From Chatbots To Automation: The AI Ladder Tech Writers Are Climbing

Every workplace now has that person who says, “We should be using AI for this,” as if that sentence has somehow completed its education and can now support itself.

For tech writers, that advice is useless on its own. The question isn’t whether we should use AI. The question is where it fits, what kind of work it should touch, and how much chaos we’re willing to invite into the building.

This post from our friends at CherryLeaf is useful because it separates AI use into four distinct approaches: web-based chatbot agents, no-code AI app builders, command-line AI systems, and workflow automations. It also makes clear that these approaches differ in skill needed, risk level, and organizational ambition.

That’s an important distinction, because too many teams talk about AI as though it were a single appliance you plug in beside the microwave. It isn’t. These are four different ways of applying machine assistance to work, and they don’t solve the same problems.

What Tech Writers Should Do With This

The practical takeaway is simple. Don’t ask, “How do we use AI?” Ask, “Which part of our work are we trying to improve?”

👉🏾 If you want faster thinking and quick experiments, start with a chatbot agent.

👉🏾 If you want a repeatable tool colleagues can use without your supervision, build a no-code app.

👉🏾 If you need AI to inspect technical systems directly, look at command-line tools.

👉🏾 If you want to change how documentation work flows across the organization, then workflow automation is the real play.

CherryLeaf’s conclusion is that chatbot agents and no-code apps are the lowest-risk entry points, while workflow automation offers the biggest transformational upside but requires more planning, stakeholder alignment, and change management.

So don’t begin with “transform documentation.” Intead, begin with one painful, repetitive task. Estimate a project. Classify release notes. Identify missing prerequisites. Trace which product changes actually need user-facing content. Start there, see what breaks, and pay attention to where the upstream information is too vague, too messy, or too dependent on tribal knowledge to support automation.

Because that’s the real gift of AI in documentation. It doesn’t just speed things up. It reveals where your process has been held together by hope, habit, and one extremely patient person named Karen.

And Karen, frankly, deserves a break.🤠