In 2022, artificial intelligence can drive cars, map the spread of infectious diseases, and recommend your next binge-worthy show. Some AI is even composing music and painting.
By some estimates, we could achieve “singularity”—or the point at which computers are proactively and exponentially improving themselves as the dominant intelligence on earth—by 2045.
Others don’t think this will ever happen. We think there’s no point worrying about it yet.
We’re wondering how it could impact marketing. Could certain marketing roles or responsibilities eventually be replaced by AI? Specifically, how will it impact content creation? With so much of a marketer’s work already living on digital platforms powered by AI (Google, social media platforms, marketing automation software, etc.), could AI-generated content ever replace human-generated content?
We wanted to know. Fortunately, our client Hiya wanted to know, too.
Human vs. Machine
Hiya is a SaaS voice performance platform that reduces spam calls and provides extremely impactful caller ID services to enterprises. We fed various AI machines content prompts for Hiya and gave the exact same prompts to the content team at 97th Floor. Take the quiz to see if you can pick out the human-created content.
We took these results back to Hiya to see what they thought about the AI content. Jonah-Kai Hancock, Hiya's Vice President of Demand Generation, noted that "Any time you are asking someone to read a blog or engage in an email or watch a webinar you are asking for their time and I don’t think that the AI does a really good job explaining what I would get out of that time.”
Rachel Bascom, Head of Content Marketing at 97th Floor, was surprised by what the AI could do. She shared, “The blog article from AI may rank fairly well. We could use it for SEO and it might please an algorithm, but I don’t think it would sell anything anytime soon. A content marketer is thinking beyond an algorithm in a way that AI can’t do. Yes, the AI piece might rank well, but what happens when someone opens that link? Human writers can think about the content journey and create something engaging, educational and conversational.”
Rachel is also feeling assured that she, a living breathing content marketer, will get to keep her job after this experiment.
No surprise here—we all felt that the AI content lacked personality. Especially in Hiya’s industry where personal touch is central to their product, this AI content could never fly.
But honestly, that’s what we expected. Here’s the process most marketers face when trying AI out for the first time:
- Sign up for a free trial at a number of content creation platforms,
- Feed prompts to a cold AI,
- Become underwhelmed with what was spit out, and
- Cancel the subscription, unimpressed (and a little validated, because most content marketers are rooting for the humans in this dogfight).
Maybe you’ve had similar experiences. AI is, most often, not where it needs to be for marketers, and many marketers feel that their existing, non-AI process for content creation is effective. To many, adding AI seems like an unnecessary disruption of that process.
Hancock shares, “It would be a lot more work for me to figure out how to make AI work. Unless my content team came to me and said ‘hey we really want this and here’s why,’ I don’t see this happening right now.”
Content-generating AI is still unproven, and marketers are justified in hesitating to invest.
But is it possible we’re not giving AI a fair shot? It’s possible the marketing industry needs to invest more time and money into AI before it can help us to improve our content.
Give Your Relationship with AI Some Time
Realizing great AI-assisted content requires investing time into the tool.
Kate Bradley Chernis knows all about that. Chernis is the founder and CEO at Lately, an AI-based content generation platform creating dozens of pretested social posts to promote your brand’s longform content. Kate shared this with us: “If artificial intelligence was a human, it would be about three months old. It can’t sit up on its own, can’t feed itself, can’t do a lot of things. It requires human intervention to even exist. Without humans, it's just automation—we have to guide the AI along in the process.”
Laura Smous is the VP of Product Marketing at Verblio, a content creation marketplace and platform powered by human writers. We asked Laura about how writers should be using AI and she assures us that “There are a ton of places where AI can provide a really great assist, but it’s not replacing humans in the way that people fear.”
So, will AI take content marketers’ and copywriters’ jobs? Never. AI has major limitations. That said, there is no doubt it is quickly finding its way into the content production process. Marketers who don’t start experimenting and discovering the value AI can bring to their content could be disadvantaged.
Paul Roetzer, founder of the Artificial Intelligence Marketing Institute, forecasts that “A lot of marketers are going to sit back and in three years think ‘wow, this software is way better than it was.’ Then there’s going to be a segment of marketers who understand the potential of more intelligent software and they’re going to find those tools today and get a multi-year headstart on their peers who are still afraid of the topic.”
So where do marketers begin? How do marketing teams invite AI into their processes? We propose 3 key opportunities:
1. AI-Backed and Data-Backed Research
Market research is time-consuming and expensive—it’s also the least predictable aspect of content creation. It could take 2 hours or it could take 15 minutes. But it’s obviously crucial in providing content that resonates with your audience.
Laura Smous believes, “Content research can be assisted by AI, ensuring that some of the foundational ideas in content are not only backed by data but that they actually come from data as opposed to instinct. We can actually get some validation from AI research before anyone starts writing or looking at a brief.”
When Tomorrow Sleep appeared as a new startup in their market, their own high-quality content was pulling about 4,000 visitors per month. Anxious to scale up their content and connect with their audience, Tomorrow Sleep tapped into multiple AI-backed and non-AI-backed data content research tools. After discovering the topics their audience responded to and what their competitors were doing with these topics, Tomorrow Sleep was ready to launch new content that would rank and resonate with customers. The new AI-informed content resulted in 40,000 monthly site visitors - a 10,000% increase in less than a year.
Want more expert marketing content sent right to your inbox?
Join our newsletter and never miss a beat.
The AI didn’t write any of Tomorrow Sleep’s new content, but it propelled the marketing team in the right direction. Because of the new insights from AI, they could be completely confident in their content strategy, and their remarkable results further justified their research and content.
AI and even some of the newest data-backed tools can identify trends and keywords to focus on, generate topics, uncover what competitors are saying and identify high-value content for your audiences. It can analyze tremendous amounts of data - even open-ended data - with speed and efficiency, delivering key insights to decision makers before a decision is made.
Palomar is 97th Floor’s patent pending software for analyzing contextual, semantic data in real-time. Palomar’s SERP Intelligence crawls through all of the content on the web that competes with your content and after thorough analysis, it will not only tell you what to speak on and how to speak about it.
Another essential tool for marketers is SparkToro (founded by Rand Fishkin, original founder of Moz). SparkToro aggregates the most comprehensive overview of audience data on the internet revealing demographics, behavioral traits, topics discussed publicly online, and other key data points so that we can pick up on how our audience thinks, what they consume and ultimately how to help them purchase intelligently.
97th Floor recently took on a client facing unfounded public criticism and negative press, desperately in need of reputation management. We learned from Semantic Analysis in Palomar that a specific thought leader’s writing was negatively impacting public sentiment. We learned from SparkToro where specifically our audience was consuming this content. Our content teams knew that in order to rank on this issue and correct the misinformation, we had to debunk what was coming from these sources. Over the course of ten months, this research-backed content helped pull our client towards a positive public sentiment. Without this intelligence, our content could not have correctly identified and addressed the issues threatening our client.
AI or not, marketers are severely under-leveraging the tools and data available to them.
Do This: Let data-backed tools analyze data and deliver insights to you. Don’t shy away from this bias-free, super-efficient way to discover the seeds in your data that lead to golden content. Spend your time strategizing around reliable data, not finding it.
Tools to Try:
2. Defeat "the blank page"
“Humans are bad at getting started. They’re bad at doing that first step towards that task.” That’s Laura Smous again and we’re all feeling quite seen by her comment. And maybe a little relieved that other humans are also like this.
Getting down an outline, a first draft, a content brief - going from nothing to something - can be daunting. But if an AI cranks out that first piece of writing for you, you can start acting as editor and creative, launching off of that writing into something more exciting without losing hours watching your cursor blink on an empty page.
The Associated Press was one of the first news organizations to use AI in reporting by integrating AI for news gathering, production and distribution beginning in 2014. By allowing AI to help draft content and amp up volume, AP reporters had more time to “experiment with new projects and establish thought leadership.”
Rachel Bascom shares, “In the past 9 years at 97th Floor, I’ve written a lot of content on a lot of different topics. Ten minutes can very quickly turn into thirty minutes or an hour when you’re just struggling to get started. Using AI-generated content as something to start with would make a huge difference.”
Where would you spend that 10 minutes? What about 30 minutes? An hour? What new projects would you start?
Do This: Add an AI draft to your workflow. Let a cold machine write something bad. Then turn that into something great.
Tools to Try:
3. Continually "consult" the AI
Yes, AI can speed up our processes—but we want more from it than efficiency. We want AI to help us create better content. There’s so much discussion about improving AI content, but could AI also teach us a few things? We posed this idea to Kate Bradley Chernis, and she shared two cases where Lately’s AI did just that.
One Lately client fed his blog into Lately’s AI and was horrified by the social posts he got back. Rather than condemning the machine for producing bad content, the client went to the Lately word cloud associated with his writing, examining which words resonate with his audiences across every timeframe, channel and campaign. He realized that his blog post was just bad. With new intel from Lately’s AI, this client rewrote a more focused blog post. The Lately social posts that came from this new content were spot on.
Gary Vaynerchuk, now one of Lately’s advisors, tested Lately by having his team create an entire Twitter channel (@garyveetv) with hours and hours of content they pumped into it. Initially, Lately’s content boosted the channel's engagement by 12,000%.
Vaynerchuk’s team also reported an 80% agreement between the quotes Lately pulled and what they would have pulled themselves. His team went back and forth with the AI, feeding Lately tons of content to learn from and then comparing their own content with whatever Lately produced for wildly successful results.
Both Lately clients consulted the AI to make sure their content was on track with their marketing goals, using each interaction as a data point to guide the content forward.
Some AI services like Grammarly can edit copy, checking for grammar, spelling and weak writing. Marketers should also consider AI that analyzes for consistency in style, tone, terminology and content goals.
Laura Smous admits that “Humans are very bad at consistency. Humans think if they have a script or a pitch that they use that they deliver it the same way every time, or that their follow-up is at the same intervals and we’re actually pretty bad at understanding if we’ve done that.”
AI can help solve this problem in your content, building coherence across all of your content so that your audience recognizes and trusts each piece you create.
Do This: Collaborate with AI. Consider the relationship symbiotic. Check back in while editing and before publishing to make sure your content is consistent, accurate and focused.
Tools to Try:
Give Love AI a Chance
AI models have not yet proven themselves to be a sure-fire investment of marketers’ time and money. Content AI isn’t quite where we want it to be now, but maybe that future dream of AI-assisted content collaboration is only realized by marketers who will put in a little more time and a little more money and a little more feedback. It is called machine learning, right?
Either way, marketers are not leveraging existing tools (AI and non-AI) nearly as much as they should be and the only way AI will ruin jobs is if marketers don’t begin learning how to collaborate with it.