How can I support both voice input and buttons for the same question in the new Agent canvas?
Hi everyone 👋 I’m building a questionnaire in the Agent canvas where seniors can either speak their answer or click a button (e.g., “Yes, no change / Eating a bit less / Much less / Clarify / Repeat”).
Right now, my Q1 block only allows me to connect to one next block (either Buttons or a Prompt). I’d like both to run in parallel so users can choose voice or button without duplicating the question.
• Is there a way to have a Prompt block show buttons (Quick Replies) so I can combine both?
• Or is the only option to duplicate the block into two flows (one for voice, one for buttons)?
Would love any guidance or examples 🙏
2 Replies
Hey there! You don’t have to duplicate the question – you can let users either speak OR click the same options in a single flow.
1️⃣ Use a Choice step (preferred)
• Drop a Text/Speak step that asks the question.
• Follow it with a Choice step.
• Add a trigger for each answer ( “yes, no change”, “eating a bit less”, … ).
– Attach an intent/utterances to each trigger so the agent recognises spoken input. – In the same trigger, click ➕ under Buttons and type the label so the web-chat shows a button. Now the same step will fire whether the user clicks or says the phrase – no parallel paths needed. 2️⃣ Keep a Button step + turn on “Listen for other triggers” • Add your buttons as normal. • Toggle Listen for Other Triggers → ON.
• Create intents for the spoken equivalents in your NL-model.
While the Button step is waiting, the agent can still match those intents, so voice & click both work. 🔹 What won’t work: the Prompt/Text/Speak step itself can’t display quick-reply buttons, so chaining Prompt → Buttons is the only way if you use Prompt alone. So pick either Choice step (all-in-one) or Button step with “listen” enabled, and you’ll support both modalities without duplicating the question. Docs reference: Choice step guide – https://docs.voiceflow.com ("Choice step" section) Button step guide – https://docs.voiceflow.com ("Button step" > Listen for Other Triggers)
– Attach an intent/utterances to each trigger so the agent recognises spoken input. – In the same trigger, click ➕ under Buttons and type the label so the web-chat shows a button. Now the same step will fire whether the user clicks or says the phrase – no parallel paths needed. 2️⃣ Keep a Button step + turn on “Listen for other triggers” • Add your buttons as normal. • Toggle Listen for Other Triggers → ON.
• Create intents for the spoken equivalents in your NL-model.
While the Button step is waiting, the agent can still match those intents, so voice & click both work. 🔹 What won’t work: the Prompt/Text/Speak step itself can’t display quick-reply buttons, so chaining Prompt → Buttons is the only way if you use Prompt alone. So pick either Choice step (all-in-one) or Button step with “listen” enabled, and you’ll support both modalities without duplicating the question. Docs reference: Choice step guide – https://docs.voiceflow.com ("Choice step" section) Button step guide – https://docs.voiceflow.com ("Button step" > Listen for Other Triggers)
thanks for the answer. I have already setup the two separate choices so was hoping to be able to not have to redo the steps. If there is no way to do it as two parallel steps I will have reinput it all.