Your Internal Communication Isn't Being Ignored. It's Being Processed in Ways You Didn't Design For.

There's a difference. And it changes everything about how you write.

New research dropped recently that I haven't been able to stop thinking about.

Fresh Intranet commissioned an independent study of 1,000 professionals across the UK and US — 500 internal comms professionals, 500 employees — to understand what actually happens between the moment a message is sent and the moment it lands. They called it the Employee Attention Recession.

I'd call it confirmation of something most comms leaders already feel but can't quite name.

The number that should reframe everything

Only 12% of employees read internal communications in full.

But here's the twist: this isn't disengagement. 91% of employees say internal comms feel relevant to them most or all of the time. They want to engage. The problem is volume — 83% say they receive too much internal content, 35% say far too much to keep up.

Your message isn't being ignored. It's arriving in an already-full room and being dealt with accordingly.

What "dealt with" means now

Here's where it gets more interesting — and more uncomfortable.

Employees aren't just skimming — increasingly, they're handing your message to AI. 88% have used AI to summarise internal communications, 58% do it regularly, and AI-generated summaries were the most common way employees first encountered company news — ahead of the intranet, ahead of email from leadership, ahead of Teams.

The message that leaves your hands and the one that arrives with your audience are no longer necessarily the same message.

The researchers call this the AI Visibility Gap, and it sits inside a wider Intent

Gap: 95% of employees are confident AI summarises their comms accurately, while 92% of IC professionals worry AI summaries and distort the meaning or tone. Neither side has real visibility into what's happening. Both are operating on assumptions.

This is not a new problem in Internal Communication.

I've spent years pointing out that your message gets filtered through a multi-audience, skimmed between meetings, compressed into one line, lost in a Slack thread about lunch. What's new is a second filter stacked on the first.

Before, the risk was one reader with partial attention filling the gaps with their own interpretation. Now there's also AI doing the same thing, faster and at scale. Your message isn't processed by one reader anymore — it's processed by a sequence of filters, none of which were designed with your intent in mind.

Before, the risk was human summarisation, someone reading half your message and filling the gaps with their own interpretation.

Now, there's also AI summarisation, faster, more scalable, and just as capable of compressing away the thing that mattered most.

Your message is no longer being processed by one reader with partial attention. It's being processed by multiple filters, in sequence, none of which were designed with your intent in mind.

The gap nobody's measuring

Ask IC professionals how they track performance and you'll hear: employee surveys, engagement metrics, AI dashboards, intranet analytics. Notice what's missing — any direct measure of whether the content was understood. Gallagher's 2026 State of the Sector report found the same thing: 70% of IC professionals are still tracking outputs, not outcomes.

42% believe their communications land well but have limited evidence to prove it. Only 33% have strong evidence content is understood as intended. That gap — between believing it landed and knowing it landed — is where most messaging problems live.

The solution isn't more writing. It's different architecture.

The instinctive response is to rewrite: shorter, cleaner, less jargon, a stronger subject line. Then it goes out into the same broken journey and gets the same result. More content doesn't solve an attention problem — it compounds one.

The single biggest factor in whether an employee understands something isn't who sent it, but how well it was designed.

Content that survives this environment — human skimming, AI compression, the manager who forwarded it with a one-line summary does a few things.

It leads with the most important information, makes intent explicit instead of relying on inferred tone, and uses structure so the point survives compression, whether that compression comes from a skimming reader, or an AI summary.

Before your next send, ask two questions

The message you sent is not the message they received. That's the multi-

audience messaging problem — and it has a solution.

What's the one thing this message must survive? If it gets reduced to a single sentence by an AI, or someone forwarding it on — what does that sentence need to contain?

Are you measuring understanding, or just delivery? Open rates tell you the message left the room. They don't tell you what happened to it on the way.

The full Employee Attention Recession report is free to download from Fresh Intranet. Research conducted with 1,000 UK and US professionals by Fresh Intranet, 2026.

If you're a communications leader whose messages keep leaving the room and not surviving the journey, let's fix that.