Which assistive tech do you predict?
Currently we predict all major screenreaders and keyboard only users via interaction analysis. This happens after a visit and sites can not take any action based on this prediction. While there are many types of assistive tech and the AT counts will not reflect them all, most technology is focused on these two use cases as they will have the biggest impact across the board.
Can you identify users?
No, they are 100% anonymous. We don't keep any information besides interaction sequence. In addition, we group user sessions together, effectively losing all identifying info to discourage cross comparison using other analytics tools.
Is it secure?
Nothing sensitive is sent over the wire. We mask all alphanumeric keys (and their length) and only send an event list of synthesized identifiers of interactive elements. Feel free to ask us for the runtime code and we'll send it over.
What if I don't have high traffic on my site?
ALLi will continue on working properly with low numbers, but you may need to wait a little longer for predicted concerns. We don't need a lot of data to create an AT simulation - just a couple dozen user journeys. Businesses cannot legally ignore a11y efforts due to a low number of AT users. We're very open about captured AT numbers being lower than in actuality.
Does this replace the need for audits and automation?
Not at all! We hope to be your new bug catcher and maintanence strategy for when things are missed. Manual audits are costly and time consuming. Automation only catches around ~30% of issues at best. Overlays are false advertising. We recommend a full manual audit and then a combination of Automated and ALLi.
