Writing with GenAI for Professional Purposes

The use of Generative Artificial Intelligence (GenAI) to produce texts for professional purposes, like emails and reports, is increasing rapidly. The more people use it, the more accepted it becomes, even though criticism pops up here and there. Rarely do team leaders know what it actually means when their people use GenAI to write texts at work.

I’ve been working closely with GenAI-supported writers from nearly all sectors since 2022. I’m not going to argue whether using GenAI is good or bad, and of course I’m assuming that issues of data protection, confidentiality, and copyright are covered. The important question for companies and leadership is how their staff use GenAI to produce text, and what that means for their organization.

Because some voices are hesitant. Some companies and governments continue to forbid it, and we’ve even heard tech companies claim they don’t need more tech specialists, or “math wizards,” as GenAI will take over these tasks anyway; what they need is smart and skilled people. This raises the question of whether the use of AI-generated text production can help us figure that out.

So, how do people actually use GenAI for writing? And what does the management, who’s expected to take decisions on how to handle AI policies, need to know about their staff using GenAI in their professional communication?

A text is a series of pieces of information coded in linear form. Yes, I do mean “coded,” as every human capable of speaking and writing is a “coder.” And there’s always two sides to a code: the content and the form, or the what and the how. Whenever we write, we first need to know what we want to communicate before we decide how we communicate it.

Prompting now allows writers to get their text from the software instead of writing it themselves. They indicate what content and style they need, and if they wish, they can even have the text sound as if it had been written by Shakespeare himself.

That sounds fantastic at first. It may, however, backfire later on. Because writers must know what they’re doing.

So, first things first: how do we actually write a text? Writing is a highly individual process. Some start with bullet points, others with the running text. Then they reread the text and refine the wording until they conclude the final version. There’s no perfect procedure, there’s only a perfect result, and the more skillful the writer, the faster they’re done.

While writing that text, the writer assumes responsibility for the text. As they proceed, they decide what to say, in which order, and how, because they’re aware of their addressee’s interests. This process entails a lot of decisions and usually happens under pressure, and most people find it hard to constantly take such decisions.

Accordingly, leaving such decisions to GenAI usually feels like a relief.

For people who struggle with taking such decisions, GenAI accelerates the creation of the first draft. That generated version is, however, just the first version, because GenAI can also rephrase the text. And rephrase it yet again – indeed, ad infinitum. The text is never finished until the writer decides to close the process. So, it’s still the writer’s responsibility to decide that the text is good-to-go.

The problem for most inexperienced writers is that they don’t know how to evaluate a text. In fact, they’re likely to find it okay because it’s grammatically and orthographically correct. That’s the effect of skillful rhetoric: we’re likely to trust an eloquent speaker more than an ineloquent one, and the same happens with generated text. When the text sounds correct and well-worded and the pressure to finish it is high, inexperienced writers are likely to accept a text regardless of its quality.

This decision-making process is determined by the writer’s skillset and mindset. Accordingly, there are generally four types of writers:

First, there’s the skilled writers with a strong mindset of assuming responsibility for their message. They know what they’re doing. When these writers use GenAI, they’ll get a text that is easy to revise and quickly sent off. No problem. They might even refrain from using GenAI altogether because they know their job anyway.

Second, there’s those with limited writing skills yet a strong mindset who know their job and know when it’s done. Their limited writing skills may be due to them writing in their second language or because they’ve always found writing a pain in the neck. These writers are often techies or other kinds of specialist, and they generally use GenAI to convert their bullet points into running text.

Then there’s those who are skilled writers but lack the mindset of assuming the responsibility for the message. They get lost in a loop of revisions, because they constantly question their message and style. Their texts are often unspecific because they feel insecure and find it difficult to finish the text. With guidance and some training, GenAI can help them save time.

The fourth type are writers who lack the skillset of writing and don’t really know what to say. Yet, they still need the job done. Since they only have a vague idea of what to write and how, they use GenAI to produce a series of grammatically correct sentences with unspecific messages. Their text sounds okay at first but hampers further collaboration or gets rejected further down the line. For leaders, it’s important to closely monitor employee development in this case and work on communicative skills.

So, writing for professional purposes is an issue of writing skills and mindset. GenAI can help users write faster provided that they know whether the outcome meets expectations and makes accurate statements. That’s of course also a question of seniority.

GenAI for text generation is not like data crunching, which GenAI truly masters by doing things in seconds that would otherwise take weeks. If we don’t know how to write an accurate text, GenAI won’t solve it for us. The skills we’d develop would be prompting skills, not communicative or leadership skills. Studies have indeed shown that the frequent use of GenAI makes users’ knowledge shallower, and all the smart guides giving advice on how to write with just a few prompts won’t make people smarter but better prompt engineers.

To answer this question, we mustn’t forget that people write for professional purposes. For employees who aren’t skilled writers, there’s simply not enough time to practice and become skillful writers at work. So, from a leadership point of view, it’s sensible to endorse the use of GenAI because it will remove pressure.

Yet, leaving them without support would be short-sighted. GenAI is a shortcut to reach a goal, and since writing is a skill that takes years to develop, GenAI can help people become more efficient in less time, and it can help those who struggle keep up with their workload. When the staff gets the support they need, GenAI can also strengthen decision-making and align processes and mindsets.

I recommend leadership to encourage their teams to use GenAI for text production, but to closely guide them along the way. Providing professional support will be key, because there will be uncertainties and questions, and learnings should go beyond prompt engineering to have a lasting effect. Because the value of well-developed writing skills for professional purposes keeps increasing.