ADHD × AI: Building Systems Around the Brain You've Got

Got diagnosed with ADHD as an adult, after fourteen-plus years leading engineering teams across Australia, Asia and India and quietly wondering why I could ship multi-million-dollar programs but couldn't keep a personal to-do list alive past a Tuesday. The diagnosis didn't change anything in the world. It just renamed the thing I'd been working around the whole time.

Most productivity advice is written for brains that can follow a linear system. Mine can't, and trying to force it ate years. The shift was giving up on fixing the brain and starting to design systems around the way it actually runs. AI is what made that possible at a level that genuinely wasn't on the table even two years ago.

Everything below is in active use across EDGE and becomingscott, not theory and not a framework. If your brain works like mine, at least a few of these will land like finding a tool you didn't know existed.

An ADHD brain produces ideas constantly and at the worst possible moments. Mid deep work, mid conversation, mid shower. The standard advice is "write it down and come back to it later," which sounds fine until you notice that writing it down is already a context switch, and for my brain that switch costs about twenty minutes of re-entry into whatever I was doing.

The working fix is a zero-friction voice capture that feeds straight into Claude. The idea gets spoken out loud, transcribed, categorised and filed somewhere I won't look at it again until the daily planning block. The idea is safe, focus stays intact, and nothing gets lost in the gap between having the thought and acting on it.

The thing that makes it work is treating capture and processing as completely separate events. My brain wants to chase every idea the second it appears. The system's job is to give me a way to acknowledge the idea without engaging with it.

This is the one nobody describes properly. ADHD isn't a confusion about what to do. The task is clear, the priority is obvious, and the activation energy to actually begin still feels impossible some days. The screen stays blank and the hour evaporates.

The way AI breaks this is by removing the blank page entirely. Instead of starting from nothing, the work starts from a rough first draft. Claude generates the scaffolding (an outline, a draft version, a structured opening) and the task quietly converts from "make something from scratch" into "improve something that already exists." That conversion is the actual difference between staring at the screen for an hour and shipping inside a focused block.

The "push against" wording is doing real work in there. My brain activates better in reaction than in creation. Tearing apart a mediocre first draft is genuinely easier than producing a good one from nothing, and the system is built around that fact instead of pretending otherwise.

Running multiple businesses at once means jumping between totally different contexts several times a day. For a neurotypical brain that's maybe five minutes of re-orientation each time. For mine it can run to thirty, and by the time I'm finally back up to speed, something else has already pulled my attention sideways into the next thing.

The workaround is Claude projects with the full business context loaded in. When the day shifts from EDGE work to becomingscott work, there's no mental rebuild. The project already knows what EDGE is, what's in flight, what stage things are at and what the priorities are. Drop in mid-thought and the AI meets you there instead of asking you to brief it for ten minutes first.

Pair that with end-of-block notes. Before leaving any task, 60 seconds goes into telling Claude what was happening, where the work stopped, and what the next step is. Tomorrow, or even later that afternoon, opening that thread puts me straight back inside the work. The re-entry cost drops from thirty minutes down to about two.

ADHD brains burn through decision-making energy faster than neurotypical ones do. By midday the tiredness isn't from the work itself, it's from the constant micro-decisions about what to do next, in what order, for how long. Every option feels equally urgent and equally important, which is another way of saying nothing actually gets prioritised.

The fix is aggressive pre-decision. Every evening during shutdown, tomorrow's single priority gets set. Not five things ranked by importance, one thing. Claude reviews what's in progress across EDGE and becomingscott and recommends the highest-leverage single action for the next day. The morning version of me doesn't get to decide. That decision was already made the night before, while the brain was still functional.

The "tell me why" is the part that matters. My brain needs to actually believe the priority outranks the twelve other things competing for attention by 9am, and the reasoning is what gives me something to hold onto when the morning's shiny objects start showing up.

Hyperfocus is the ADHD superpower that nobody bothers warning you about the downside of. When it lands, six hours of building can pass without a blink. The catch is that hyperfocus doesn't discriminate between important work and unimportant work. Entire afternoons have disappeared perfecting things that genuinely did not need to be perfect, and the feeling of productive flow makes it almost impossible to notice in the moment.

The counter is timed check-ins. Every 90 minutes a notification asks one question. Is what you're doing right now the most important thing? If the answer is yes, the work continues. If the answer is some version of "honestly, not sure anymore," that's the signal to surface, check the priority list, and recalibrate. It sounds almost too simple but for this brain it's the difference between productive hyperfocus and a beautifully over-engineered thing nobody asked for.

Every ADHD person knows this pattern in their bones. New system, enormous initial energy, solid execution for roughly a week and a half, then the novelty wears off and the whole thing quietly joins the graveyard. Personally I've done this with task managers, note apps, journals, and at least four different morning routines, all of which looked promising at the time.

The only version that's actually stuck is one built around AI in a way that adapts to the day's energy instead of demanding a consistency I can't reliably deliver. On high-energy days the full operating system runs (priming, deep work, content, ops, the lot). On low-energy days there's a minimum viable version: one priority task, one journal capture, one piece of content. The AI scales the scope up or down based on what I tell it about the state I'm in that morning.

The system survives bad days because a bad day is already designed into it. That's the difference between this and every productivity system I tried before. It was built by someone whose brain actually works this way, for that exact brain, instead of borrowed from someone whose default mode is the one mine is trying to fake.