Why effective AI communication starts with the recipient

I don’t know about you, but when I read something I can tell was clearly produced by a large language model, it puts me off immediately. I lose interest, and I sometimes lose trust in whoever sent it, especially when they clearly haven’t put any thought into it.

That reaction is becoming a lot more common. AI has made it cheaper than ever to produce words, slides and plans, and people are drowning in “workslop,” content that looks polished at first glance but has not substance beneath it.

It’s becoming so common that 40% of workers say they’ve received AI workslop in the past month, and dealing with it costs those workers nearly two hours on average.

That invisible tax adds up to millions of dollars a year for a large organization. Our CEO, Victor Riparbelli, recently put another name to this in an internal memo: “AI-sloppification.”

AI is making content cheaper to produce without making it any more useful to the person receiving it.

The Cost of “Workslop”

The root problem is that we forget why we communicate in the first place. We don’t write memos, decks, or updates for their own sake; we write them to change something in the person on the other end: inform them, train them, get them to act, or draw insight out of them.

The recipient is who matters. The only real measure of successful communication is whether it did that job for them, not how quickly or cheaply it was produced.

Once you take the recipient’s side, it becomes immediately apparent that AI cannot be viewed as a shortcut and needs to be a test. That philosophy has led to two internal rules on my team.

Two Rules for Mindful AI Usage

Rule one: you cannot use AI for writing emails, reports, or decks unless what you produce is both shorter and richer than what you could have written without it. If a tool makes you faster but the result is longer or more generic, you haven’t saved any work; you’ve only shifted the work of distilling it onto the reader.

That’s exactly the workslop survey finding in practice: one person saving ten minutes editing can end up costing ten readers ten minutes each.

Rule two: be intentional about the format that actually serves the message, not the fastest one to produce. Text is often right for something quick.

But through AI we can now use richer formats for the things that really matter, and that’s when AI video can make a huge difference. A short walkthrough, a roleplay, an interactive coaching moment can land something a written document never will.

How We’re Applying These Principles

We tested this recently with an internal employee survey on AI tools. Instead of sending people a form and asking them to rank or check off which tools they use, we ran it as a conversation with an AI avatar, asking people to talk through specific moments a tool actually helped them and following up live on what they said.

We cared far less about the number of tools someone named than the texture behind it – why it helped, when it didn’t, what almost worked. That conversation gave us richer insight than any survey form could have.

We’ve used the same approach in onboarding. All of our new employees get hands-on with the product from day one, and that matters. But what matters more is the visibility that gives us into how people actually onboard: where they get stuck, what clicks fast, what needs to be framed differently for different people.

That insight is what lets us shorten ramp-up time and get new hires to real productivity faster.

None of this makes AI less useful. Quite the opposite, it raises the bar for using it well and makes it more useful to more people. When the cost of producing communication drops to nearly zero, the only thing separating good communication from noise is whether it was built with the recipient in mind. This requires discipline, and we’re building it into how our entire company works.