Do you trust artificial intelligence? How much? Would you trust it to write a social media update for you? An email? A company blog post?

And how much do you trust a human who knows what they’re talking about? I can imagine that out of these two, you’d trust the human more. Perhaps this is because you know that a human is able to validate the information, or perhaps because a human is more able to put themselves in the reader's place.

Since the boom in chat AI, we have seen this technology evolve. We received many questions from our customers - “how can we implement AI in our work and take advantage of all its benefits?” As the lead UX designer of Tridion Docs, I was curious to see how we could leverage AI to improve the author’s day-to-day experiences while keeping it trustable.

In addition to the trust concerns, technical writers may not have a clear picture of the productivity gains AI can bring to them personally in their daily work and might even see AI as a threat to their jobs. Generally, however, authors embrace progress, love technology and enjoy working with it, so it is critical for them to understand how AI can enhance their job, not replace it.

Here I want to respond to these concerns and share my thoughts about this.

Let's start with a quick analysis of the typical writing process, such as for writing documentation for new equipment. 85% of the time is not spent on writing text, but on clarifying details, communicating with the team, accumulating this gathered knowledge, and structuring the information. And in this process, human can typically assess the quality of the text from both an emotional and logical standpoint, which can be challenging for AI. AI could create text that "seems" correct in terms of grammar and structure, but it may miss the subtle nuances that make text truly good.

You, the author, are not going anywhere because you bring these critical elements to any writing project:

• the audience composition and level of knowledge;
• how the audience will use the information;
• the specific context of the documentation.

No matter who or what is producing the text, useful content needs planning and a lot of details, which only come from a human. Don’t worry that AI is able to compete with your thinking!

 LLM (large language model) - is a type of artificial intelligence (AI) program that can recognize and generate text, among other tasks.

But in order not to underestimate the abilities of AI, I would like to point out what currently large language models (LLM) do very well:

• Simplifying/adjusting text - it can suggest alternative wordings or change the structure of sentences to improve the text’s readability.
• Organizing/structuring - it can help you organize information by categorizing and grouping related content. So you can ask LLM to create hierarchical structures or tables to make the information easier to perceive.
• Quick analysis of large volumes of data to make an overview or prepare a summary.
• It is good at languages - it can check your content in terms of grammar so that authors can feel more confident while writing in another language.

So when it came to bringing AI into our product to help authors, we kept these advantages in mind.

Our new tool Draft Companion is called Companion and not Author for a reason – because the author here is you, and Draft Companion will help to make your work easier and not replace you.

Draft Companion works with the draft content you have already created, not by creating content from scratch. Even if you have just a rough outline, it can help you. In this way, you get the best overall quality - combining your knowledge, empathy, and intuition with the intelligence of the LLM. One way to look at the LLM is as having a personal peer editor, a second pair of eyes you can consult at any time.

To interact with a chat AI, the user enters prompts - requests that explain what is required from the LLM. And Draft Companion has quick prompt buttons that we believe you’ll use the most:

• Improve readability - suggest alternative wordings, such as to be less technical, or adjust the sentence structure to raise the readability score;
• Summarize - quickly analyze large volumes of data and offer a summary;
• Check grammar - identify errors in grammar, spelling, and punctuation.

and of course, you can write your own specific prompts. And finally, you can apply a suggestion whenever you’re ready or copy and tweak it manually.

So, what now? Effective collaboration between an author and AI brings out the best qualities of both. Use Draft Companion to adjust your drafts or even to generate ideas, and then edit, supplement, and style the text.

Try Draft Companion and see how effectively and confidently content authors and AI can work together!