AI and marketing have great promise to complement each other. Used well, artificial intelligence can increase the efficiency, velocity, and agility of nearly all kinds of marketing work.
I’ve recently spent loads of time combining my own decades of marketing expertise with the brave new world of AI as I support the marketing efforts for the Agile Marketing Alliance, and I’ve come back with five use cases that nearly all marketers can copy.
Check out what I’ve learned, complete with highly detailed video walkthroughs for each use case.
There’s been a lot of speculation that AI will replace copywriters, including those who write blog posts. In my opinion, that’s unlikely to happen.
AI can make us more efficient in writing, but it can’t replace good writing.
Generative AI, like ChatGPT, is based on a technology called a large language model, or LLM. LLMs predict the next word or short phrase in a sentence based on their training.
The companies behind the technology train the LLM on large amounts of material: the Internet, various books and journals, etc. This library gives them a wide range of knowledge, but it limits them to what’s already been written about particular topics.
If you ask ChatGPT, or any LLM, to generate an 800-word blog post on a particular topic, say how to improve your marketing workflow, you're going to get something that sounds very familiar.
LLMs are predictive models, which means the content they create is predictable. That’s not necessarily bad, but it’s not automatically good either.
These blog posts may work for SEO purposes for a short time, although I suspect Google will soon figure out the difference between AI-generated content and human-generated content.
Once they do, they’ll no doubt rank the former lower than the latter.
Humans figure it out almost immediately. They know when a blog post is original, thoughtful, and creative, and will likely reward great content with shares and likes.
So as a general rule, I don’t recommend using generative AI to write blog posts.
I’ve found it much more useful to use Generative AI to create outlines for blog posts.
My approach when I have a blog post to write is to begin with my own outline for the post. I then ask ChatGPT to generate an outline with a prompt that's as specific as I can make it.
For example, my prompt might look something like this:
You are an expert Agile marketer. You’re familiar with the Agile marketing manifesto and with both the benefits and challenges of adopting Agile ways of working. Write in a compelling and professional style at an 11th-grade reading level.
Your first task will be to write an outline for an 800-1000 word blog post on the topic of Agile marketing in healthcare. Emphasize the business reasons that healthcare organizations should adopt Agile marketing. Give some examples of healthcare companies that use Agile marketing.
I then compare my original outline to the one generated by ChatGPT.
Often, I find that I’ve forgotten something that appears in the generated outline, so I add it. Then I write the blog post myself, in my own unique voice, with my own unique point of view on the topic.
Of course, I use AI, in the form of a tool called Grammarly, to check my spelling and grammar.
If you’d like to learn more about how to generate good prompts and use ChatGPT for writing blog posts, check out the first episode of my AI and Agile Marketing series.
As you explore how best to integrate AI into your content creation, keep in mind that it impacts your legal rights.
The courts have found that AI-generated material, both text and images, isn’t entitled to the benefits of copyright protection. So if you use AI to generate your blog posts, they’re not your intellectual property; anyone can copy them.
In some cases, this may not matter, but in most instances, you’ll want to proceed with caution.
I’ve found AI to be very helpful in creating transcripts of videos and podcasts, and then generating SEO-optimized descriptions for those videos and podcasts.
We use a tool called Descript to edit our podcasts and to produce transcripts of our podcasts and videos. We then use a prompt like the following to generate an SEO-optimized description:
Write an SEO-optimized YouTube description, including a title, short description, highlights, timestamps for the entire video, and hashtags using the attached video transcript.
Make it fun, 11th-grade reading level, and interesting with appropriate emojis.
If you’d like to learn more about how to do this, check out episode 3 of my AI and Agile Marketing series.
Whether you’re a B2C or B2B marketer, you have contact with customers. And those customers almost certainly tell you their opinions in their own words.
So why not use that language to improve your own copywriting?
Customers are more likely to respond to copy that they identify with rather than copy that is written from the point of view of someone inside the company.
For example, we’ve used a tool called OctoParse to scrape data about why people joined the Agile Marketing Alliance and we’ve used it to generate copy for our website.
We usually don’t use it directly, but directionally.
Taking this approach has proven to be very helpful. It gets us out of our own heads and into the mindset of the customer, helping us create meaningful website copy that converts.
Marketers can also use this approach to analyze large amounts of unstructured data (think reviews) and see trends and patterns.
If you’d like to learn more about how to do this, check out episode 4 and episode 5 of my AI and Agile Marketing series.
Like all time-strapped marketers, we try to repurpose content when we can. If we have a blog post, for instance, we might turn it into a video. If we have a great podcast, we might extract snippets from that podcast to promote it.
AI can help make this repurposing much more efficient, and it can do things that were previously out of reach.
For example, we use a tool called Lumen to generate short videos from our blog posts. It’s quick, (generally less than 5 minutes), and the videos are engaging and informative.
Here’s a walk through of how it works: episode 6 of my AI and Agile Marketing series.
Even more amazing is AI’s ability to take a single longer video and transform it into multiple, snackable, vertical videos to promote that video.
Let’s say I have a 5-10 minute video from a self-paced course. Using a tool called OpusClip, I can generate 8-10 thirty-second clips from different parts of the video.
Now I have a ready-made library of videos that I can share across multiple channels.
If you’d like to learn how to do this, check out episode 7 of my AI and Agile Marketing series.
Of course, AI isn’t just for written content. And its visual capabilities can be a big help to marketers.
Whether it’s a stock photo or original creative, marketers leverage a wide variety of images across their content and campaigns. AI can now generate images, and while it’s not as creative as what the best humans can do in many cases, it is good enough.
There are two major AI tools for generating images: MidJourney and Adobe Firefly.
The choice is more a matter of preference; both tools have their strengths and weaknesses.
While Firefly is in beta, you can’t use the images for commercial purposes. This should change soon as Adobe takes Firefly out of beta.
If you’d like to learn how to generate images using AI, check out episode 8 of my AI and Agile Marketing series [link to be provided next week].
AI has great utility right now for marketers who want more efficiency in their workflows, with even more promise in the future.
It can help us do more, faster. It also helps us gain insights into our customers by analyzing vast amounts of data to see trends and patterns.
But AI also has limitations.
It’s predictive, not creative, and it can sometimes give us incorrect information.
Marketers have to use it judiciously, understand its shortcomings and use it as a tool to augment human capabilities and creativity, rather than a technology that will replace us.