Build the future of communications.
Start building for free

Introducing Twilio Autopilot — A conversational AI platform to build bots that work


There’s a huge gap between the promise and reality of bots. Bots tend to get stuck in loops and fail to achieve the task at hand. If they do manage to break out of the loop and handoff to agents, they often lose the context of data collected to that point.

Today, developers have to choose between software-as-a-service bot builders focused on particular use cases with limited customizability or, natural language understanding services that are essential but not sufficient to build a complete conversational flow.

Today, we're announcing Twilio Autopilot — a conversational AI platform to build intelligent bots, IVRs, and Alexa apps that work.

Twilio Autopilot is made up of three building blocks:

  1. A Natural Language Understanding engine that analyzes user intent.
  2. A Conversational Application Platform to build sophisticated workflows.
  3. An Omnichannel Hub for broader reach across channels such as SMS, Voice, Whatsapp, FB Messenger, Alexa, and more.

Get started with Autopilot today.

A New Approach — Decoupling Business Logic and Style

We've modeled Autopilot separating business logic from style, similar to best practices for building web and mobile applications. Decoupling the two reduces the complexity of coding for all edge cases. Developers can focus on business logic on one side and styling/personalizing a bot on the other side. With this unique approach, developers can create bots that can have a tone a voice that defines their personality.

Autopilot features such as the Natural Language Router and Autopilot Actions handle business logic, while Stylesheets can be used to define the factors that give bots a personality.

Autopilot Features

A. Analyze intent with Natural Language Router to redirect issues

Some inquiries could be entirely self-service, such as, “I want to book a flight.” Others are partly self-service, like “I want to buy life insurance.” More critical inquiries such as “I need advice on what kind of mortgage I should take ” need immediate agent support. Autopilot's Natural Language Router takes customer issues and chooses the right task for the job.

B. Control the flow of conversations with Autopilot Actions.

Autopilot Actions are JSON commands or instructions that tell Autopilot what to do. With Autopilot actions, your virtual assistants can say something to the user, show something to the user if the device has a visual display, remember something for context, collect data, or redirect so you can control the flow of your dialogue.

Here are some examples of Autopilot Actions:

1. { Collect }

Data collection is one of the most common tasks that bots perform. The { Collect } action enables you to ask questions to users and efficiently collect answers.

Here’s an example of how to use { Collect } to collect answers to multiple questions:

    "actions": [
            "collect": {
                "name": "travel_questions",
                "questions": [
                        "question": "Do you like traveling?",
                        "name": "likes_traveling",
                        "type": "Twilio.YES_NO"
                        "question":"What's your favorite city?",
                        "name": "favorite_city",
                        "type": "Twilio.CITY"
                "on_complete": {
                    "redirect": ""

2. { Hand Off }

The Autopilot { Hand Off } action enables you to transfer a user from an Autopilot session to a human agent, passing along all the context of the conversation.

Here’s an example of how to use { Hand off } to transfer the Autopilot session to a Programmable Voice TwiML URL:

        "actions": [
                        "handoff": {
                                "channel": "voice",
                                "uri": ""

Build AI-Powered Assistants

Use Twilio Autopilot to build, train, and deploy bots that work.

Bots that are intelligent. Build intelligent IVRs, bots, and Alexa apps that are powered by Twilio built Natural Language Understanding and Machine Learning frameworks.

Bots that keep learning. Autopilot provides the training pipeline and interfaces to train your bots using real-time data, making your bot smarter over time.

Bots that can contextually handoff to a human. Smoothly transition a bot conversation to a conversation with agents when necessary, passing along the full context of the interaction.

Bots for every channel. Autopilot responses are adapted to provide the best experience on nearly any channel including IVRs (Interactive Voice Response), SMS, Chat, Alexa, Slack, and Google Assistant. Developers can build once and deploy across multiple channels instantly.

Join us on Nov 1st, at 10 AM PT for Keeping it Human: Bots, AI, and Customer Experience, a live webinar with the Autopilot team. We’re talking about delivering better customer service with artificial intelligence.

How Does it Work?

Introducing Twilio Autopilot — now in Beta

Autopilot uses a request/response model similar to TwiML. When a request comes in it performs the natural language analysis to route the user to a task. Tasks are configured with URLs that point to your application. Your application then responds using the Actions, and Autopilot does the work to make sure they work across all channels.

Getting Started

Want links? We’ve got links. To learn more, visit the Twilio Autopilot homepage. Autopilot’s pay-as-you-go pricing model ensures that you only pay for what you use.

We’ve also put together docs that include quickstarts and full API reference documentation so that you don’t get lost along the way ;)

If you want to see Autopilot in action, join us for a webinar on Nov 1st, 10 AM.

Head over to your Twilio Console to start using Autopilot.

We can’t wait to see what you build!












Sign up and start building
Not ready yet? Talk to an expert.