modalities
One adaptive agent on every channel
Customers can call, chat, text or email you, and the same adaptive agent picks up every time.

Flexible pricing
Scale your voice agent with confidence across your organization and use cases
with a transparent pricing structure.
Voice, browser and in-app calls
The channel PolyAI started with. Every call runs on the same agent you deploy everywhere else.
- LIVEself-servevia ccaas
Voice
Answer and place calls on the numbers you already have, or buy new ones in Dialog Studio. Calls can route through the contact center you already run.
- Buy numbers in Agent Studio or bring your Twilio numbers
- SIP connection to Genesys, Five9, NICE CXone, Amazon Connect, Zoom and Dialpad
- Outbound calls for reminders and follow-ups
- Handoff to your live agents over SIP
- LIVEself-serve
Web calling
A click-to-call button on your website. Visitors talk to your agent in the browser, with no number to dial and nothing to install.
- Browser audio over WebRTC
- One script tag, or add it through Google Tag Manager
- Branding, copy and consent set in one editor
- Hosted preview link for testing before launch
- LIVEDeveloper-led
In-app voice
Live voice calls inside your iOS or Android app, answered by the same agent that runs your in-app chat.
- Two-way WebRTC calls
- PolyVoice for iOS, ai.poly:voice for Android
- You build the UI, PolyAI runs the agent
- Works alongside the in-app messaging SDK
Chat on your website and in your app
Text conversations where your customers already are, with rich answers and handoff to your live chat team.
- LIVEself-serve
Web chat
A chat widget for your website that runs the same agent as your phone line, with room for richer answers than voice allows.
- Images, videos and interactive charts in the chat
- Localised for customers in multiple languages
- Signed-in identity passed in with verified context injection
- Handoff to Salesforce, Zendesk, NICE CXone, Amazon Connect, Webex and Five9
- LIVESelf-serve for developers
In-app messaging SDK
Headless libraries for Swift and Kotlin. You design the chat experience in your app and PolyAI provides the agent underneath.
- Responses stream in word by word
- Sessions resume across app launches and dropped connections
- Handoff events so your app can show queue status
- Install with Swift Package Manager, CocoaPods or Maven Central
- LIVEDeveloper-led
Messaging API
Connect the agent to any messaging surface you run yourself, including Genesys Cloud through Open Messaging.
- WebSocket events for messages and handoff
- Genesys Cloud Open Messaging, with handoff to an ACD queue
- Verified customer context at the start of each session
- The same knowledge and flows as every other channel
Messages on the customer's phone
Two-way conversations over SMS, RCS and soon WhatsApp, started by the customer or by your CRM.
- LIVESelf-serve in Agent Studio
Two-way SMS
Customers text your number and get answers from the agent that picks up your calls. You can also start conversations from your CRM.
- Inbound, outbound and one-way triggered messages
- PolyAI provisions and runs the Twilio setup for you
- STOP, START and HELP keywords and opt-outs handled
- Built for TCPA, GDPR and PECR, CASL and DLT rules
- LIVESelf-serve for developers
RCS
Branded business messaging on Android and iOS, with images, rich cards and reply buttons your customers can tap.
- Rich cards, carousels, images and video
- Suggestion chips that feed straight into the agent's flow
- A verified, branded sender profile
- Falls back to SMS when a device or carrier can't receive RCS
- Coming Oct'26Implementation, then self-serve
WhatsApp
Answer customers on WhatsApp Business with the agent that already handles your calls and chat.
- Inbound conversations on your WhatsApp Business Account
- The same knowledge base and guardrails as voice
- Opt-in and opt-out handled in line with Meta policy
- Images and PDFs in the conversation
Your support inbox, answered
Email is often the last channel a support team still works entirely by hand. PolyAI now reads it too, and it works on the mail setup you already have.
- NEWImplementation, then self-serve
Email
Connect your support address and the agent replies to routine emails on its own. Anything it can't settle reaches your team as a prepared case, with a summary and a suggested reply.
- Works with Microsoft 365, Google Workspace, Zendesk and Salesforce
- Replies come from your own domain and thread properly
- Keeps the context of a thread for two weeks when customers reply
- Every inbound email is screened for spam and viruses
- For now, emails with attachments go straight to your team

Security and data
Email often carries personal details, so here's how it's handled
Stored in your region
Email is kept in the AWS region that matches where you're registered: US, UK or EU.
Checked on the way in and out
Inbound mail is screened for spam and viruses. Replies are DKIM-signed and sent from your own domain.
Every decision on record
An audit log keeps each email, each reply and each routing decision. ISO 27001, SOC 2 Type II and HIPAA.
Talk to customers in their own language
75 languages available on the PolyAI platform, 45 with full end-to-end voice support, and 18+ live in production today. One agent can handle several languages from a single project. Content is translated automatically, and your team can override any translation by hand.

Already answering for some of the busiest service teams
35,000
labor hours saved by handling outage calls with a PolyAI agent.
391%
return on investment over three years, with payback in under six months.
Customers can switch channels without starting again
Most companies run a separate bot on each channel, and none of them know what the others said. PolyAI runs one adaptive agent across all of them, so it already knows the customer when they switch.

Get to know our platform
Wren works alongside the rest of the stack. Here is where to go next.
- PlatformExplore the studio
Dialog Studio
PolyAI's no-code builder for voice agents and conversational AI, and where Wren lives and works. Design and orchestrate every conversation: flows, tools, and brand voice, from no-code to full code without changing platform.
- Latest modelMeet the model
Dialog-RSN-1
The audio-native frontier model behind the conversations Wren builds. It reasons directly over raw call audio and responds in under 300ms in production.
- DevelopersRead the docs
ADK and APIs
Agents as code. Build and version with the Agent Development Kit, connect tools and data over MCP and REST, and run our models as an API endpoint. Everything Wren does, scriptable.
Works with the contact center you already run
Calls, chats and messages hand off to your live agents in the tools they already use. Nothing needs to be migrated.
Explore integrations
Frequently asked questions
Which channels are live today?
Phone, web calling, web chat, the in-app SDKs, the Messaging API, two-way SMS and RCS are live now. Email is new, and WhatsApp is planned for 2026.
Do I need a separate agent for each channel?
No. You build one agent in Agent Studio and switch channels on for it. Each channel can have its own greeting and formatting, but the knowledge, flows and guardrails are shared.
Does the agent remember customers when they change channel?
Yes. Agent memory recognises a returning customer by their phone number or email address and keeps the key facts from earlier conversations for up to 90 days, so they aren't asked the same questions again.
Can the agent hand a conversation to my team?
Yes. Calls hand off over SIP or through your CCaaS, and chats and texts hand off to Salesforce, Zendesk, NICE CXone, Amazon Connect, Webex, Five9 or Genesys Cloud. Emails the agent can't resolve arrive as a prepared case.
How is it priced?
Every channel draws on the same usage balance, so you don't buy each one separately. A messaging or email session costs the same as a minute of voice, it's only billed when the customer replies, and you can move volume between channels freely. Carrier fees for SMS, RCS and WhatsApp are passed through at cost.
How long does it take to go live?
Web chat and web calling are a script tag, so you can test them the same day. SMS depends on carrier registration, which varies by carrier. RCS needs your brand verified first, which can take a few weeks. Email needs a forwarding rule and a few DNS records, so your IT team's DNS change sets the pace.
How is customer data handled?
Conversations are stored in the AWS region that matches where you're registered: US, UK or EU. Inbound email is screened for spam and viruses, replies are DKIM-signed from your own domain, and every input, output and routing decision is kept in an audit log. PolyAI is certified to ISO 27001 and SOC 2 Type II and supports HIPAA.


