There is no shortage of technical writers and documentarians pointing out that AI is not replacing our profession, but it is changing it. If you follow my posts, you know I agree. AI is disrupting how we work, and used well, that disruption can be a real advantage.

The question I get most often is how. To me, the answer starts with intent. There is a big difference between using AI because it is available and using it with a clear purpose. When a company understands what it wants AI to support – and gives people the right tools and access – it becomes a force multiplier.

Most of us come in mid-stream

Most documentation work does not start from a blank page. More often, you join a company that already has years of content in place. That content may have passed through many hands, picked up different writing styles, and drifted into inconsistent formats over time.

The goal is usually to bring everything into a shared standard, but the scale can be overwhelming. In the past, that meant converting documents in batches. New or urgent content followed the updated guidelines first, while older material waited until someone had time to get to it.

With AI, that work can move much faster. You can define a rule set for formatting, grammar, style, reading level, terminology, and structure, then use your preferred AI tool to convert existing content against that standard. With the right prompting and review, you can do it without losing content fidelity.

You can’t fix what you can’t measure. 

A maturity model gives each document a clear score against defined criteria such as ownership, structure, quality, and upkeep. I used Campell and Swisher’s model as a base and inventoried our existing documentation repository to assign a maturity score to each article. That score shows where a document is strong, where it falls short, and what needs to improve. Instead of a vague goal like “improve the docs,” the work becomes concrete and trackable. The gaps become a prioritized to-do list, and progress can be measured over time.

Using AI, you can refresh those scores against live content, update titles, add newly completed documents, raise scores where work is finished, and sync the related tasks.

Because I built the process as a reusable tool using Claude Cowork, it can run again whenever needed.

Build a smart assistant on top of it. 

Even a clean, well-organized library still depends on people knowing where to look.

When teams are busy, they do not have time to search through pages, and too much knowledge can stay trapped in people’s heads. Building a smart assistant helps close that gap. It lets someone ask a plain-language question and get the right document quickly, without guessing where to search.

My company leverages the full Microsoft ecosystem, so it made sense to build mine in Copilot Studio. The important part is that the assistant only works because the content behind it is structured and maintained. Good documentation makes the assistant useful; messy documentation makes it unreliable.

Keep the project on track

As we head toward a big launch, there are a lot of moving parts. You’ll have tasks, owners, priorities, and deadlines. Rovo helps me keep track of all the tasks on my Kanban board in Jira.

---

The theme is simple. I use AI across the full lifecycle of the work: rewriting inherited content, measuring and tracking quality, powering the assistant that sits on top of it, and handling the communication that drives adoption. Each piece supports the next.

And I’ve only just begun.