Can AI Actually Help With ADHD? A Real Therapist Tested It.

I opened Claude, typed in four scenarios that come up constantly with my ADHD clients, and let it answer. Then I judged every response.

Not to prove a point. Not because I’m skeptical of AI in general. Because a lot of people with ADHD are quietly using these tools and I wanted to see what they’re actually getting.

Here’s what I found.

The Setup

I typed each scenario the way a real person would — no coaching, no clinical framing, no loaded context. Just the problem.

That matters. Because the quality of AI output isn’t just about what AI can do. It’s about how people actually use it. And most people aren’t prompting it like a researcher. They’re typing in a problem and hoping something useful comes back.

Test 1: Procrastination

The prompt: I’m having one of those days. I cannot stop procrastinating. Can you give me a few tips to get started on a task I need to start? It’s a homework assignment. Please help.

Claude gave five suggestions:

• Just start for two minutes

• Break it into the smallest possible first step

• Clear your environment

• Use a body double

• Forgive yourself and just begin

My take: Two and three are genuinely useful. Starting for two minutes is a real technique — the activation energy is the hard part, not the work itself. Breaking into the smallest first step is solid. Body double is underused and underrated, even if it takes friction to set up.

“Forgive yourself and just begin” is the weakest one. Not because self-compassion is bad — it’s not. But it’s not going to get you started on a homework assignment right now.

The bigger issue: no real explanation. Why does two minutes work? Why does body double work? Without that, it’s just a list.

Verdict: Decent at best.

Test 2: Forgetfulness

The prompt: I keep forgetting my wallet every day when I go to work. This is causing real problems because I don’t have my ID and it’s hard to get into the building. Please help me stop forgetting so much.

Claude’s suggestions:

• The nuclear option: don’t carry a physical wallet

• Create a launch pad (one spot by the door for wallet and keys)

• Habit stacking (attach wallet-grabbing to something you already do)

• A physical reminder (sticky note inside the front door)

• Get a Tile or AirTag

• The night-before rule: pack your bag the night before

My take: “Launch pad” is a real concept — picking one fixed home for the thing so you never have to remember where it is. Habit stacking is solid too. If you never forget your shoes, and your wallet lives in your shoe, that’s a real system.

The nuclear option (ditch the physical wallet) misses the point. That’s not really building a skill. And shoving an AirTag in your wallet is a solution almost nobody is going to actually do.

The deeper issue: forgetting your wallet is usually a working memory problem. None of these tips name that. They’re surface-level fixes without any understanding of the underlying mechanism.

Verdict: 2 or 3 good ones. The rest are filler.

Test 3: Hyperfocus and Relationships

The prompt: I find myself coding for hours on end and I get into the flow state. It’s really hard for me to stop, but I keep missing plans with my partner — like dinner or time together. Can you give me some tips to improve this?

Claude’s suggestions:

• Set hard stop alarms (not reminders — actual alarms labeled “leave now”)

• Use a parking note before you close out (write down exactly where you are)

• Timebox your coding sessions

• Make your partner part of the system

• Schedule coding like you schedule plans

• Recognize what’s at stake

My take: We start losing ground here. Hard stop alarms are okay — they need to be audible and labeled clearly, so that part tracks. Timeboxing is the problem, not the solution — the whole issue is that people with ADHD struggle to stop when the timer goes off.

Making your partner part of the system is actually not bad. Inviting someone in — not to babysit you, but because you’ve named what would help — is a real thing. It’s taking ownership and creating accountability.

“Recognize what’s at stake” is the weakest suggestion. Anyone in this situation already knows what’s at stake. Knowing didn’t stop it from happening.

Verdict: Starting to miss the real issue.

Test 4: Hobby Hopping

The prompt: I get interested in so many different things. I stick with something for one week, then switch to another. I go really deep, I switch to another, and it just feels like I’m switching hobbies all the time and losing all my free time. I think this might be ADHD related. Can you please give me some tips?

Claude’s suggestions:

• Embrace being a “multipotentialite”

• Create a hobby shelf (write down the new interest instead of abandoning the current one)

• Set a minimum commitment window

• Limit active hobbies to two at a time

• Schedule interest hopping

• Track what you’ve done, not just what you’ve abandoned

• Talk to a professional about ADHD

My take: The last one is the best suggestion in the whole set — Claude recognized it couldn’t actually help here and pointed toward professional support. That’s the right call.

The rest aren’t terrible as general life advice. But they’re not ADHD-specific. What’s missing is any real understanding of why hobby hopping happens with ADHD — the novelty-seeking, the dopamine drop once something is learned, the way interest functions differently in ADHD brains. Without it, the tips just feel like advice from someone who Googled the problem.

Verdict: The most important tip was “go see a therapist.”

So Can AI Actually Help With ADHD?

Somewhat. Here’s the honest picture.

Some of the suggestions across these four tests were genuinely good — two-minute rule, body double, habit stacking, launch pad, hard stop alarms, inviting your partner into the system. Those map to real skills. If you’re not already using them, they’re worth trying.

But AI doesn’t really understand ADHD. It doesn’t know why these things work. It gives you a list without the reasoning, which means if the first tip doesn’t work, you’re stuck — because you don’t know why it didn’t work or what to try next.

The other thing it can’t do is personalize. ADHD is highly individual. What works for one person won’t work for another. A therapist helps you figure out which executive function is being impacted and builds a tool around your specific brain — not a generic list of seven things.

The best way to use AI for ADHD, if you’re going to use it: bring what it gives you to a real provider. Let the list be a starting point, not a solution. Process it with someone who can help you figure out what actually fits.