If you’re optimizing for ChatGPT and other LLMs, it’s best to use the name of a real person when publishing blog articles on a company’s website.
One of my clients recently asked which was better: using a real person’s name as the author of a blog article, or simply listing the company’s name?
This is a great question because blog articles are one of the best ways to quickly build up your EEAT (Experience, Expertise, Authoritativeness, and Trustworthiness) on a topic. I like blog articles for answer engine optimization (AEO) because they’re conversational, can be hyper-focused on answering a single, well-defined question, and don’t take long to write. So when you’re working in AEO or SEO, this is a question that should come up all the time. That said, it was surprisingly difficult to find a straight answer!
I definitely had a preference. My gut feeling was that attributing the article to a real person would be better for AEO, since LLMs like citing sources with defined expertise in a topic. Sure, brands can be associated with a topic, but can they really be experts? Real people can give keynotes, teach classes, write for industry publications, and have unique conversations on LinkedIn.
But was I right? It took some digging to find out.
Marketers are split on which is better
The first thing I did was pop a question into Google: “Should my blog author be a person or a brand?” I crossed my fingers for a clear answer from Google’s AI Overview.
I immediately remembered that LLMs don’t always like to take sides. In fact, this is a trend I’ve been seeing a lot more lately. I have no evidence to know if it’s true or just my experience, but Google AI Overviews and ChatGPT seem to be getting more diplomatic in their responses, often careful to offer a pro and a con side for questions like these instead of a decisive response.
So I scrolled down to look at the sources and see what my fellow marketing practitioners were saying. Opinions were split, and no one seemed to have any data to back up the answer. And weirdly, most of the top results were from at least two years ago. For a question that seemed pretty basic, I was surprised there wasn’t already a best practice figured out.
ChatGPT saves the day…kind of
What did I do next? I went to ChatGPT, of course.
You might think I use ChatGPT all day, every day, based on the kind of work that I do. But I usually stick to it for research or generating boilerplate language for things like client bids and business planning. It’s great at research (even if it does sometimes hallucinate), and you’ll see why in a minute.
One of my favorite things to do is ask ChatGPT how I can pander to it. So I asked this:
Asking ChatGPT how to optimize for ChatGPT might be a hack, but it’s smart if it works, right? ChatGPT went on to say that it’s best to use a person’s name as the author byline, plus clear brand association. In my case, that would look like: Tehra Peace, Co-Founder of On Marketing.
It also told me that “LLMs and search engines favor identifiable human authors,” and that keeping the company name visible can associate the content with your brand in embeddings and search indices.
But how could I trust that ChatGPT was telling the truth? Here’s the thing: ChatGPT remembers things about you, even from other conversation threads. For example, elsewhere in its answer ChatGPT offered me an example of a byline that used my client’s brand name, likely because I had done research on its product in another thread. Could it have picked up on my bias for human authors from other conversations we previously had?
Was ChatGPT now pandering to me?
LLMs have a bias toward human authorship
So I followed up with another question that I encourage everyone to use to save themselves from hallucinations: Do you have any data or sources?
And that’s how I got to the real jackpot—a 2025 research paper (with multiple humans in the author byline, by the way) that discussed attribution bias in large language models. The study found that when source documents include metadata about the author, LLMs show a bias toward human authorship. Here’s a quote from the paper:
Our findings indicate that metadata of source documents can influence LLMs’ trust, and how they attribute their answers…Moreover, we show that these models carry a bias towards human authorship against LLM authorship: they are more likely to attribute their answers to documents that are explicitly labelled as having been written by humans (even if the documents are actually generated by LLMs).
By the way, that last part makes me cringe. I really don’t advocate for having an LLM write a blog article and then slapping a person’s name on it. The real person will be 10x more insightful than an LLM, and they’ll be more interesting, too.
This study isn’t a perfect match to our work in AEO. Rather than using the consumer-facing version of ChatGPT that we’re familiar with, the researchers set up their own retrieval-augmented generation (RAG) pipeline and prompted LLMs to generate citations to the source documents. Still, it tells us something useful about how these black boxes operate.
When to use a company’s name as the author in a blog article
All that said, before you can use a real person’s name as the author of a blog article, you need to have a real subject matter expert on hand. And no, this can’t just be the head of content marketing. It has to be someone who is knowledgeable about the topic.
It also has to be someone who’s sure to stick around. Imagine building authority and thought leadership for a person who leaves your company a few months later. Yikes! All that work you did to build expertise will now follow your expert right to their next job assignment. Nice for your author, but bad for you.
So, if you need to hedge your bets, it’s a good idea to attribute your article to the company instead of an individual person.
And by the way, if you need help ghostwriting your blog articles, we offer real, human-written content here at On Marketing. Even the busiest experts have enough time to write a few bullets points on a topic, and we can take it from there.