AI as Task Decomposer

Picture the task that has been sitting on your list for three weeks. "Do the taxes." "Write the proposal." "Sort out the garage." Notice that none of these is actually a task. Each one is a category containing dozens of smaller tasks, and your brain, looking at the category, sees only a wall. The wall is the problem. You cannot start a wall.

Of the four mechanisms by which AI helps ADHD brains, this one is the most immediately useful, because it goes straight at the thing that makes starting feel impossible.

Why the wall stops you

ADHD comes with real difficulty in what researchers call task decomposition, the act of breaking a large goal into ordered, doable steps. For a neurotypical brain this often happens automatically and invisibly. For an ADHD brain it can be genuine work, and the work has to happen before any of the actual work can begin, which means you face two hard tasks where you thought you faced one. Faced with that, avoidance is the rational move.

The 2025 ISCAP analysis of 45 studies named task decomposition as one of the four core mechanisms by which generative AI helps, describing how these tools break complex work into manageable, discrete components. This is not a small feature. For an ADHD brain, the decomposition is frequently the entire barrier, and handing it to something that does it instantly removes the exact step where you were getting stuck.

What this looks like in practice

The move is simple. You take the wall and you ask for the first brick. Not the plan, not the whole breakdown, just the smallest possible first action.

"Do the taxes" becomes, after one sentence to a chat window, "open last year's return and find the income section." That you can do. It takes ninety seconds, and once it is done you are inside the task rather than outside it, which is a completely different psychological place to be. The wall has become a doorway, and you have already stepped through.

The reason this works is the reason the whole blog keeps returning to: the barrier is the start, not the continuing. Decomposition shrinks the start until it is smaller than the urge to avoid it. AI just makes the decomposition instant and effortless, so the one hard cognitive step that used to trigger avoidance no longer has to be done by the brain that finds it hardest.

The clinical signal

There is early evidence that professionals see this potential too. A 2025 study from the University of Palermo worked with seventeen neuropsychology and neurorehabilitation experts to design and evaluate prompts for using a large language model to build executive function support plans for people with ADHD. The notable part is who was doing the evaluating: clinicians, not technologists, treating AI-assisted decomposition as a serious tool worth designing carefully. The field is young and the study is preliminary, but it is a real signal that the decomposition use case is more than a productivity-blog trick.

How to use it without losing the skill

One honest caveat, which the final post in this series develops fully. If you let AI do every decomposition forever, you may never build the skill yourself, and there will be moments without a chat window. The healthier pattern is to use AI to get unstuck, then notice how it broke the task down, so that over time you are learning the move rather than only outsourcing it. Use it as a tutor that shows its work, not only as a vending machine that hands you steps.

But on the days when you are frozen in front of the wall, do not be precious about it. Ask for the first brick and start. A task begun is worth more than a skill theoretically preserved while the task rots on the list.

Next in the series: AI as cognitive scaffold, holding the working memory your brain keeps dropping.