I have walked past more kiosks than I can count. In museums, airports, restaurants, hospitals, just to name a few. Beautiful machines. Big touchscreens. Slick animations. And not one of them I could actually use. Not to buy a ticket, not to look up an object, not even to pay a bill.
It happened so many times. I went to the airport and just wanted to be independent like anybody else. I found the accessible control, plugged in my headset, and nothing. I went to a fast food restaurant, and I had specifically checked in advance which one has an accessible kiosk. It was there, it started speaking, but I could not navigate the menu or restart the system. I went to the hospital with a friend who was driving, and I wanted to pay for the parking ticket. It not only had no voice output, but the system was so complicated that even the two of us together could not figure it out, and we gave up. I went to several museums with the promise of an accessible kiosk, and it did not work.
This post is about these types of situations. No lecture on why accessibility matters, just a working definition, the legal requirements, and a test you can run this afternoon. It is also the first in a series. Here I stay with the concepts, and the deeper technical pieces will get their own articles.
A working definition
An accessible kiosk is a kiosk every visitor can operate independently. Sighted or blind. Hearing or deaf. Full mobility, limited mobility, cognitive disability. Every visitor, independently. That last word is doing all the work, and I will come back to it.
Notice what the definition does not say. It does not say "a kiosk with a screen reader installed." It does not say "a kiosk that meets the ADA." Those are means, not the goal. The goal is that a visitor who cannot see the screen, cannot hear the audio, or cannot reach the top of the display can still complete the same task as everyone else, without asking anyone for help.
I will talk mostly about blind visitors in this post, because kiosks tend to fail blind visitors first and most visibly. But the same standard applies to every disability, the deaf visitor who needs captions, the wheelchair user who cannot reach the screen, the visitor with a cognitive disability who needs plain language and no timeouts. They matter equally; blindness is just my lens.
Two halves, and both have to work
A kiosk is two products in one: the physical hardware and the software. Accessibility fails on either half, and it usually fails on both.
The physical half is about the body. Can a wheelchair user roll up to the kiosk, with clear floor space in front of it? Can a seated visitor reach every control, including the top of the screen? Is there a standard headphone jack, and does plugging in headphones actually route the audio to the jack instead of the speaker? Is there a tactilely discernible keypad for entering numbers? Are the kiosk's edges detectable by a cane, or does the unit jut into the walking path as a hazard? These are settled questions, reach ranges, clear floor space, protrusion limits, tactile keypads are all specified in the ADA Standards and in EN 301 549, but I keep encountering kiosks that get them wrong, which tells me the specs are not reaching the people who design the exhibits. Mainstream kiosk setup guides, in my experience, treat accessibility as a throwaway line, one passing mention of an "ADA-compliant typeface", while the real requirements live in installer checklists that exhibit designers never read.
This hardware exists, and it is not exotic. Take a tactile keypad with a headphone jack that works with any screen reader. A six-key version lists at $300. When a vendor tells you this is some impossible custom engineering, that number is worth knowing.
And here is the interaction every accessible kiosk should offer. You plug in your headphones, and the accessible session starts: speech begins, a welcome message plays, and everything stays private through the jack. You unplug, and the session ends. No menus to find, no mode to enable, no staff to flag down. Plugging in the headphone is the whole interaction.
And this half is not only about blindness: a deaf visitor needs every spoken message available as text, and a visitor with limited dexterity needs controls that do not demand precise taps or pinches.
The software half is about the screen reader. Every function of the kiosk must be operable through speech output and a keyboard or tactile controls: every button labeled, every state announced, every error message spoken. Focus must move logically. Nothing may depend on a gesture a blind visitor cannot perform, like swiping to a precise spot on a flat screen. Timeouts are a classic failure: the kiosk resets to the attract screen after ninety seconds, which is plenty of time for a sighted visitor and nowhere near enough for someone who has to listen through every option before choosing. The same thinking extends to other disabilities: every audio message needs a visual equivalent for deaf visitors, and the interface should use plain language, consistent navigation, and generous timeouts for visitors with cognitive disabilities.
The independence test
Here is the test I apply to every kiosk: can a blind visitor, or any visitor with a disability, complete the task without asking staff for help?
Not "can staff help them." Not "is there a phone number to call." Can the visitor do it alone, start to finish.
Independence is the whole point. A kiosk exists to let visitors serve themselves, to skip the line, to browse at their own pace, to avoid bothering anyone. The moment a disabled visitor has to flag down a staff member to do what everyone else does alone, the kiosk has failed at its one job, for that visitor. Staff assistance is a backup, not a strategy. And it is worth saying plainly: many disabled visitors will not ask. They will just walk past your kiosk, the way I have walked past many.
The legal backdrop
You should know the rules that already exist, because your vendor certainly claims to.
In the United States, Section 508 of the Rehabilitation Act, section 402, covers "closed functionality": when a product does not allow the user to install their own assistive technology, which describes nearly every kiosk, the product itself must provide the accessibility. Speech output has to cover everything on screen, and there has to be volume control.
The ADA Standards, section 707, set requirements for self-service machines including speech output, tactilely discernible input controls, and privacy of the audio (section 707.4): a blind visitor's transaction must not be broadcast to the lobby. I like the bluntness of section 707.5: speech shall be recorded or digitized human, or synthesized. Recorded audio and synthesized speech are both explicitly allowed.
In Europe, EN 301 549 section 5.1 covers closed functionality the same way, and section 5.1.3.2 says the audio has to work through a standard headphone jack, without requiring the use of vision. The jack itself has to be findable by touch. The European Accessibility Act is bringing enforceable requirements to self-service terminals across the EU, and it has covered them since June 2025.
One thing worth knowing, especially for museums: a city or county museum answers to Title II, which covers state and local government programs. A private museum answers to Title III, as a place of public accommodation. Figure out which one your museum sits under before you argue with anyone.
I am not your lawyer, and whether each of these applies to your museum depends on your funding and jurisdiction. My point is narrower: the technical requirements are already written down. You do not have to invent them. When a vendor tells you accessible kiosks are impossible or exotic, they are telling you they have not read the standards.
This is not meant to be a thorough legal review, just the references that matter most. We will unpack them properly in another post.
What it is not
An accessible kiosk is not a touchscreen with a screen reader bolted on as an afterthought. I have heard this many times, and it always ends the same way.
The screen reader is installed, but it talks over the attract-loop audio. Or it reads, but the "Begin" button is an unlabeled image, so there is nowhere to go. Or everything works until the payment step, which is a third-party widget the screen reader cannot see. Or, my personal favorite, the speech comes out of the kiosk's speaker at full volume, announcing the visitor's purchase to everyone in the store.
Bolt-on accessibility fails because the kiosk was designed for eyes first and ears never. Accessibility has to be in the design from the first sketch: in the choice of controls, in the information architecture, in the audio routing, in the timeout policy. Retrofitting is always more expensive and always worse. If you are buying, buy it accessible. If you are building, build it accessible.
The spectrum: information, ticketing, payment
Not all kiosk tasks are equally demanding, and you should know where yours sits.
An information kiosk, what is on this floor, where is the restroom, is the lowest bar, and even it requires labeled controls, speech output, and sane timeouts.
A ticketing kiosk raises the bar: selecting a date, a time slot, a ticket type, a quantity. Every option must be announced, every selection confirmed, every error recoverable without starting over.
A payment kiosk is the hardest. Now you need a tactilely discernible keypad for PIN entry, audio privacy through the headphone jack so the transaction is not overheard, and a payment flow that a screen reader can complete, which rules out most off-the-shelf payment widgets. Payment is where I see the most failures, and it is also where the legal exposure is sharpest, because a blind visitor who cannot pay independently is being denied the service.
Scope your requirements to your task. Just be cautious when a vendor sells you an "accessible information kiosk" and then bolts a payment step onto it later.
Who is responsible? Whoever writes the purchase order.
The venue blames the vendor, the vendor blames the platform, the platform blames the payment provider. Meanwhile the kiosk sits in the lobby, unusable.
Put bluntly: the leverage is in procurement. Vendors respond to purchase orders, not to goodwill. If accessibility is not in the RFP, it will not be in the product. So it belongs in the RFP. Ask for a current Accessibility Conformance Report (the document formerly known as a VPAT) against Section 508 and EN 301 549, and verify it, because I have read conformance reports that describe a different product than the one being sold. Make acceptance testing with disabled users a condition of final payment, not an afterthought. Name the standards by number: 508 section 402, ADA section 707, EN 301 549 section 5.1. Vendors who know what those numbers mean are the vendors you want.
And if you already own kiosks that fail: retrofit where you can, replace on a schedule where you cannot, and avoid buying another inaccessible one. Every purchase is a vote.
A five-minute self-test
You can run this on any kiosk in your building this afternoon. You do not need me, and you do not need a lab.
Pull up a chair, and try to reach every control, including the top of the screen.
Find the headphone jack. Plug in headphones. Does the audio switch to the jack, or does it keep coming out of the speaker? On a well-built kiosk, plugging in should also start the accessible session, with speech and a welcome message. Unplugging should end it.
Turn the screen away, or close your eyes, and try to complete the kiosk's main task using only the audio. Can you? (If your kiosk has no audio interface at all, you already have your answer.)
Is every button and option announced with a clear label, not "button", not silence?
Does the kiosk time out while you are still working? A timeout that fires while a visitor is listening is a denial of service.
If the kiosk takes payment: can the transaction be completed without the details being visible or audible to bystanders?
If your kiosk fails any of these, you do not have an accessible kiosk. You have a project.
Most kiosks fail, and that is not a technology problem. The standards exist. The hardware exists: headphone jacks, tactile keypads, screen readers that run in kiosk mode. What is missing, in most cases, is a purchase order that required it.
My suggestion: test your kiosks this week, if you can. Run the five-minute test, note what fails, and put accessibility in writing the next time you buy anything with a screen on it. Your visitors with disabilities are already in your building. The question is whether your kiosk works for them, or just sits there looking modern.