For people living with vision impairment, accessibility is often associated with familiar tools such as magnification, larger text, higher contrast, dictation, and screen readers. All of these tools remain valuable, but they represent only part of what is now possible. Some of the most useful solutions were never designed as assistive technology at all. Mainstream devices, utilizing the ever-increasing capabilities of artificial intelligence, are creating new ways to reduce visual demand, simplify everyday tasks, and increase independence. Perhaps accessibility is no longer just about making something easier to see, but finding a different way to interact with it.
Rethinking accessibility
Traditional accessibility often focuses on adapting an existing interface: making text larger, increasing contrast, or else adding speech. But another approach is to change the interaction itself.
As my own vision has changed, I have increasingly looked for ways to reduce the visual input a task requires. Sometimes that means replacing a series of mouse movements with a single button press; sometimes it means using audio instead of looking at a screen. In other situations, automation can remove several steps altogether.
Vision impairment affects more than the ability to read small text. It can increase visual strain, slow navigation, and make complex digital environments more demanding.
The aim, then, is not simply to make technology more visible. It is to make the task itself less visually demanding.
Changing the interface
Sometimes the most effective solution is not replacing an interface but reshaping it. Many websites may technically meet accessibility standards while still being difficult to use with limited vision. Small buttons, subtle borders, and densely packed menus can all remain challenging to users who are visually impaired, even with magnification.
One of my most useful discoveries has been a browser extension that allows me to modify how individual websites appear on my screen. By changing the underlying CSS (the code that controls how a webpage looks), I can enlarge controls, strengthen borders, increase spacing, or even make an element more visible when the pointer moves over it. The website itself remains unchanged; I have effectively created a personalized accessibility layer over it.
Similar thinking can be applied elsewhere. Large, highly visible widgets can replace small smartphone icons, while carefully positioned windows on a larger monitor can reduce unnecessary searching and navigation.
Rather than accepting an interface as it arrives, increasingly we can tailor it to our own visual needs.
From screen to touch
Reshaping an interface can make it easier to see, but another option is to move some interaction away from the screen altogether.
I discovered this almost accidentally while looking for physical buttons that could launch applications on my computer. Devices such as Stream Decks, popular with broadcasters and content creators, allow buttons to be assigned to specific actions. MIDI controllers used in music production offer something similar, with tactile knobs, dials, pads and buttons providing physical alternatives to small on-screen controls.
Even something as simple as copy and paste can benefit. Rather than locating small on-screen controls, I have permanent copy and paste buttons assigned to my Stream Deck, reducing repeated mouse navigation across almost every application I use.
The same principle can be applied in more specialized environments. In my own music production, digital audio software can involve complex screens filled with small menus and controls. By mapping commonly used functions to physical controllers and touch surfaces, around 80 per cent of my workflow can now be performed without navigating those interfaces visually.
When a repeated digital action can be assigned to a distinctive physical control, touch and muscle memory can begin to replace visual navigation.
Automating the everyday
Physical controls can reduce the need to navigate an interface, but automation can go further by removing some steps entirely.
On my computer, I use automation software to combine sequences of actions into a single keyboard shortcut. One command can open several applications, position their windows across my monitors, and launch the websites I regularly use. What previously required multiple visually demanding actions becomes one easy repeatable process.
The same principle works particularly well on a smartphone. Apple's Shortcuts app allows a series of actions to be triggered by a large widget, button or voice command. A shortcut might start a favourite playlist and connect to headphones, while Siri can read a message without requiring me to look at or handle the phone directly.
These may sound like small efficiencies, but their value accumulates. Removing repeated visual navigation can reduce effort across work, communication and everyday life. Once established, automation can turn complicated digital routines into predictable actions requiring very little visual input.
Case study: the weekly shop
Almost two years ago, I created an AI-driven iPhone shortcut to simplify a routine grocery delivery. Sainsbury's sends a long confirmation email listing the order, substitutions, unavailable products and items with a shorter shelf life.
With one button press, the shortcut extracts that information, uses AI to identify the relevant changes and reads the result aloud in under ten seconds. There is no need to visually navigate or listen through the original email.
I built the shortcut when I still had considerably more central vision and have never needed to modify it since. As my vision has deteriorated, its value has increased. Sometimes the most effective adaptation is one that is already in place before it becomes essential.
When listening replaces looking
Reducing visual interaction raises another possibility: replacing some of it with sound.
Dictation and text-to-speech now form a large part of my daily workflow. I dictate emails, messages and longer pieces of writing, while natural-sounding text-to-speech allows me to listen rather than continually read from a screen. Browser tools such as Speechify can extend this approach to websites and AI platforms.
Hardware matters too. I use an open-ear SHOKZ headset with a microphone for dictation and listening. Because my ears remain uncovered, I can stay aware of my surroundings while working or moving through everyday activities.
Combined with voice commands and physical shortcuts, this creates an audio-first workflow. The screen remains available when I need it, but it is no longer the default route to every piece of information.
For someone with limited vision, listening can become more than an accessibility alternative; it can become an efficient way of interacting with the digital world.
Do I need the interface at all?
The next step may be to remove the interface from some interactions altogether. AI connected to services such as email and calendars increasingly allows natural language to replace visual navigation. I can ask when my next appointment is, find space in my calendar, identify important emails, summarise a message or draft a response without opening the applications themselves.
This does not make human judgement less important. AI is most useful when we decide what information matters, what task needs simplifying and what we want the technology to do. For people with vision impairment, its value lies partly in reducing unnecessary visual and cognitive workload.
None of this means abandoning magnification, screen readers or traditional assistive technology. It means widening our understanding of what accessibility can be, and incorporating the features that work individually for us into our daily lives.
Perhaps the question is no longer simply, "How can I access this interface?" Sometimes the more useful question is, "Do I need to interact with the interface at all?"
