Are we writing for people or bots? Rethinking intranet content and analytics in the age of AI
Intranets used to be simple. You published a page, someone read it, and you measured how many clicks it got. But things are moving quickly.
Today, with AI-powered tools everywhere across the digital workplace, employees don’t always have time to visit intranet content directly. Increasingly, people can now access an intranet chatbot to ask a question, and receive an instant answer based on content it finds on your intranet.
This changes everything. It forces us to ask two big questions. Are we writing for humans, or for machines who then re-processes the content and deliver it to humans? And are we then measuring the right things?
Content is now serving two purposes
Most intranet content still fits into two main categories.
There are structured content pages, like policies, procedures, help guides, and FAQs. These are factual, evergreen, and built to support daily work.
Then there are news articles and updates, like leadership messages, employee stories, or internal campaigns. These are designed to connect, inform, and inspire. They give employees the bigger picture, show culture in action, and explain the “why” behind decisions.
This distinction remains critical. Structured pages are the backbone that supports work. News stories build engagement and trust.
Enter the age of hyper-personalisation
With AI in the mix, we’re seeing something new. Instead of employees having to read the same intranet article, AI can now tailor information for each person, based on their role, location, or interests.
That means internal communicators might start producing two deliberate versions of the same message.
The official “company all” version, which is the crisp, approved story that goes on the record. It is aligned with brand, checked by legal, and is consistent across the entire organisation. This is what lands in inboxes, sits on the homepage, or gets referenced at town halls.
A richer “AI source” version, which is more detailed and structured. It might include extra context, common questions and answers, examples for different departments, or scenarios that give AI the material needed to personalise responses.
For example, the official announcement might say;
“Our new hybrid work policy begins 1 August. It gives teams more flexibility while supporting collaboration. Talk to your manager about what this means for you.”
The AI source version might add;
“This means:
• Sales teams can continue remote client meetings but must log location in CRM.
• Frontline staff are expected on-site, with local exceptions.
• International employees may adjust hours to overlap with HQ.
• Managers get updated one-on-one guides next week.
• FAQs cover equipment stipends, meeting etiquette, and travel approvals.”
This richer material allows the intranet bot to respond accurately.
So, when a sales rep asks; “Do I still have to come into the office for client meetings?” the bot has the information needed to supply an accurate answer. When a frontline worker asks, they get different guidance from the AI’s answer.
Writing for both humans and bots
So, are we writing for people, or for bots? The reality is we are writing for people who might reach our content through a bot.
This does not mean making everything robotic. It means crafting content that is rich enough for people to care about, while also structured enough for AI to personalise and deliver in new ways.
For structured content, clarity and consistency matter more than ever. The cleaner and better organised these pages are, the easier it is for AI to pull the right answer.
For news and stories, the goal is still emotional and cultural connection. These are written for people first. They help employees understand strategy, feel recognised, and see themselves in the company’s journey.
It is not one or the other. Communicators keep telling powerful stories. At the same time, they also build the deeper, structured resources that allow AI to personalise and scale those stories.
Why linking news and content matters
There is one catch. Often, news stories also contain factual information. A piece about a new parental leave policy, for example, needs to link to a structured page that lays out eligibility, process, and contact details.
If these are not connected, problems can arise. Bots might give outdated or inconsistent answers. Employees could get frustrated, slow down their work, or make the wrong decisions.
The best approach is to make sure every news story that touches on a process or rule links clearly to a structured, up-to-date reference page. That way, the story provides human context and the reference page gives factual certainty, both for employees reading directly and for bots pulling answers.
How measurement needs to evolve
All of this changes how we measure success.
For years, intranet analytics focused on page views, unique visitors, time on page, and search terms. These are still useful, but they only tell half the story.
As the intranet bot overtakes the direct intranet access in usage, and becomes the primary way employees access content, we also need to measure how well that ecosystem is working.
Look at chat inputs. What exactly are employees asking the bot? These are honest signals about what they care about or struggle to find.
Review bot response quality. Are the answers accurate and helpful, or is unclear content causing problems? Imagine the consequences of the intranet bot using an outdated policy to confidently tell employees about entitlements that no longer exist.
Track source usage. Which pages is the bot actually using to build its answers? Some of your most important content might have few direct clicks, yet power hundreds of chatbot interactions.
For news and stories, keep watching engagement metrics like views, comments, shares, and sentiment. These show whether people are connecting with your messages on a human level.
At SWOOP Analytics, we are working to support this bigger picture. We provide powerful SharePoint intranet analytics that reveal how employees navigate, what they read, and how they engage with news. Now we are developing ways to complement this with intranet bot analytics, so you can see what questions people are asking, which pages the AI relies on to answer, and how effective those answers are. This will give organisations a far more complete view of how structured content, and news stories, drive the digital employee experience.
Final thoughts
We are not writing for bots instead of people. We are writing for people who might reach our content through a bot.
Keep your official stories clear, consistent, and on brand. Build richer, structured content that gives AI what it needs to personalise and scale. Link everything together so employees, and the bots that serve them, always have the full picture.
Measure both direct engagement and these new AI-driven interactions.
Let the bots handle the delivery. You still own the connection.
