- Twilio Autopilot is now Generally Available.
- Autopilot lets you build, train and deploy AI-powered bots on any communications channel.
- Autopilot bots interface seamlessly with contact center software like Twilio Flex.
- The following features are now available as part of the General Availability of Autopilot:
- Bot Templates: Jump start your development with pre-built bot templates for common use cases and verticals.
- Simulator: Test and troubleshoot bot behavior. The Simulator is available in the console and supports testing over Voice and Chat.
- Custom Channels: Connect Autopilot bots to custom messaging channels like proprietary in-app chat or web chat.
- Event Webhooks: Monitor the performance of your bot in real-time with webhooks for key conversation events.
- Speech Optimization: Improve the accuracy of your IVR by optimizing the speech transcription for the data you’re expecting.
- New task training interface: A set of new training tools in the console to make it easier and faster to create and train new Tasks and Fields.
Bot Templates are a set of pre-built and pre-trained bots for common use cases. They allow you to create a fully functional bot from the console with a few clicks and zero code. We currently have templates for use cases like Appointment Scheduling, Lead Qualification and FAQs, and verticals like Banking, Insurance and Hospitality with more templates to follow over time!
Create a bot from a template with these easy steps:
- Go to the Autopilot console here.
- In the left hand navigation, go to Build a bot.
- Select the template of your choice and select create.
- In a few moments, you’ll see a fully functional bot in the console.
- Test the bot out using the Simulator!
The console Simulator is the quickest way to test and troubleshoot your Bot’s behavior during programming and training. It currently supports Voice and Chat.
- The Task tab in the chat mode shows you which task was triggered by each utterance. It lets you add or remove samples from the task or create new samples from the utterance if the inappropriate task was triggered, and displays any Fields detected in the utterance.
- The Memory tab displays the data stored in your bot’s memory over the course of the conversation. If you’re using the Collect Action, all data collected from the user will be displayed here.
- The Dialogue Info tab displays relevant metadata related to your conversation.
- Switching to the Debug View and selecting the Dialogue state for each message lets you play back the conversation to see how tasks are being triggered and data is being added to your bot’s memory.
Autopilot natively supports a number of messaging channels like SMS, WhatsApp, Facebook Messenger, Twilio Programmable Chat and Slack. With Custom Channels, you can connect Autopilot bots to proprietary channels like your own in-app chat or web chat applications, or third party messaging apps that Twilio doesn’t natively support yet, like Discord, Intercom, Telegram or Viber.
Learn how to use the Custom Channel endpoint in the docs.
You can now monitor the performance of your bot in real-time with webhooks for key conversation events. Simply subscribe to the events you’re interested in and configure a webhook URL. You can set up Event Webhooks through the console or programmatically through the API.
Autopilot currently supports the following events:
A user starts a dialogue with your bot.
A user ends the dialogue.
Fired each time a Collect question is answered.
Fired each time Autopilot requests the Actions JSON.
Fired each time an Action is executed. For example, if your action includes a ‘Say’ and a ‘Listen’, this event will be fired twice.
Learn more about Event Webhooks in the docs.
Speech Optimization optimizes the transcription generated by Twilio’s speech engine using the Built-In Field Types used in the training data for your bot. This improves the performance of your IVR by ensuring information in the user’s speech needed to complete their request, like ticket numbers or dates, are transcribed accurately. For example, if your IVR asks the user for their ticket number and the speech engine transcribes it as ‘zero oh o nine triple 3 seven’, ASR optimization ensures the ticket number you receive from Autopilot is 00093337.
It also reduces the chances of the wrong task being triggered. ASR Optimization is currently supported for these Built-In Field Types:
New Task Training Interface
Autopilot’s brand new training interface makes it faster and easier to train your tasks using Samples and Fields. We’ve improved the discoverability of Field Types in the console and made it easier to learn how to create Custom Field Types. We’ve also added examples to the interface to make it easier to learn how to use Samples and Fields when training your bot.
To start using the new training interface, click the Train button on a Task in the Autopilot console.
The road to GA
Since launching Autopilot in October 2018, we’ve also shipped a number of features to improve how Autopilot works with contact centers, increased the number of channels we support and improved the developer experience.
Here’s a full recap:
Autopilot <3 Contact Centers:
- How to build a chat bot with Autopilot
- How to hand-off messaging conversations from an Autopilot bot to your contact center
- How to build a conversational IVR with Autopilot
- Build WhatsApp bots with Autopilot
- Use Autopilot with Programmable Chat and Twilio Studio
- Build Facebook Messenger bots with Autopilot
- Build Google Assistant Actions with Autopilot
- Autopilot CLI.
- New Task Editor.
- <Connect><Autopilot> TwiML Verb.
- Field Validation.
Read about all of these features here.
We’re very grateful to our customers and the Twilio developer community for their feedback and support over the past year, and excited about continuing to make it faster and easier to build useful bots with Autopilot.
We can’t wait to see what you build!