American businesses are facing a writing crisis that predates artificial intelligence but threatens to undermine its potential. Employees who can’t write clearly will feel this the hardest, and this fundamental skills gap is becoming the hidden bottleneck to our AI-powered future.
This crisis hits particularly close to home for me. Nearly a decade ago, I was building an AI and machine-learning writing platform for US Higher Education, and the statistics we uncovered were genuinely alarming. The sheer amount of remedial learning and practice needed to bring college students up to based writing competency revealed how late in the game we were in solving his problem.
What made building this product so exciting for me was my background in writing: as a teacher teaching writing across different contexts, like ESL and A Level Literature & Poetry and ESP, and as a development editor analysing and refining manuscripts for international language learning textbooks.
This experience has shown me that writing isn’t just a communication skill, it’s a thinking skill. Now, as AI becomes the dominant interface for professional work, that same clarity threatens to undermine every interaction we have with these powerful tools.
The Number Don’t Lie
The statistics paint a grim picture that predates the current AI revolution. Reports from 2016 show that nearly a decade ago this was a critical issue with a study from CollegeBoard, a panel established by the National Commission on Writing, showing us that blue chip businesses were spending as much as $3.1B on remedial writing training every year. Today, poor communication costs US businesses approximately $2T annually, equating to over $15,000 per employee.
For decades, managers have complained about emails that bury the main point in paragraph three, reports that read like academic dissertations (from graduate students used to writing academically now having to adopt a different style with no training) and memos that somehow say nothing in 500 words. The rise of instant messaging and social media may have accelerated the decline, but the fundamental issue runs deeper. Many workers simply never developed the ability to organise thoughts logically and express them clearly.
When employees can’t express ideas clearly, projects stall, deadlines slip, and opportunities vanish. And this isn’t just at the junior level, even the most experienced employees add to this financial cost by not being able critique language - they know what it should look like but they don’t know how to get someone else there.
Why Clear Writing Matters More in the AI Era
Even writing that heading makes me feel like I’m belabouring the point but the writing crisis stems from gaps in fundamental linguist knowledge. Most employees never learned the cognitive processes that underlie effective communication.
Linguistics research reveals that writing proficiency depends on metalinguistic awareness: the ability to think consciously about language as a system (systems thinking… product management… it’s all starting to come together). This includes understanding how syntax affects meaning, how word choice shapes perceptions and how structure guides comprehension.
The work of linguist Stephen Krashen on language acquisition show us that writing improves through comprehensible input and meaningful practice, not through rote grammar drills. Yet most workplace writing training focuses on surface-level rules rather than developing deep linguistic competence.
Similarly, research in discourse analysis demonstrates that effective writing requires understanding genre conventions, audience expectations, and rhetorical moves. These skills transfer directly to AI prompting, where success depends on understanding how different phrasings trigger different response patterns. In short, for leaders who are unclear, unable to give direction or provide a vision of their expectations to junior colleagues, don’t expect artificial intelligence to understand them any better.
Pedagogical Insights We’re Ignoring
Educational research offers clear guidance on developing writing expertise that applies directly to AI interaction skills.
Writing researcher Linda Flower’s work on cognitive rhetoric shows that expert writers engage in complex planning processes, considering purpose, audience and constraints before drafting. This same cognitive approach is essential for effective AI prompting.
Th process writing movement pioneered by researchers like Donald Graves and Lucy Calkins, emphasised that writing is thinking made visible. This insight is crucial for AI users, who must articulate their thoughts clearly enough for a machine to understand and extend them.
Composition theorist James Britton’s research on writing development found that expressive writing (writing to discover what you think) must precede transactional writing (writing to accomplish specific goals). Many workplace writers skip this crucial development step, jumping straight to task-orientated communication without developing the underlying thinking skills.
The AI Amplification Effect
Poor writing skills don’t just limit human communication, they amplify when filtered through AI systems. A poorly written prompt generates poor output (or, one of my favourite phrases: garbage in, garbage out) which then gets shared with colleagues, clients and customers. The original writer’s unclear thinking gets multiplied across the organisation.
This creates what we might called “cascading incoherence”. When AI sysmtems trained on unclear human input generate responses that are then used as input for further AI interactions, the degradation compounds. Clear writing becomes a critical quality control mechanism in AI workflows.
As AI tools become more sophisticated, the ability to critique and refine their output becomes paramount. This requires not just knowing what sounds right, but understanding why it sounds right - a skill that depends on deep linguistic knowledge.
I remember times when my leaders would tell me to rewrite things with no clear direction on how or arbitrarily telling me what to write with no clear explanation. That doesn’t help people learn, I didn’t know how to do it by myself because I had no idea why I was rewriting it other than someone telling me it was “wrong”. It wasn’t until, after many months of just being told what to do, did I suddenly realise they were asking for edits to rephrase things for an audience they knew better than me. Without that metalinguistic awareness or even a “we’re doing this because…” you’re not helping people learn.
The Hidden Skills Shortage
The writing crisis reveals a broader problem: we’ve traded communication as a soft skill rather than a technical competency. But in an AI-mediated workplace, writing becomes a form of programming. The precision required to promote an AI system effectively mirrors the precision required to write clear code of technical documentation.
Organisations investing heavily in AI training (hopefully with the help of an instructional designer) while ignoring writing fundamentals are like companies upgrading their hardware while running buggy software: the more powerful the system, the more dramatic the failures when the human interface breaks down.
What This Means for Business Leaders
Smart organisations are already recognising this connection. They’re treating writing training not as a nice-to-have communication skill, but as essential AI literacy. They’re teaching employees to think of prompt as instructions, outputs as drafts requiring revision, and AI interactions as collaborative writing processes.
This means identifying leaders who struggle leading people or giving clear directions to junior members. Leaders who rely on mind-reading and vague communication will struggle the most in the age of AI. If people can’t understand them, machines definitely won’t. Alarm bells should ring if AI is blamed for poor output: garbage in, garbage out.
It also means writing skills should evolve with role seniority. Junior team members need to be strong writers because they’re often creating the content. But as individuals move into leadership roles, their focus should shift toward the ability to assess, guide and critique others’ writing. If your people leaders can’t model and support those skills today, the gap in both succession planning and writing capability will only continue to grow.
The companies that solve their writing problem first will have a significant advantage in deploying AI effectively. They’ll be able to extract more value from AI tools, generate high quality outputs and maintain better quality control over AI-generated content.
Writing As Strategy
The solution isn’t more grammar workshops or style guidelines. Organisations need comprehensive writing development that addresses the cognitive and linguistic foundations of effective communication.
This means teaching employees to understand their own thinking processes, to analyse audience needs systematically and to structure information for maximum clarity. It means helping them develop the metalinguistic awareness that allows them to understand why certain phrasings work better than others.
Most importantly, it means recognising that in an AI-powered workplace, writing skills are not optional. They’re the interface between human intelligence and artificial intelligence. Get that interface right, and AI becomes a powerful amplifier of human capability. Get it wrong, and AI becomes an expensive way to scale up confusion.
The writing crisis isn’t going to go away on its own. But for organisations willing to address it head-on, it represents an opportunity to build a significant competitive advantage in the AI economy. The question isn’t whether you can afford to invest in writing skills. It’s whether you can afford not to.
This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit thestrategiclinguist.substack.com/subscribe
