BotFramework-WebChat
BotFramework-WebChat copied to clipboard
[Screen Readers-Ask HR Virtual Assistant]: Screen readers(NVDA/JAWS) are narrating the suggestion invoked by the user multiple times upon invoking the Suggested Questionnaire button present in the “Ask HR Virtual Assistant” chat window.
Is it an issue related to Adaptive Cards?
- [X] Yes, this is an Adaptive Card issue but it is specific to Web Chat.
Do you have any screenshots?

What version of Web Chat are you using?
Latest production
Which distribution are you using Web Chat from?
Bundle (webchat.js)
Which hosting environment does this issue primarily affect?
Web apps
Which browsers and platforms do the issue happened?
Browser: Edge (latest)
Which area does this issue affect?
Chat history
Is this an accessibility issue?
- [X] Yes, this is an accessibility issue.
Please describe the bug
PWD Impact: Screen reader users are getting impacted if Screen readers (NVDA/JAWS) are narrating the suggestion invoked by the user multiple times upon invoking the Suggested Questionnaire button present in the “Ask HR Virtual Assistant” chat window.
Test Environment: OS: Windows 10 version 21H1 (OS Build 19043.1110) Browser: New Edge Version 92.0.902.62 (Official build) (64-bit)/Chrome Version 92.0.4515.107 (Official Build) (64-bit) URL: AskHR Virtual Assistant (azureedge.net)
Repro Steps: Step 1: Enable Screen readers (NVDA/JAWS) and open the above URL in the Chrome/Edge Browser using valid credentials. Step 2: 'AskHR Virtual Assistant' Home page gets opened. Step 3: Navigate to the list items suggestions using down arrow key and select Benefits verify the screen readers (NVDA/JAWS) behavior upon invoking the suggestion questionnaire button present in the “Ask HR Virtual Assistant” chat window.
Actual Result: Screen readers (NVDA/JAWS) are narrating the suggestion invoked by the user multiple times upon invoking the Suggested Questionnaire button present in the “Ask HR Virtual Assistant” chat window.
Ex: NVDA/JAWS narrating as "Suggested Actions Container: Is empty………………………………complementary landmark form landmark form landmark message input box edit type your message blank Suggested Actions Container: Has content Press Alt+ Shift + A to select Laya Pension (Mercer)Well being program ESPP other benefits None of these/ask a different question "
Expected Result: Screen readers(NVDA/JAWS) shouldn't narrate the suggestion invoked by the user multiple times upon invoking the Suggested Questionnaire button present in the “Ask HR Virtual Assistant” chat window.
Ex: NVDA/JAWS should narrate as "Select an option to continue this conversation Suggested Actions Container: Has content Press Alt+ Shift + A to select 6 list items button Laya, button Pension(Mercer), button Well being program, button ESPP, button other benefits, button None of these/ask a different question"
MAS Reference: MAS 4.2.1 – Object Information
Do you see any errors in console log?
No response
How to reproduce the issue?
- Navigate to ...
- Click on ...
- Type "Benefits" in the send box
- select "Benefits enrollment" button.
- observe the
What is the expected and actual behavior?
Expected: Screen readers(NVDA/JAWS) shouldn't narrate the suggestion invoked by the user multiple times upon invoking the Suggested Questionnaire button present in the “Ask HR Virtual Assistant” chat window. Ex: NVDA/JAWS should narrate as "Select an option to continue this conversation Suggested Actions Container: Has content Press Alt+ Shift + A to select 6 list items button Laya, button Pension(Mercer), button Well being program, button ESPP, button other benefits, button None of these/ask a different question"
Actual: Screen readers (NVDA/JAWS) are narrating the suggestion invoked by the user multiple times upon invoking the Suggested Questionnaire button present in the “Ask HR Virtual Assistant” chat window. Ex: NVDA/JAWS narrating as "Suggested Actions Container: Is empty………………………………complementary landmark form landmark form landmark message input box edit type your message blank Suggested Actions Container: Has content Press Alt+ Shift + A to select Laya Pension (Mercer)Well being program ESPP other benefits None of these/ask a different question
Adaptive Card JSON
No response
Additional context
No response
#A11yMAS;#A11ySev2;#HCL;#WCAG4.2.1;#Accessibility;#Win10-Edge(Chromium);#BotFrameworkWebChat-Aug21;#Linked:Bug8007103;#LinkedBug;
@ramfattah - Do we have any tentative fix ETA for this bug.
This is fixed in 4.15.3. I tested this on 4.15.4.
When a bot message arrived with a few suggested action buttons, say, "Yes, I am." and "No, I am not." NVDA will narrate:
"Bot said, please select one of the actions below. Yes, I am. No, I am not. Message has suggested actions. Press ALT SHIFT A to select them."
And in Narrator, it is very similar. Bolded are different.
"Bot said, please select one of the actions below. Yes, I am, button. No, I am not, button. Note, message has suggested actions. Press ALT SHIFT A to select them."
@compulim As verified issue is still reproducible with NVDA. NVDA is announcing the information twice.
@compulim As verified issue is still reproducible with NVDA. NVDA is announcing the information twice. Can you please check
@compulim Issue still exists using NVDA upon invoking the Suggested Questionnaire button is being announced more than once, i.e., twice announcement is observed for NVDA