The Effectiveness of AI in B2B Thought Leadership Marketing.
Thought leadership is an essential part of any B2B marketing strategy. That's why successful companies inevitably have a blog section that is densely populated with content relevant to their industry. In the old days before AI, this used to be a LOT of effort. Companies had to have a dedicated researcher who would research relevant topics, write articles, obtain necessary approvals, and upload them to the company's website. One post a week was the most that a company could realistically handle.
But the game has changed. With ChatGPT and other generative AI tools, topics can be generated in seconds, and an article can be written equally as fast. So now it's quite possible to post a blog post or more every day. Many companies have figured this out and are doing exactly that.
So, how has this affected the effectiveness of thought leadership marketing in B2B companies?
Given the topic, I couldn’t resist asking ChatGPT, “How effective is AI-written content in thought leadership marketing for B2B companies?” The answer I got was eerily similar to this article's thesis, which doesn’t make it any less true. ChatGPT said that it was effective at delivering content cheaply and rapidly, making keyword-rich content for SEO, and supporting thought leadership campaigns. She admitted that the pitfalls of using ChatGPT (in this case) were: lack of original insight, credibility risks, and brand voice inconsistency.
Google recently made a change to its search algorithm in its March 2024 Core Update. The goal was to “reduce unhelpful, unoriginal content”. They never explicitly said this was aimed at deprioritizing AI-written content; in fact, they’ve explicitly said they don’t penalize content just because it was written by a robot. However, since they do prioritize originality, which AI does not do well, it's safe to assume that human-written content can have an advantage over its digital competitors.
There are pages and pages of articles on this very topic. So much so that I’m starting to doubt whether the world really needs one more. However, they all seem to converge on the same conclusion as ChatGPT.
CodeStringers, as a B2B company with a marketing program, also has a thought leadership program. We have a whitepapers section, which is human-written, and a blog section, which is a mix of AI-written, AI-assisted, and human-written content.
The human content is created by me, our CEO, or one of our other subject-matter experts. The AI-written content is generated using ChatGPT. The AI-assisted content is created in a hybrid– first a human creates an outline, then ChatGPT writes the first draft, then a human edits it to make it sound more consistent and “warmer”.
For what it’s worth, we recently tried to find an AI tool that could make our content sound warmer and “fool” the AI detectors, but we found it wasn’t worth it. As of the time of this writing, none of them are sophisticated enough to generate original or engaging content as well as a human can.
Since I have access to Google Analytics reporting for our internal content, I decided to look at which type of content had the highest website performance, as measured by pageviews.
Here are the results (March 1 - May 29, 2025) for the thought leadership content:
Author | % of page views |
Human | 69% |
AI-assisted | 32% |
AI-written | 10% |
I learned a few things from this exercise that I’d like to share with the world:
There is value in all three types of content. Human content performs the best but it's the toughest to create, AI-written content performs the worst but it's the easiest to create, and AI-assisted falls somewhere in the middle.
Consequently, it's best to employ a combination of the three methods in your thought leadership marketing program.
Use humans to create the content that is actual thought leadership, meaning the content that you want to show the world as representative of your organization’s original thoughts. Use AI to create your SEO-spam. Use a combination of the two for everything else.
This is the same conclusion that ChatGPT gave me.
There are a bunch of articles online that drew this same conclusion.
As a result of #3 and #4, I’m questioning whether the world needs one more article about this topic.
But I’ll post it anyway.
Can you guess what kind of content this article is? Was it human, or AI written?
Watch the 6-minute guided demo, then book a walkthrough: we read your CRM, ad accounts and content and show what a governed system would do differently. You pay only if you proceed. Or see how we approach it.
More on the same problem:
Content that states a position the company is willing to defend, drawn from its own experience: what it has seen work, what it disagrees with, what it would advise a buyer to do. A summary of what everyone already says is content, not thought leadership. The test is whether a competitor could have published the same piece; if so, it carries no signal about who you are.
Not for being AI-generated. Search engines say they rank content on helpfulness and originality, whoever wrote it. In practice, AI-written pieces that restate common knowledge tend to rank poorly because many near-identical pages already exist. Content with original data, specific examples and a clear point of view ranks better, and that is harder to produce with a generic prompt.
Let people write the pieces that carry the company's judgment: opinions, case experience, advice. Use an AI model to draft explanatory and definitional pieces from a human outline, then edit for accuracy and voice. Use it freely for research, outlines and first drafts of anything. Keep a human responsible for every published claim, because the model does not know what is true about your business.
Pageviews show reach but not effect. Track time on page, return visits, newsletter signups from the piece, and whether prospects mention it in sales conversations. Tag inbound leads with the content they read. Compare human, AI-assisted and AI-written pieces on the same measures over a quarter. The pieces that produce conversations are the ones to make more of, whatever they cost.
Not reliably. Detectors produce false positives on plain human writing and miss edited AI output, so they are not a sound basis for editorial decisions. The better question is whether the piece says something specific and true that a reader could not get elsewhere. Readers notice generic content regardless of what a detector reports.
"Something isn't working." "We're outgrowing our tools." "We're afraid to make the wrong move." No-Risk Discovery is a short, practical conversation that gets you clarity before you commit to anything big. We'll tell you if we're a fit. If we're not, we'll tell you that too.
Say who you are and ask. A person replies within one business day.