It’s the question my CEO asks me every couple of months, usually when he’s looking at our company LinkedIn page’s less-than-dizzying engagement numbers:
“If barely anyone likes or comments, why are we still investing in this?”
And to be fair, it’s a reasonable question. Company pages on LinkedIn are a bit like having a big, shiny billboard, but on a road that doesn’t get much foot traffic. You still keep it clean, up-to-date, and well-designed, even if only a handful of passers-by actually stop to read it.
But instead of mumbling something vague about “brand presence” and “long-term awareness,” I decided to take a different approach. I asked four large language models - ChatGPT, Gemini, Microsoft Copilot, and Perplexity - the exact same question:
When do tools like ChatGPT, Gemini, Perplexity, and Copilot refer to a company LinkedIn page? What kind of information do they surface and for which queries?
It was simple, direct, and not designed to trick them. The results? Let’s just say: some were helpful, some were similar, and one misunderstood the assignment entirely and went down a “how to post on LinkedIn” rabbit hole!
Copilot was the teacher’s pet of the group. Structured bullet points, neat categories, even tool-specific behaviour. It laid out situations where it would reference LinkedIn, from company background queries to leadership lookups, and the exact data it would pull, like employee counts, recent posts, and follower numbers.
Gemini had a similar vibe to Copilot but with a touch more polish, especially for B2B scenarios. It talked about company research, employment information, and lead generation queries. It also pointed out that LinkedIn is often the go-to for real-time updates, like product launches or CEO posts.
ChatGPT took a methodical approach, breaking things into three neat sections:
It also threw in a strategic nugget: your LinkedIn “About” section is often the version of your brand story that AI tools surface first.
And then there was Perplexity. This one didn’t just misunderstand the question; it packed a bag and went on an adventure about using AI to automate LinkedIn content creation, complete with mentions of Google Sheets and Zapier workflows. Interesting? Yes. Relevant? Not exactly what I’d asked for.
Despite their different personalities, three out of the four gave broadly similar answers on when they refer to a company’s LinkedIn page:
1. Company Facts & BackgroundIf your official site is sparse or outdated, AI tools lean heavily on LinkedIn because it’s structured, searchable, and generally up-to-date.
While there’s consensus on the basics, each tool also had quirks:
Different Priorities
Content vs. Data
Some tools treat LinkedIn purely as a data source. Others see it as a source of content, meaning your posts could influence the AI’s summary of you.
Here’s the honest bit:
Even if engagement is low, the game is changing. SEMrush research shows that by 2027, 75% of searches will be through LLMs. And when those LLMs look for the most current, structured snapshot of your business, your LinkedIn page will be one of their go-to sources.
In other words, your company page might not be where conversations happen, but it’s increasingly where first impressions are formed.
Think of your LinkedIn company page less as a place to “go viral” and more as your AI-friendly business card. Keep it updated, keyword-rich, and reflective of your best messaging.
It’s not there to win a popularity contest; it’s there so that when an AI introduces you to your next big lead, it’s quoting the good stuff.
And to my CEO’s question - yes, we’re still posting. Not because there’s a crowd gathered under our billboard every day, but because when the main road eventually gets rerouted past it - and it will - we want to be ready with the best possible message.