Humans have been skeptical of most large shifts in technology from the steam engine, automobiles, typewriters, the telephone, and of course generative AI and everything in between.
With the release of GPT4, the hype around generative AI is unlikely to die down.
As marketing professionals, we’re often faced with the nebulous question, “What does this mean for us?”
Clients and peers alike have been asking this question and discussing what the future of marketing, particularly content and SEO, will look like and how they should adapt.
Once we dug deeper, we identified three different questions wrapped up into one. We’ll break them down and share our position here:
- How do generative AI tools impact how marketers produce content? How does it change how Omniscient Digital produces content?
- What is the impact of large language models (LLMs) generative AI tools on our jobs?
- How do Google Bard and Microsoft Bing + ChatGPT impact people’s search behaviors? What does that mean for content and SEO marketers?
While they’re all related, let’s address them individually.
How do generative AI tools impact how marketers produce content?
We’ll start here because we’re a content and SEO agency and thus have been asked this question many times in the last few months.
Generative AI tools like Jasper and ChatGPT are just that, tools. They enable us to do more of what we already set out to do, more efficiently.
Here’s John Carmack, founder of Keen Technologies, building artificial general intelligence (AGI) reminding us of this:
You can give an amateur athlete the most advanced, expensive equipment, but the equipment doesn’t make an athlete.
Tools influence the “how to do” more than it influences the “what to do.” The “what” is the strategic piece that requires a human being to assess tradeoffs and risk.
Once you’ve defined a strategy and what content needs to be produced, generative AI can help you produce that content. How it helps is up to you.
As the saying goes, “Cheap, fast, and good. You get two.”
Generative AI helps you go fast.
If you have a low bar for quality and want to produce a lot of content in a short time, it can also help you produce content cheaply. But what if you have standards for quality?
While generative AI tools alone are unlikely to produce good content that resonates with readers, they can help you produce more high-quality content. It’s just not cheap because, whether you outsource it or not, it requires an editorial touch—the application of hard-earned, human expertise and skill.
So how do we use generative AI tools to produce great content?
Let’s start by looking at the stages of producing an article:
- Ideation – You come up with an idea
- Research – Research the idea and find supporting stats, arguments, and counterarguments
- Composition – The actual writing
- Editing – Polishing, strengthening the writing, adhering to brand voice
- Distribution – Making sure people can find the content
For purely human-generated content, it’s not uncommon that our time is mostly spent at the composition stage. We have deadlines to meet and goals to hit. We can’t spend too much time researching and editing. We need to publish.
Generative AI can help us get part of the way to producing a great piece of content.
But a human needs to be involved to make sure the content empathizes with readers, delivers a unique point of view, represents the brand it’s written for, weaves in data analysis, and relates to the customer journey.
Most importantly, a human is needed to verify facts and sources.
Generative AI can help us produce more great content faster but, until it can do everything a human writer and editor can do, not necessarily cheaper.
Podcast of us talking about generative AI tools and why it might actually be great for the industry:
What is the impact of generative AI tools on our jobs?
This is the more interesting question. These tools are going to make us significantly more efficient and effective, maybe even enable us to be more creative.
Here’s an example query I asked ChatGPT via our Slack integration.
Remember that tools enable us to be more efficient and effective. They help us do what we already seek to do.
A hammer helps us accomplish a task we had already set out to do more efficiently: drive a nail into two pieces of wood to connect them.
The hammer doesn’t tell us to connect the two pieces of wood.
Similarly, AI helps us accomplish what we set out to do: write a high-quality blog post that generates leads.
Sure, we can ask these tools to give us some ideas, but as we all have likely seen, most ideas aren’t great. We still need to sift through, prompt, and apply our own knowledge and experience to get what we need.
The AI doesn’t tell us what questions to ask to figure out the right content to produce. It doesn’t tell us the puzzle we’re trying to solve.
So here are some examples of how we’ve used generative AI in our work:
- Build domain models
- Give a starting point for a content brief
- Analyze search results
- Generate headline ideas
- Help rewrite our crappy first drafts (show me a writer who writes amazing first drafts)
- Generate keyword ideas
- Summarize and repurpose content for different formats and channels
- Update and optimize content to increase traffic
How might Google Bard and Microsoft Bing + ChatGPT impact consumer search behavior?
This is the most speculative of the questions and seems to be what’s got marketers shaking in their boots.
Our take: Too early to call and therefore too early to react.
These tools are also highly flawed and subject to hundreds and thousands of iterations over the next years.
Microsoft has limited the number of questions you can ask the Bing chatbot per session. Since we began drafting this document, they also limited you to 50 total sessions per day.
Google’s Bard famously flopped its first demo. While working on this draft, they’ve introduced generative AI in Google Workspaces.
Since drafting this (I swear we haven’t been spending that long drafting, there’s just so much happening!) GPT4 was also announced.
We can’t predict what’s going to happen over the next few years, months, or even weeks.
When it comes to search behavior here are a few questions to consider that need to be answered before we can meaningfully react:
- How will these generative AI tools cite their sources? Will they link to them?
- If these models analyze the entire corpus of the internet to present information, who owns the content these AI chat interfaces present?
- What system will be implemented to give credit to the sources?
- Are stack-ranked lists of links still going to be a thing?
- What does it mean for Google’s Ad platform?
- Will “Bing it” replace “Google it”?
- Most people aren’t using AI tools yet (yes, we’re in a tech bubble), when will broad adoption occur?
There are more questions than answers right now.
As content marketers and SEOs, we wouldn’t try to build a strategy based on speculation on a technology that’s changing literally every day.
For now, focus on what you can control.
Crystallize your messaging and positioning, produce high-quality, helpful content, and keep experimenting with these tools.
