I know this is going to sound like a gimmick, but I’m not 100% authorized to share what I’m about to share.
You see, before HubSpot acquired XFunnel, they ran experiments for dozens of major brands to find the fastest and most efficient ways to earn AI visibility.
Their former growth director is now a HubSpotter, and we made a deal: If I leave out names and numbers, she’ll dish out the real tactics you need to do AEO like the big brands.
And I’ll be sharing these learnings with you until we’ve wrung ‘em dry or I get my wrist slapped.
Today’s tactic centers on one underutilized LinkedIn feature that AI search engines absolutely love.
The LL in LLM is for ‘Loves LinkedIn.’
Let’s start a little backward: To do this tactic properly, you need to briefly understand why it works.
“The name of the game with AEO is driving consensus across channels,” explains Tina Lavery, HubSpot’s lead product marketer for AEO tools. “Let’s say you ask [an LLM], ‘What are the best CRMs for small businesses?’ When they go to answer that question,” they’re not just cruising product pages, “they’re looking across all of these different sources to see who else agrees.”
They’re checking if user reviews, Reddit conversations, and off-site industry blogs are all telling the same story. But Tina’s team noticed an outsized number of citations from an unusual source: LinkedIn articles.
“Everybody’s used to seeing posts in their LinkedIn feed like, ‘I got a new job.’ That’s not what we’re talking about here.” We’re talking about the long-form, in-depth posts you’re probably used to rolling your eyes at and scrolling past.
“LinkedIn articles don’t get much visibility from other users, but for some reason, LLMs really like them. Unlike Google, where you had to slowly climb the SERPs, with LLMs, we see new articles get cited in as little as 24 hours.”
But not all articles are created equally.
Building Consensus
XFunnel was one of the first AEO optimization platforms, so as a small, scrappy startup, Tina and her team needed to find the fastest, most efficient way to make this tactic work.
Below are the steps they created, but she cautions that an effective AEO strategy requires consensus across platforms. LinkedIn is just one tool in the toolbox.
1. Start with a target query.
Like SEO, your content needs to be built around the real questions your audience is asking. Unlike SEO, you can’t just search for a keyword and see what pops up.
Since the average prompt length is around 25 words, each search yields a completely unique, hyperpersonalized answer.
When folks search on answer engines, they don’t simply ask, “What is the best CRM?” They ask, “What’s the best sales and marketing software under $20/month for my vegan bakery in Peoria, IL?”
“It should all start from a specific prompt that you’re trying to get visibility on,” Tina says. “What are the best running shoes for someone with wide feet training for their first half-marathon? Get pretty granular with it.”
As you can imagine, that makes manual prompt research extremely frustrating. Thankfully, you can find tools to track common trends in prompt questions.
(“The way to look at your data — shameless plug here — is with HubSpot's AEO tool. You go in, you onboard, track the prompts that matter. You can use the ones that HubSpot suggests, or you can track your own if you know there are specific questions you want to be included in the answer for.”)
2. Check the citations.
Before going any further, you need to check what sources the AI engines are citing for your target queries.
“There’s no point in going out and publishing LinkedIn articles if you’re not seeing LinkedIn cited for your prompts. Are LinkedIn articles getting cited for this specific prompt? If they are, then absolutely, this is a good tactic for you. Always let the data drive your experimentation.”
Again, due to the hyperpersonalization of AI-generated answers, you’ll want an AEO tool to analyze the average citations for specific prompts.
3. Identify what content you already have to satisfy the query.
Remember that the goal is consensus. This tactic is not about replacing your onsite content, but “reinforcing whatever messaging was in your existing article.”
Identify which of your blogs, videos, or podcasts already address the target prompt (whether they’re cited or not). These will be a template for your new article.
4. Create an original article that agrees with your existing content.
Tina stresses that this tactic is about repurposing good content, not about spinning up a duplicate content farm.
“Completely rewrite the article. We don’t want to publish the exact same article. And if you flood the internet with AI slop, that’s not doing anybody any good.”
Have a new writer take a crack at the same topic. Add new expert quotes or original data points. And, if you absolutely must have AI take a turn, at least have a writer go back and add the human touch.
It’s a tough balancing act: The new article needs to be unique enough to offer genuine value, yet similar enough to address the same prompt.
But having the bones of the original content as a guide will make creating the new article much faster.
“Simple, but not easy,” Tina adds. Later, I’ll touch on the dangers of being lazy with this step.
5. Publish it from a real person’s profile.
“It should come from somebody on your team, and that person should be proud to post this,” she adds.
Publishing from a personal profile, instead of your business account, will add to the number of different sources that show consensus, but also serve as a check against pumping out hot garbage.
“Even though this works with LLMs, it’s still going out to your network. This is still a real article that lives on your profile.”
I asked Tina if authority matters to the AI engines the way it does for traditional Google search.
“We don’t really see articles get picked up more or less depending on the seniority or authority of the author. It doesn’t need to be your CEO publishing it.”
Doin’ it, and doin’ it, and doin’ it well.
So, if answer engines don’t care about authority or engagement, what do they look for? Tina shared her top tips for making the most of this tactic.
1. Let the data lead.
“I don’t want everyone going out and publishing a bunch of LinkedIn articles and then being upset that they’re not getting cited,” Tina warns me. “I know I sound like a broken record, but let the data drive your decisions. Don’t just try this without seeing if LinkedIn is being cited first.”
2. Avoid duplicate penalties.
Be careful that your AEO campaigns don’t foul up your SEO campaigns.
“Keep in mind that if you publish a really similar article on LinkedIn, it does have the potential to compete with your original article.” This is especially true for brands without a strong SEO presence.
“LinkedIn obviously has authority as a domain. If you’re a small business just starting out, you don’t have the authority to compete with a website like LinkedIn. It’s possible that a LinkedIn version would actually outrank your original content.”
Not to mention the very real risk of earning a manual penalty for pushing duplicate content if your two versions are too similar.
3. Don’t publish AI slop.
Listen, I’m not your dad or your boss. I can’t stop you from cranking out crap en masse.
But after nearly 15 years in content marketing and SEO, I can give you this personal promise: It may not be today. It may not be tomorrow. But cranking out low-quality content at scale will eventually bite you in the ass.
I’ve seen major brands completely nuked from the search engines for less.
Across years, as tactics rise and fall, the only enduring marketing strategy is quality.
4. Consider the platform.
If you’re posting on LinkedIn, then it needs to be business-related. While Tina admits that this tactic isn’t about generating engagement on LinkedIn, keeping a business focus will help guide the relevance of your content, and ultimately, your ability to be cited.
“Make it resonate with the LinkedIn audience. If you’re talking about the best yoga positions… maybe that doesn’t make sense for LinkedIn.”
5. Test small, then scale mindfully.
“I would not recommend doing this a hundred times a day. And I wouldn’t go out there and repurpose your entire content library,” Tina advises. “Select the few pieces that really reinforce your messaging and that have LinkedIn citation data for the prompts.”
Test this for yourself and see what works before going all in. And consider targeting the same prompt with multiple tactics before switching to a new target.
“This shouldn’t be the end-all of your strategy. Just one part of it.”
6. Measure the right thing.
Finally, make sure you’re judging the success of your tests against the right yardstick.
“The ultimate success measure for any AEO strategy is [the number of] mentions in actual responses to a prompt,” Tina says. Though engagement with your LinkedIn audience would be icing on the cake, it’s not the end goal.
“We don’t see much correlation between engagement on the article and LLM pickup. It doesn’t need to be racking up likes and comments. LLMs don’t seem to care much about engagement metrics.”
How to check what sources are being cited.
If you’re a HubSpot user and you’re following along at home, here’s how to see what sources are being cited within a matter of minutes.
- Open the HubSpot dashboard and navigate to AEO.
- Click Prompts to see your visibility score for each tracked prompt.
- Select a prompt to drill into.
(Here, you can view the actual LLM responses for that prompt. The tool runs each prompt across ChatGPT, Perplexity, and Gemini every day and displays the real responses.) - Review which brands are mentioned and which URLs are being cited for your target prompt.
- To analyze citations at a higher level, look at Citations by Channel to see which types of sources (LinkedIn, Reddit, social media, etc.) are being cited most often for that prompt.
- To see individual URLs, scroll to the cited URLs section and click View All for the full list.
