness. And what I mean by that is: clearly, we’ve talked about personalization for years as marketers, they’re now starting to talk about how can I use this if I’m going to personalize and create more variants of things and be able to get more variants. Testing out the interactions for our audience and being able to see ‘does this make a difference for us.’ You’re starting to see people think about this as helping us speed this process. In the past, some limiting factors for campaigns or being able to really customize the experience—it’s now within reach.” – Lisa Gately
Email marketing is a great area to explore personalization using AI.
“It’s helped, too, that there are some of the visual email marketing tools like what Knak recently came out with, where you don’t have to be down into the bowels of your automation system and understanding how to program things. You can just select some copy and say I want this copy to vary based upon the customer segment. You can have as many customer segments as you realistically can create differentiated copy for or create different images.” – Erika Heald
Think of generative AI as your creative partner.
“With AI, you can actually test yourself and say, I have this copy in place, and I think it’s good for this audience, but is there something I’m missing? Can you optimize a little bit more for that audience? And frequently, AI will give you great suggestions for how to change a word or something to look for in an image that you hadn’t really thought about. And then all of a sudden, you’re able to do that personalization at scale, where you can talk to all your different audiences without having to create 50 separate emails that have to go through approvals.” – Erika Heald
Q3: How are content marketers using generative AI? Are you seeing differences between B2B and B2C use cases?
Lisa works with B2B marketers of all different backgrounds, and their generative AI use varies depending on their organization and role. That said, she sees several common use cases:
“What I see overall is great interest in the content lifecycle, no surprise, no matter what their responsibilities [are]. It could be from thinking about planning and using generative AI not just for generating things. I hear people who are using it to summarize information. Or if you’re using it to do some research, it’s a thought partner.” – Lisa Gately
Teams use generative AI to accelerate their content planning and optimization.
“How do I make this more specific? How do I work on topic selection? How do I work through and might fill in some gaps? Where can I stand out?” – Lisa Gately
Marketers also realize that generative AI should not create entire assets.
“I know everybody is thinking about: I just want some headline ideas, help me with an outline, help me with the beginnings of a brief, help me with some drafts. What a lot of marketers realize is this is not something that’s going to give you a whole asset.” – Lisa Gately
Generative AI addresses our skills gaps, especially in creative work.
“You’re able to produce text, images, audio, video. I see tremendous power for creative teams. We know a lot of content teams are not distinguishing, but definitely in the realm of some imagery. And what is possible for people who have had concepts in their mind and maybe they’re not a designer, you’re able to get to some things quickly.” – Lisa Gately
Audio transcripts from events, webinars, and podcasts can be created using generative AI. This extends to translation, too.
“This [also] has the power to get the transcript for marketers who are working with events or recordings and they’re trying to then repackage or reformat across formats. Events, webinars, podcasts, [there are] really great examples of people being able to do accompanying material for that quicker.” – Lisa Gately
“Back to customization, the localization translation. We know that marketers have used machine translation to some degree, but I see generative AI opening up some interest there, too.” – Lisa Gately
“In one of my last in-house roles, I was in charge of content for an entire global team. Sometimes, we would have to wait for weeks [for translations] because our regular translators were busy. We didn’t have a tool that was good enough for that first version. So it could take weeks before we could get something to them. And, inevitably, that means that somebody would create their own version of something that was not good and not what we wanted to have out there. Just being able to have things like translations or transcripts be readily, immediately available [is a major benefit].” – Erika Heald
And generative AI makes it easier to repurpose content for new formats.
“Think of the things where you’ve put work into a primary format. If you’re crossing over and doing something else — social posts, landing page, email, blog post, related articles — there are other things that you can do; you’re reusing a lot better.” – Lisa Gately
Lisa explains the differences between B2B and B2C buying cycles and generative AI uses below:
“In B2B we know that buyers tend to buy in groups. It’s much more likely you are dealing with lots of individuals who are doing some research, whether they’re visiting your website or consuming and engaging your content, as opposed to just an individual. In B2B there’s complexity of how do your marketers and sellers work together. We know there is a lot of content fueling engagement. From the time that you are working with somebody in the very beginning stages, across marketing and sales are you able to capture some of these signals? I see a great retooling of systems across marketing and sales. How do we start to learn about our audience? And based on some recommendations or what we know, can we start customizing it and deliver a more contextual experience?” – Lisa Gately
“From B2C, what we’re hearing is yes, there’s some interest. It might be more in that dynamic, creative optimization. And I think for some of the big brands, they’ve been doing this to some degree, it’s now going to get faster. One recent conversation is about how do you really hook on some things at the speed of social media. We were talking before this call about Barbie and things that are in pop culture. B2C people will have a lot of interest on that note of customizing and personalizing or being able to respond when there’s a really timely opportunity. You can start customizing and coming out with content that just really would not have been possible in the past.” – Lisa Gately
Q4: Are there risks of generative AI that marketers should be aware of?
Generative AI poses several risks, most centered around how copyrighted data or intellectual property is used to train models and generate outputs.
“There is so much optimism, but there are really thorny problems that there is no set answer coming soon on.” – Lisa Gately
“Your intellectual property. Are we going to see some lawsuits or some rulings that come about? How are these models trained on copyrighted data? I hear of marketers saying their legal team is very concerned—are we inadvertently infringing on copyright? Or are we working on something that’s really important to our business? It’s really high-investment content. We want to protect the copyright for that.” – Lisa Gately
“It’s worth talking to everybody who has a role in content. Here’s brand guidelines and our brand voice. I shudder [thinking] about everybody just doing their own prompting of what they think your brand voice is to inadvertently good intentions. But they may be sharing some information in prompting and using these tools that really shouldn’t be going outside the company.” – Lisa Gately
Companies must train their teams and partners to understand the risks of generative AI, including that models are flawed and can present factually inaccurate information (known as hallucination).
“Be very thoughtful about how you’re using generative AI. Have those conversations not just with your employees, but your vendors, your freelancers, anybody who is involved in the process. Everyone talks about hallucinations or trust. I think [deep fakes] are going to become the next area, especially in an election year.” – Lisa Gately
Evaluate all generative AI outputs for potential bias. This spans written and visual content.
“How do you make sure that you are protecting your company, your brand? You really want to have a check on accuracy. The more that generative AI is used, you have to have some humans who are checking for editorial and judgment-based skills. Do we recognize any signs of bias?” – Lisa Gately
“Some marketers tell me their company is not using imagery or video yet, but I know technology companies are moving really fast there.” – Lisa Gately
And assess if generative AI is inadvertently adding more work to different parts of the content cycle. This is often fueled when teams use many generative AI tools without a clear purpose or governance over which tools should be used.
“Are you starting to see some points along the way in the content cycle where someone is using AI, and it’s just pushing more work to another stage in the cycle? Because you start to recognize it, and you have to go back and rework and talk with people about what we’re really trying to do here. How do you get a handle on the content you have reusing it? All of the things [companies] are talking about in terms of unapproved tools, I also see this as opening the door where it’s not creating the efficiency that some marketers or companies think.” – Lisa Gately
“[It creates issues when] it’s not being integrated with other tools. It’s being used on a standalone basis. So it doesn’t know what is in your content inventory. It hasn’t read everything. Most marketers haven’t taken the time to train their own generative AI, they’re using someone else’s model. So it doesn’t know the whole breadth of their offering.” – Erika Heald
On a positive note: Generative AI does not replace the need for human expertise.
“A lot of marketers may ask, what is this going to do? [Will it] disrupt our jobs? But I still see there’s going to be a need for people who have that skill: You understand your audience, you understand your company goals, your messaging, your positioning, your brand voice. Do you have a handle on content analytics? How do you manage some of the content, all of these collaboration challenges that happened in our lifetime before generative AI? Generative isn’t going to take over and make this easier. You still have to have some of that content operations [skill] and a lot of these skills and expertise.” – Lisa Gately
Q5: How can content marketing teams create guidelines or structure around their use of generative AI?
Organizations need a documented brand voice to ensure generative AI outputs maintain that voice.
“When used properly, AI can be a tool that helps people in all of their customer interactions to maintain the brand voice. That means you have to have documented your brand voice. [Without that] how on Earth can you expect other people to be creating content for your organization? And if you haven’t thought about what is that differentiated voice, how do you sound different, what elements about the way you construct your content that is different than everybody else—you’re going to have really boring, generic content that sounds like everybody else’s. And nobody is clamoring for more of that.” – Erika Heald
Identify a clear business problem that generative AI can solve before implementing it.
“First of all, a lot of education across the marketing team. Not just the content team proper, but everyone you’re going to be working with for stakeholders. Raise a lot of awareness about what does this do for us. Then it’s stepping into: Do you have some idea about what business problem this is solving. We’re not using generative AI for the sake of using this technology. There’s some direction about what you use it for.” – Lisa Gately
Assess what internal capabilities you have and choose a specific use case to start testing generative AI.
“Second, as a company you [need to] assess what capabilities you have and where you might step into this. Content is one of the low hanging fruit. We know the complexities there. Start working with others and get more specific about use cases. Look at your content maturity and think about the areas that are well suited for your organization. Start there, start framing projects with others and involving the experts. SEO, email, your blog, there are many different areas where you may choose to start working first.” – Lisa Gately
Partner with leaders in IT and security to create a policy around your generative AI use.
“Is there some kind of company policy? Is there any specified, accepted list of tools. I talked to a lot of marketers that are going to be talking with IT or security or others across the organization.” – Lisa Gately
Have a clear purpose for each generative AI tool you use, and consider purpose-built tools for specific needs.
“We see a lot of capabilities coming to market right now. But then, also, do you consider some other purpose built tools? If your company has an agreement with OpenAI or Microsoft or Google, there are multiple areas where you could do this work. It’s about identifying what type of work gets done where.” – Lisa Gately
Q6: How should marketers decide where generative AI can improve their workflow?
Create a core team of people who will start experimenting with AI and identify the right use cases.
“Form some type of core team. Usually, there are some team members who are identifying these specific areas where you’re going to test or pilot using generative AI. I tell people apply the same rigor as a project where you’re going to be giving some sort of a readout or sharing what you learned to marketing leadership.” – Lisa Gately
Treat it like a scientific experiment. Assess your baseline and see if and how generative AI improves your existing processes or boosts your results.
” If you are testing out your hypothesis, capture some sense like it’s an experiment of here were the inputs or prompts, here are the outputs. Here’s actually an area where we’re trying to rate or observe, did this do what we thought? Is it doing this at an acceptable level? It may take some time for the marketers who are first using it, but be able to compare your baseline process. That cycle time in producing the content, is it also effective, is it performing compared to our human-generated content, is it giving us anything?” – Lisa Gately
Find use cases that work for your organization and your unique needs.
“Use cases are very subjective. Just because you were hearing about some other organization and what they’re doing, choose a few areas that are helping you get some result, and be ready to talk to others in the team about what you found.” – Lisa Gately
Continually talk about your efforts—both good and bad—and build a library with your successful prompts and lessons learned.
“I hear of some of the organizations moving forward. It’s a continuity of people are sharing not only best practices, but they’re trying to set up a prompt library or they’re trying to have some ‘show and tell’ time of not only showing here’s what works, but here’s something I discovered that I don’t want to do it this way again.” – Lisa Gately
Q7: What generative AI tools do you recommend for content marketers?
There is no set list of tools for all organizations.
“I want to give the biggest caveat that there’s no set list for all organizations. We all have different starting places and paces for bringing new technology.” – Lisa Gately
Start by seeing what internal capabilities you have.
“Look through your existing martech stack. There might be some existing capabilities there. Ask those vendors what’s coming.” – Lisa Gately
Lisa encourages you to consider your options across the content lifecycle, but don’t get overwhelmed thinking you have to try everything.
“You’re not going to find one tool that does absolutely everything across the content lifecycle. There’s a huge array of things. Don’t get overwhelmed thinking you have to experiment with everything. You might place a couple of strong bets, and that’s looking at what you already have access to and maybe a few other tools.” – Lisa Gately
Focus on areas that support your organizational goals and objectives.
“Look at what your objectives are for the year. And if one of your objectives is to get more ROI out of your events, for example, then that means you need to do more content repurposing. So maybe you’re going to be looking at a tool like Descript so you can take those videos and get a really fast transcript and be able to create a ton of social media snippets.” – Erika Heald
We share a few generative AI tool considerations in this blog post sharing 70+ marketing templates and tools to boost your results.
Q8: What pitfalls should marketers avoid as they evolve their use of generative AI?
A solid content foundation is essential for generative AI success. Start by reviewing and refining your brand voice, personas, templates, and taxonomy.
“Your starting place will affect your success. Some of this is if you have a handle on the content you have. What are the examples of the best content to repurpose or think about for training? You as a content person are in a really good spot to be able to identify a lot of that.” – Lisa Gately
“If you suspect you have reason to review your taxonomy and make some changes with tagging, also think about that. Also, again, the messaging, positioning, your templates, and brand voice. Even sharpening on personas.” – Lisa Gately
Understand how you will measure your success.
“I’ve talked with some really savvy clients who are thinking about how are we getting ready if we want to measure AI generated versus human generated content. You’re thinking about that from the start, not just waiting until people are producing drafts of different things and stuff starts flying around.” – Lisa Gately
Establish clear workflows to support the introduction of generative AI capabilities. It will take time for your team to get up to speed.
“This is not overnight, instant success. You’re trying to bring people on board as a team. Think about your workflows. Prepare people. We know a lot of teams are overworked, and you’re already doing, let’s say, the next big thing for your company’s launch and you’re doing these pilot projects. Be ready to communicate and get some help. You may have some experts who can work with you on this. But it’s keeping rigor around it, because you’re going to be doing some kind of sharing or readout eventually to tell people this is why we are recommending and we’ve got this approved use case.” – Lisa Gately
Encourage constant evaluation and information sharing. A regular meeting can help with this.
“How do you come back and look at how you’re going to be evolving little by little? People are going to have findings along the way. Is there some kind of—whether you have a regular team meeting or a content carousel—way to talk about how it is helping you? Or if you got some early observations to help troubleshoot where some teams need help.” – Lisa Gately
“With the pace at which they’re releasing new models or new functionality, that alone means you need to be evaluating how this stuff is working pretty regularly, and adapting your approach. If not, you’re gonna get really blindsided when all of a sudden there’s a new version that does things very differently.” – Erika Heald
Be prepared to address your team’s change fatigue.
“I think people are underestimating the change management and the communication that’s going to be involved. Change fatigue is going to be on a lot of people’s minds.” – Lisa Gately