The Atlantic Says Reading Is Ending. Your Deliverable Is a Document.

I once spent twenty minutes explaining my leadership theory to a former COO. She listened to all of it, patiently, and then said: "It sounds like the problem is teams are behind schedule and you go in and fix that."

I said yes, that's right.

She said: "Just say that. Don't say any of the other stuff."

I think about that regularly, usually while looking at something long I've just written. She wasn't telling me the twenty minutes were wrong. She was telling me she'd extracted everything she needed in about ninety seconds and the remaining eighteen and a half minutes were for me.

The reader you're assuming

Almost every knowledge business delivers its value in the form of a document. The findings report, the deck, the brief, the strategy memo, the requirements doc. We argue about length and structure and whether to lead with the recommendation, and underneath all of that sits an assumption nobody examines: that on the other end there is a competent, willing reader who will sit still for the thing.

That assumption is getting harder to defend, and there's now workplace data on it rather than just the feeling we all have.

Research summarized by Strategic this year found that just 12% of employees read internal communications in full. The other 88% skim, filter, ignore unless a manager flags it, or hand it to an AI tool and read the summary. Drawing on the IC Index 2026, the same analysis reports that most employees now have ten minutes or fewer per day to engage with internal communications, one in five has no meaningful time at all, and the single most common answer is five minutes a day – a number that's been shrinking every year since 2023.

The finding I keep coming back to, though, is about why. The strongest predictor of whether someone reads a communication in full turned out not to be the subject, or the sender, or the format. It was the cumulative weight of how much arrived before it. Thirty percent of respondents named volume as their top reason for not reading.

Volume doesn't just compete for attention. It destroys it.

Aliteracy, not illiteracy

The cultural backdrop is real too, and worth being precise about.

Rose Horowitch's essay in the August issue of The Atlantic makes the case that the age of reading may turn out to be a short anomaly rather than a permanent condition. The decline, she reports, cuts across age, gender and education – even the groups that historically read most have seen a collapse. New York Times best sellers now have sentences about a third shorter than a century ago. The explosion of textual fragments, as she puts it, has come at the expense of sustained attention to longer works that convey rich and complicated information.

A 2025 iScience study of two decades of American Time Use Survey data, as summarized by Johan Osteyn, found the share of US adults reading for pleasure on a typical day fell from 28% in 2004 to 16% in 2023. Fortune, citing YouGov, reports two in five Americans read no book at all in 2025.

Horowitch is writing about literacy and culture, not about whether your executive team reads your deck. The bridge between those is an argument I'm making, and the word that makes it defensible is aliteracy – the capacity to read fully intact, without the habit or the appetite for doing it deeply. Your CFO can read. Your CFO is not going to read forty pages.

That distinction matters because it changes the fix. If people couldn't read, you'd simplify. Because people can read but won't sustain it, the question is what earns the sustained attention you're asking for.

The part AI makes worse from both ends

Here's what makes this a right-now problem rather than a slow cultural drift.

The capacity to absorb documents is contracting. And in the same eighteen months, the cost of producing a document has fallen to roughly nothing. Anyone can generate a competent forty-page report in an afternoon, so everyone does, and all of it lands in the same inbox belonging to the same person with five minutes a day.

That's the mechanism the IC Index data identified, running at machine speed. Volume destroys attention, and we just industrialized volume.

Which means the individual quality of your document is no longer the thing that determines whether it gets read. Your carefully-built forty pages are competing not against other people's careful work but against sheer arriving mass. You can write the best report in the organization and lose to the eleven mediocre ones that showed up before yours.

What the document was always for

I have a rule about communication that I've made a lot of teams tired of hearing: it has to come from action and lead to action. You communicate what was done, what needs to happen next, and why – and if a piece of communication doesn't have an action on both ends of it, it's decoration.

I break this rule constantly in the direction of over-communicating. On remote teams especially, my instinct is to put everything out there so nobody lacks context, and what I've learned the hard way is that too much information creates its own fog. People can't read it all, and a team drowning in updates isn't better informed than a team with none – it's differently uninformed, and more tired.

The COO was applying that rule to me. My twenty minutes came from theory and led to nothing she could act on. Her one sentence came from what she'd heard and led directly to a decision about whether to keep talking to me.

So the useful question about your deliverable isn't how long it should be. It's what action it comes from and what action it leads to, and whether the person who needs to take that action can find it in the time they'll actually give you.

What I'd actually change

Not "write shorter decks." That's the obvious answer and it's wrong, or at least insufficient – a shorter document that still assumes a patient reader has the same problem in less space.

The change is to stop treating the document as the delivery and start treating it as the record. The delivery is the twenty minutes in the room where you say the one sentence and answer the hard question, or the conversation where someone gets to argue with you. The document is what that conversation leaves behind so it can be checked, referenced, and disagreed with later by people who weren't there.

That's a reversal of how most of this work is scoped and sold, where the document is the deliverable and the meeting is the formality around it. It also explains something otherwise strange about the research: 88% of people not reading in full is a catastrophe if the document is the product, and largely fine if the document is the artifact of a conversation that already landed.

Every knowledge business is about to find out which of the two it's been selling. The tell is whether anything happens after you hand it over.