Internal Comms’ “Content Machine” Problem Now Has an AI Layer

Internal communication has spent years arguing it’s more than a content machine, only to be measured like one. New independent research from Fresh Intranet shows AI has inherited that exact same blind spot, and made it harder to see:

Internal communication produces more content and more data than at any point in its history, and it still struggles to answer the one question that actually matters: “Did any of it land the way it was meant to?” That uncertainty isn’t new. What’s changed is what stands between the message and the answer.

Anthony Monks made the case that internal communication tends to focus on guaranteeing transmission, not understanding: that most of what gets counted as measurement – opens, clicks, attendance – tells us whether a message went out, not whether it changed anything. 

New independent research shows why that gap matters more now, not less: a growing share of the audience those metrics are supposed to account for never actually receives the message at all. It instead receives a summary of it, generated by AI, and nobody is measuring what that summary said.

Fresh, the SharePoint-based intranet platform, commissioned research firm Opinion Matters to survey 1,000 professionals across the UK and US in April 2026 – 500 communicators and 500 of the employees who receive what they send. 

The resulting report, The Employee Attention Recession, sets out to measure the distance between what gets published and what actually lands. Fresh calls that distance the “Intent Gap.” Only 12 percent of employees now read a given internal communication in full. 88 percent have used AI to summarize or interpret one; 58 percent do it regularly.

An old argument, a familiar shape

Debbie Braden drew a useful distinction in her piece between linear and non-linear communication. 

Linear work is easy to measure: announce a process, track adoption. Non-linear work, the interpretation, the trust, the meaning employees make of a message when nobody’s watching, is where most of the value sits, and where measurement has always struggled.

Fresh’s data shows that non-linear space has just acquired a new, largely invisible layer. When 20 percent of employees say they typically hand a message to AI before deciding what to do with it, the interpretation Braden described is no longer happening inside the employee’s head alone. It’s happening inside a tool the communication professional never sees.

That should concern the profession for reasons beyond accuracy. 

Emily Hecker’s piece on whether internal communication is a “BS job” argued that the function survives its own worst perceptions, the content factory, the copy-paste-send machine, only when it can point to something a message actually changed. 

That’s a hard case to make when two-thirds of communication pros privately admit they’re operating on belief rather than evidence that their content is understood as intended, and when the version employees encounter first is increasingly an AI compression rather than the initial piece. AI hasn’t created the box-ticker risk Hecker describes. But it has made the box easier to tick and harder to check.

Why this isn’t an employee laziness problem

Fresh’s data shows 83 percent of employees already feel they receive too much internal content, and volume, rather than relevance or format, is the single biggest factor in whether something gets read at all. Handing a message to AI is a rational response to what employees see as an excessive volume of input, and not necessarily about being inattentive or indifferent.

The opportunity underneath the problem

Dan Sodergren has a  more optimistic version of this take: AI’s real value is as a curation tool, filtering and prioritizing so employees aren’t drowning, delivering the right thing at the right time instead of everything at once. His point was that the choice was never AI versus no AI, since employees will use it regardless. Fresh’s numbers confirm exactly that: the curation is already happening, just without communication professionals being anywhere near it. Someone is deciding what employees see first. In most organizations right now, that someone is a tool, working alone.

That’s a genuinely fixable problem, and fixing it sits closer to communication’s core job than anything else AI touches. It means designing content that survives compression instead of losing its point in it. It means treating the post-publication journey of a message, not just its delivery, as something to measure. And it means finally answering the question Monks posed before AI made it more pressing: not whether the message went out, but whether anyone understood it once it landed.

The full Employee Attention Recession report is available here: https://freshintranet.com/ebook/the-employee-attention-recession-report-2026/