Powsoo is a drop-in widget that adds visual intake cards to AI voice platforms. When a user is on a voice call with your agent, Powsoo renders synced cards in the browser so users can tap-to-confirm emails, phone numbers, dates, and other intake fields instead of spelling them out.

How do I add Powsoo to my voice platform?

Create a widget in the Powsoo dashboard and connect your Retell account, then drop one web component and one versioned script tag into your site. You add powsoo_show_card as a custom function on your agent. Setup takes about 5 minutes.

Which voice platforms does Powsoo support?

Retell today. Vapi and ElevenLabs support are coming soon. SMS fallback works for traditional PSTN calls.

Can I customize the look of the cards?

Yes. Powsoo supports full theming via JS configuration: colors, fonts, border radius, spacing. For deeper control, use Powsoo Atoms (React components) to build your own UI.

Where does the data go after a user confirms?

Powsoo fires webhooks the moment a card is confirmed, with the field name, value, and call session ID. Drop the payload straight into your CRM, database, or backend.

Does Powsoo work for phone calls?

Pure phone calls (PSTN) don't have a screen, so Powsoo's web widget can't render directly. Instead, Powsoo offers an SMS fallback: when the agent needs to confirm an email or phone number, the caller gets a one-tap SMS link.

Is Powsoo HIPAA / GDPR compliant?

Powsoo is built with healthcare and EU customers in mind. We're SOC 2 Type II in progress and offer DPAs and BAAs on the Pro plan and above.

Add tap-to-confirm to a Vapi AI voice agent

Updated 2026-07-20

A Vapi agent runs the conversation, but it still captures structured fields (emails, phone numbers, dates) by ear, and that is where speech-to-text fails. Independent testing puts voice-AI email capture at roughly 53% correct, and even at a strong 98% per-digit accuracy a ten-digit phone number is only about 82% right, so nearly one in five callers gives a number with a wrong digit. A visual confirmation layer fixes this by letting the caller read and tap instead of spelling values aloud. Powsoo ships this today for Retell; Vapi support is coming soon. When it lands, the agent will call powsoo_show_card with the transcribed value, the caller will tap Confirm or edit a character, and the confirmed value will flow to your backend. The caller reads instead of listens, so a wrong character is obvious at a glance.

Where a Vapi agent loses data

Vapi gives you a composable orchestration layer: your choice of speech-to-text, model, and text-to-speech, wired into a conversational agent. That flexibility is its strength, but the capture problem is the same across every stack. Speech-to-text is tuned to minimise word error rate, which means it protects common words and leaves rare tokens (names, emails, account numbers) exposed.

The numbers bear this out. Best-in-class models reach 95–98% word accuracy on clean audio, but that drops on the accented, noisy, telephone-bandwidth audio real calls actually carry. For an email address, standard voice-AI capture is right only about 53% of the time. For a ten-digit phone number, even 98% per-digit accuracy compounds to roughly 82% fully correct. The agent then reads the wrong value back confidently, and the caller has to catch a single wrong character by ear.

Vapi support is coming soon

Powsoo’s Vapi adapter is on the roadmap but not yet available to connect. It will ride on Vapi’s tool-calling, so you will not rebuild the agent: connect your Vapi account in the Powsoo dashboard, add a powsoo_show_card tool to the assistant’s tools array (with async off so it waits for the tap), and drop the <powsoo-voice> web component plus the widget script onto the page hosting the call. When the assistant calls the tool, a card renders in the browser, the caller taps Confirm or edits one character, and the confirmed value flows to your backend.

Want it today? Powsoo already works with Retell, using the same tap-to-confirm pattern.

Which fields to confirm

Email addresses: the highest failure rate and usually the most costly to get wrong.
Phone numbers: digit strings where speech-to-text confuses similar-sounding digits.
Dates: ambiguous spoken forms like “next Tuesday” resolved to a concrete, tappable date.

Frequently asked questions

Can I connect Vapi to Powsoo today?

Not yet. Vapi support is coming soon. Powsoo works with Retell today. When Vapi lands it will not replace your agent: Vapi keeps running the conversation, and Powsoo handles the moment a structured field needs confirming by rendering a card the caller taps.

How accurate is voice-only data capture on Vapi?

It depends on the field. Grammar-heavy speech transcribes well, but rare tokens fail: research puts spoken-email capture around 53% correct and shows that 96.6% word accuracy can still mean only 77% exact match on a full identifier. One wrong character breaks an email or a phone number.

What about phone-only Vapi calls with no screen?

A pure PSTN call has no browser to render into. For those, use an SMS fallback: the agent texts a one-tap link so the caller confirms on their phone. See the tap-to-confirm vs SMS fallback guide.