Why I Use Claude as an Accessibility Tool

I started with ChatGPT, moved to Gemini, and chose Claude. The reason comes down to how each one communicates — and for someone with dyslexia, that is imperative.

The reality of writing with dyslexia

Writing anything — an email, a document, a reply to a message — is not a linear process for me. My thoughts do not arrive in order. They arrive all at once, in fragments, and the work is in arranging them into something that makes sense to someone else.

Through dyslexia coaching, funded via Access to Work, I was taught a structured approach:

1. Use a mind map to capture everything first
2. Draft a sentence or paragraph for each point
3. Arrange those into a logical order
4. Write the linking sentences
5. Read the whole thing back using a screen reader to check the flow

It works. It is also a significant process for something as routine as replying to an email with several questions — the kind of thing that can take up a large part of a day.

Access to Work, the UK government scheme that funded that coaching, typically involves a long wait — a long time to go without support you already know you need.

This is where tools like generative AI (gen AI) have become useful for me.

Where gen AI fits in

The workflow I use now is simpler. I open Voice Memos on my phone and speak through everything as it comes to mind — non-linear, unordered, however it arrives. The OS transcribes it automatically. I put that transcription into Claude, and it arranges it into a coherent narrative.

My thinking is still entirely my own. What changes is the structural overhead — the arranging, the linking, the re-reading — which used to take up a large part of my day, and now does not.

This is why verbosity in a gen AI tool is a barrier for me. More text to read means more processing, more time, and more cognitive load. When I tried ChatGPT and Gemini, I found both too chatty — responses that could be two sentences arrived as six, often with a follow-up question at the end that extended the exchange further. I adjusted the context and tried custom agents, and it helped, but the verbosity would still return without warning.

Claude is different. Responses are more direct. Conversations reach a natural conclusion rather than being nudged forward. For someone who finds dense or verbose text harder to process, that default behaviour matters.

The environmental question

Some people have stopped using gen AI because of its environmental impact. I think that is a valid position, and I respect the people who have made that choice based on their values.

What is sometimes missing from that conversation is an awareness of the position not everyone is in.

Opting out of a tool that removes daily barriers is a different decision depending on where you are starting from. That is not a criticism of people who have made that choice. It just rarely comes up.

An honest caveat

I am not tied to Claude, or to any gen AI tool. Things change — pricing, quality, priorities. Tools that start out useful tend to become expensive, degraded, or both. Google Search, Facebook, Twitter — the pattern is familiar enough that it would be naive to ignore it.

The employment gap data matters here too: disabled people are statistically more likely to be priced out when costs rise, and more likely to be overlooked when priorities shift. That is not a reason to avoid useful tools now — it is a reason to stay aware, keep options open, and not build a workflow so dependent on a single paid service that losing access to it becomes a crisis.

For now, Claude works. It removes barriers that are otherwise slow and effortful to get around. That is enough of a reason to use it — as long as you are aware of what might change.