Testing ChatGPT by OpenAI for Keyword Research

Automation is taking over marketing in a big way. Companies using marketing automation platforms has become the norm, with even tiny businesses adopting enterprise technologies. By now, mid-sized and large companies are mostly automating some of their marketing outreach, especially customer and prospect communications.

In 2023, there’s obviously a shift underway. Marketing leaders are becoming increasingly fixated on automating as many tasks as possible. While some marketers fear automation could one day take over their jobs, others see it as more of an opportunity. The mass adoption of marketing automation platforms like Pardot, Marketo, Eloqua and Hubspot so far hasn’t translated to reduced headcount. It has, however, made technical proficiency a requirement for the field.

Leveraging AI and machine learning for applications like content development and keyword research is hardly new, but the tools to accomplish them are becoming mainstream. The technology has mostly been used by spammers in recent years but honestly that’s where innovation always happens.

What is ChatGPT?

Launched in November 2022, ChatGPT is OpenAI Generative Pre-trained Transformer chatbot technology. What differentiates it from other chatbot technologies is its conversational nature and detailed responses. It is programmed to maintain context and provide more relevant answers the more questions that are asked and instructions given.

OpenAI’s value ballooned to $29 billion when ChatGPT’s research release was unveiled. This is either a flash in the pan or something you’ll be hearing about all year. Google is reportedly so shook that it’s rushing a similar product to market. Meanwhile, the media has started fear mongering about generative AI torpedoing everyone’s careers. The Atlantic has the hottest take so far: “How ChatGPT Will Destabilize White Collar Work.”

All that said, ChatGPT has a ton of potential that could be disruptive in ways that are uncomfortable for some. For others, it could free up time to focus on strategy and optimizing processes for revenue.

Keyword Research

Having spent some time chatting with the application like a lunatic, I’m pretty impressed. The intelligence I got in just a few minutes would’ve taken many Google searches.

Let’s say I’m tasked with developing and executing a strategy for a hyperautomation startup. To get an idea of what that actually means, I’d normally start with Wikipedia and expand from there before selecting keywords. It could take hours or days depending how foreign the concept is to me. This eats up valuable billable hours that could be better used securing quick wins to bolster confidence with the client.

If I was using Chat AI, I’d start with the obvious: What is hyperautomation?

That’s pretty good. If I was developing SEO-friendly copy, that’s a very useful chunk of text.

But what keywords are there?

That’s a great starter list. The inclusion of acronyms is especially helpful. So next I asked it to hit me with a list of intelligence automation keywords:

This is gold. To improve targeting even more, I asked which industries leverage hyperautomation:

This is all stuff that could accelerate a paid search campaign’s results. The list of industries could be used to create modifier keywords for the main keyword set.

How does it compare to popular keyword tools?

Honestly, I already prefer it.

I’ve never been impressed with any of the keyword research tools I’ve used. They either provide highly questionable results (especially Google’s Keyword Planner) or they fail to provide context. SEMRush, Moz and other platforms can be useful but their keyword and content topic suggestions are just not scalable and often irrelevant. If you’re in a niche industry, the suggestions become laughable.

At best, the mainstay keyword tools are decent guides. While most of them have APIs, they just don’t have the same potential as OpenAI’s.

How can this be automated?

Here’s an example of how the OpenAI API could be used to enhance paid search campaigns:

  1. A search keyword meets a certain threshold (conversion rate, for example)
  2. ChatGPT is asked for a list of similar keywords or industries
  3. The keywords and modifiers are added to the campaign

Some other paid search applications I can think of:

  • Automatically populate search ad descriptions
  • Insert text onto landing pages with low quality scores (also useful for SEO)

That’s where this becomes uncomfortable for some marketers. Brand standards shouldn’t go out the window, but they shouldn’t be a hindrance either. Like it or not, you’ve probably read your share of AI generated content and it’s no longer easy to detect. As for job security concerns, this still requires oversight.

As long as you’re not ridiculous about it, this should be a way to enhance your capabilities, not devalue them.

Latest blog posts