Intelligent Discovery AI Assistant is currently available as a Public Beta product and information contained in this document is subject to change. Public Beta products are not covered by a Twilio Service Level Agreement. Learn more in the "Twilio Beta Product Support" Help Center article.
The Intelligent Discovery AI Assistant is an AI-driven assistant enabling users to interact with messaging data using natural language. The Assistant helps you troubleshoot messaging-related issues, including deliverability errors. It can assist with data analysis and dive deeper into your messaging data to find potential root causes. The Assistant provides recommendations to solve messaging errors and potential issues. Customers can provide feedback on each response and the overall Assistant experience. Customers can seamlessly request to talk to a live Twilio support person if they still face unresolved issues.
Why Intelligent Discovery?
Twilio customers often face challenges locating relevant information within the Twilio Console. The current troubleshooting process can be time-consuming, potentially leading to support tickets and business disruptions. This can introduce significant costs, delays, and loss of customer trust. The longer an issue persists, the higher the cost and the greater the risk to customer relationships.
There is a clear demand for a self-serve utility that allows businesses to continue operations smoothly before resorting to support tickets. Such an assistant increases the reach of timely help to everyone, enabling customers to resolve issues faster and manage problems without losing the trust of their end customers.
Intelligent Discovery caters to both non-technical business users and technical users.
How Intelligent Discovery can help you
The Assistant can assist you with a variety of tasks related to Twilio's messaging services. Here are some of the things it can do for you:
Natural language processing and support
Take questions in natural language and answer in natural language.
Start with one of our guided flows or feel free to ask questions directly using the chat feature.
Here are some examples of the potential questions that the Assistant can support:
How many messages that I sent to Canada in the past 7 days have failed?
What are the messages which are failing because of the error code 30034?
Identify message delivery problems: The Assistant can analyze your messaging data to identify issues with message delivery, such as failed or undelivered messages. It can answer questions related to Channels (SMS, MMS, WhatsApp, RCS, etc.), Messaging services, Countries, Phone numbers, Subaccounts, Message Status, Carriers, etc.
Top error codes: Provide a list of the most common error codes affecting your messages.
Country-specific issues: Identify countries where your messages are facing delivery problems.
Phone number issues: Highlight specific phone numbers that are experiencing delivery issues.
Subaccount analysis: Analyze messaging issues specific to your subaccounts.
Retention: The Assistant supports analysis on the messaging data generated in the last 30 days.
Context: The Assistant carries the context of previously asked questions in order to formulate the next response.
Provide recommendations
Actionable recommendations: Offer advanced and personalized recommendations to resolve specific issues identified in your message history.
Knowledge base lookup: Search the Twilio knowledge base for solutions to specific error codes or issues. Learn more about Recommendations.
Get issues resolved with live agent's help
Human agent: Twilio customers with certain support plans are able to get human help if the AI Assistant cannot resolve all of their issues. They can connect seamlessly with a human agent without leaving the Assistant window. All of the Assistant interaction history is accessible to the human agents, resulting in faster issue resolution.
Continue interaction in the Twilio Help Center: Twilio customers can transfer chat history to the help center seamlessly and continue the chat in the help center if a human agent is not available in the Assistant. Customers can choose multiple options in the help center to take further actions based on their support plan.
Create support ticket: Customers can create quality support tickets with the help of the Assistant interaction history.
Contact support using phone.
Status updates
Twilio service status: Provide the current status of Twilio services, including any downtimes, outages, or maintenance.
General information
Product information: Answer questions about Twilio products and services.
Best practices: Share best practices for using Twilio's messaging services effectively.
Get started
To access this feature, users should click on the Try Intelligent Discovery Beta button in the upper middle of the Messaging Insights Console page. This action opens a chat interface where users can interact with the Intelligent Discovery Assistant.
Key features
Rapid execution: Sorted summary tables are part of responses so that users can take actions faster.
Trust and transparency: The "What's showing" button provides the query equivalent to offer transparency to customers.
Assistant boundaries: Legal boundaries are maintained.
Quick navigation to home screen: A back button is available for quick navigation to the home screen.
Interaction history: If you are logged in, signing off deletes the history, and going to the home page also removes the history.
Various ways to provide feedback: Users can give feedback using thumbs up and thumbs down options, as well as overall session feedback.
Ease of use: The Assistant uses natural language responses.
Guided flows: The Assistant provides guided flows for ease of use.
Free form chat flows: Users can engage in free-form chat flows.
AI Nutrition Facts
Intelligent Discovery AI Assistant uses third-party artificial technology and machine learning technologies.
Twilio's AI Nutrition Facts provide an overview of the AI feature you're using, so you can better understand how the AI is working with your data. Intelligent Discovery AI Assistant qualities are outlined in the following Intelligent Discovery AI Assistant Nutrition Facts label. For more information and the glossary regarding the AI Nutrition Facts Label, please refer to Twilio's AI Nutrition Facts page.
AI Nutrition Facts
Intelligent Discovery AI Assistant
Description
The Intelligent Discovery AI Assistant is an AI-driven assistant enabling users to interact with messaging data using natural language. The Assistant helps you troubleshoot messaging-related issues, including deliverability errors. It can assist with data analysis and dive deeper into your messaging data to find potential root causes. The Assistant provides recommendations to solve messaging errors and potential issues. Customers can provide feedback on each response and the overall Assistant experience. Customers can seamlessly request to talk to a live Twilio support person if they still face unresolved issues.
Privacy Ladder Level
1
Feature is Optional
No
Model Type
Generative
Base Model
Twilio Proprietary + OpenAI GPT-4
Trust Ingredients
Base Model Trained with Customer Data
No
Customer Data is Shared with Model Vendor
Yes
This is only for interaction and intent aspect from the conversation that the user has with the Intelligent Discovery AI Assistant. Twilio does not share PII with the model vendor.
Training Data Anonymized
N/A
Data Deletion
Yes
Twilio only deletes conversation data created during the interactions with the Assistant.
Human in the Loop
Yes
Customers can opt to talk to a Twilio human agent if the Assistant is not able to solve their issues, but it is not required. The human agent can also collect feedback to improve the Assistant
Intelligent Discovery for Messaging Insights democratizes actionable insights from messaging data, making it accessible to both developers and non-technical business users. By leveraging AI-driven natural language processing, this feature simplifies troubleshooting and enhances the ability to gain valuable insights from communications data. Here are some possible next steps and additional resources to help you get started: