Release Notes

You can now utilize system variables when configuring workflows without needing to pass these variables during API requests. System variables include key data such as date and time, caller phone number, and deployment type, offering valuable context and customization options for interactions.

• Enhanced Contextual Awareness: Leverage system variables to automatically integrate key data into workflows, providing a richer context for AI interactions.
• Ease of Use: Reduce the need for manual data passing and configuration by incorporating system variables directly into workflows.
• Customizable Interactions: Tailor user experiences based on dynamic information like call origin, interaction timing, and deployment context, improving relevance and user satisfaction.

This feature empowers you to build more context-aware, efficient, and dynamic workflows for your customers.

This new feature allows the AI agent to end a call from its side when specific phrases are spoken by the user. The End Call functionality is particularly useful for outbound deployments, handling voicemail interactions, automated messages from mobile providers, and other scenarios where users may request to be removed from calling lists or end the interaction.

• Improved Automation: Automatically detect and end calls based on pre-defined phrases, minimizing manual intervention and improving workflow efficiency.
• Voicemail and Automated Message Handling: Efficiently manage interactions with voicemail systems and automated messages, ensuring calls are concluded without unnecessary delays.
• User-Driven Interaction Control: Comply with user requests, such as those to end the call or be removed from calling lists, improving user experience and compliance with user preferences.

This feature empowers your AI agents to handle call interactions more intelligently and efficiently, reducing unnecessary call durations and improving spend.

This new non-interactive node type allows you to add background actions in your advanced workflows without directly interacting with the user. Non-Interactive Nodes enable your AI agents to perform actions—such as data processing, API calls, or variable updates—silently, without prompting user input or generating responses.

• Background Processing: Execute tasks behind the scenes, enhancing workflow efficiency and responsiveness without interrupting user interaction.
• Flexible Workflow Design: Integrate complex actions into your workflows while maintaining seamless user experiences.
• Improved Automation: Use this new node to automate background tasks, reducing manual intervention and improving workflow capabilities.

This addition empowers you to create more powerful and efficient AI-powered workflows that can handle a wide range of background tasks with minimal impact on user interactions.

Our platform now supports the use of variables in function calls, allowing for dynamic and context-driven data exchange between your AI agents and Internal/external systems. This enhancement enables greater flexibility and customization in workflows that require data manipulation and integration with external systems.

• Dynamic Data Handling: Use variables within function calls to dynamically pass data, improving the adaptability and responsiveness of your AI agents.
• Enhanced Customization: Tailor function calls to specific user interactions, providing more contextually relevant responses and actions based on variable values.
• Seamless Integration: Easily integrate complex data flows and interactions between your AI agents and external endpoints, enhancing the overall knowledge and functionality of your workflows.

This improvement empowers you to build more responsive and adaptable AI workflows tailored to a wide range of use cases.

Thank you!

Inbound Phone Calling Deployment: This deployment type allows you to configure and manage inbound phone calls, directing users to voice workflows through a dedicated phone number. This feature is ideal for handling incoming user requests and inquiries with tailored voice interactions.

Outbound Phone Calling Deployment: This deployment type enables customers to initiate outbound calls to users, connecting them to voice workflows. Outbound calls can be used for a variety of applications, including customer outreach, notifications, and automated support calls.

• Dedicated Phone Numbers: Each deployment is connected to a dedicated phone number, ensuring consistent and reliable communication.
• Integrated Voice Workflows: Both inbound and outbound calls are seamlessly integrated with voice workflows, allowing for customized, context-aware user interactions tailored to specific needs.
• Flexible Use Cases: Whether for inbound support, sales, outreach, or automated responses, these deployment types offer robust solutions for managing user communication effectively.

These new deployment options empower customers to leverage phone-based interactions more effectively, creating a deeper connection with their users.

Stay tuned for more updates and new features in the coming months.

Audio recordings of voice interactions are now directly accessible within the conversation logs. This feature allows for more detailed analysis and understanding of how your AI agents are performing in real-time voice interactions.

• Enhanced Insights: By accessing audio recordings, you can gain deeper insights into user interactions, uncovering areas for improvement or optimization in voice-based workflows.
• Improved Quality Assurance: Ensure high-quality user experiences by reviewing actual voice interactions and identifying any issues or opportunities for refinement.
• Seamless Access: Audio recordings are conveniently available within the conversation logs, making it easy to review and analyze without navigating away from the main interface.

This feature empowers to optimize voice interactions and ensure that AI agents provide the best possible experience.

This functionality allows to define general or system prompts that serve as foundational instructions for their AI-powered workflows. These prompts can include response guidelines, identity attributes, background information, and other overarching directives for the AI agent.

• Consistent Behavior: Ensure that all AI interactions follow a consistent set of general rules and guidelines, enhancing user experience and predictability.
• Simplified Workflow Creation: By leveraging default instructions, you can streamline workflow configuration and focus on detailed customization at the node level.
• Improved Flexibility: Adapt and modify default/system instructions as needed for different use cases while maintaining a baseline set of behaviors across workflows.

When configuring advanced workflows, these default instructions are automatically appended to each step of the workflow, enabling you to focus on fine-tuning specific node configurations without needing to repeat general setup details.

Thank you for your continued support!

This feature allows for call transfers from phone deployments to live agents, customer support teams, or other designated departments. With Transfer Call, your users can be effortlessly connected with human agents whenever needed, ensuring a smooth transition from AI-powered interactions to human support.

• Seamless Transfers: Enables to move from automated AI interactions to human conversations with minimal friction, improving user satisfaction and overall support quality.
• Flexible Routing: Calls can be transferred to various departments or teams based on user needs, providing a more tailored and effective support experience.
• Enhanced Customer Support: Empower your customer support operations by integrating live agent connections directly from your phone deployments, creating a more comprehensive support framework. This functionality helps you bridge the gap between AI interactions and human expertise, ensuring your users receive the best possible support experience.

New Features and Enhancements

Advanced Workflow Builder

• Node-Based Design with Drag-and-Drop Interface Build complex conversational flows effortlessly by connecting various nodes using our intuitive visual editor.
• Conditional Logic and Flow Control Implement conditions to direct the conversation based on user inputs, variables, or system states, enabling personalized and dynamic user experiences.
• Function Calls Integration Enhance your AI companions with custom and pre-defined function calls to perform specific tasks, retrieve data, or trigger external actions within your workflows.
• Variables Management Store and manipulate data throughout the conversation to personalize interactions and make informed decisions within the workflow.
• Enhanced Knowledge Base Integrations

Enable Retrieval-Augmented Generation to allow your AI companion to access and utilize external knowledge bases for more accurate and contextually relevant responses.

Improved Testing and Collaboration Tools

• Real-Time Testing Environment Test and debug your workflows in real-time using the updated Test Chatbot for text interactions and Test Voice Companion for voice interactions.
• Shareable Links for External Testing Generate shareable links to allow stakeholders, colleagues, or clients to test your AI companions without needing access to the OutRival platform.
• Detailed Logs and Analytics

Access comprehensive conversation logs, including transcripts and technical details, to optimize your AI companion’s performance.

Streamlined Deployment Options

• Multiple Deployment Channels Deploy your AI companions across various platforms:
  • Web Chat Widget
  • Text iFrame
  • Voice Widgets and iFrames
  • Inbound and Outbound Phone Calls
  • Upcoming Channels: Stay tuned for SMS, Slack, and more.
• Customization and Branding Tailor the appearance and behavior of your deployments to match your brand identity with customizable settings for colors, logos, greeting messages, and more.
• Session Triggers and Configuration Define how and when the AI companion initiates interactions, with options for session triggers based on user actions or events.

We hope you enjoy the new capabilities of our advanced workflow builder. Your feedback is invaluable to us as we continue to improve and expand our platform.

Happy building!

This feature allows you to securely pass user identifiers to all webhooks and function calls without exposing sensitive information to language models (LLMs). With the Metadata entity, you can ensure that proprietary and sensitive data, such as user IDs are handled securely and kept separate from the LLM interactions.

Stay tuned for more features and updates!

This feature enhances the accuracy and relevance of responses by integrating real-time data retrieval with AI-generated content.

• Improved Response Quality: By leveraging the knowledge base, your AI companions can provide more informed and context-specific answers, reducing the likelihood of errors.
• Dynamic Information Retrieval: Ensures that the AI companions can access and use the most up-to-date information available in the knowledge base during conversations.

Build more intelligent, responsive, and context-aware AI companions, capable of providing high-quality interactions tailored to your business needs!

Our platform now allows you to create a knowledge base directly within the AI Builder Tool. This feature enables you to store, manage, and retrieve information that your AI companions can use to provide accurate, context-specific responses.

• Centralized Information: Store all relevant data, documents, and FAQs in one place, making it easier for your AI companions to access and provide consistent, accurate information.
• Improved Response Accuracy: By referencing the knowledge base, your AI companions can deliver more precise answers, reducing the chances of errors or misinformation.
• Ease of Management: The knowledge base can be easily updated, ensuring that your AI companions always have access to the most current information.

Stay tuned for more features and updates!

Support for GPT-4o Mini

A compact yet powerful variant of GPT-4, this model offers advanced natural language understanding and generation capabilities, optimized for performance and efficiency.

• Compact Size: Designed to be smaller and more efficient while retaining the powerful capabilities of GPT-4, making it ideal for applications with limited computational resources.
• High Performance: Despite its compact size, GPT-4o Mini provides high-quality natural language processing, enabling sophisticated and accurate interactions.
• Efficiency: Optimized for performance, GPT-4o Mini delivers fast response times and efficient processing, suitable for high-demand applications.

Support for Llama 3.1 405b

We have also added support for Llama 3.1 405b, Meta’s latest large language model for voice modality. This model brings enhanced language processing capabilities and is designed for high-performance applications.

• Advanced Language Processing: Llama 3.1 405b excels in understanding and generating complex language constructs, making it suitable for a wide range of sophisticated tasks.
• High Performance: Built for demanding applications, this model offers robust performance and can handle large volumes of data with ease.
• Versatility: Ideal for applications requiring high-level reasoning, contextual understanding, and detailed content generation.

Previously, sharable links were available for voice interactions only. We have now added support for chat mode as well. This allows anyone to test your AI companion by accessing it through a sharable link. This feature is perfect for demos and for sharing with internal teams that are not part of the builder, providing a seamless way to showcase your AI companion’s capabilities in both voice and text formats.

• Integration Options: When configuring the iframe/bubble widget integration options, customers can now:
  • Choose voice only
  • Choose text chat only
  • Choose both voice and text chat (default toggle is on)
• Dynamic Script Updates: Based on the selected modality settings (toggle), the system will automatically update the iframe/bubble script.
• Sharable Link UI: The user interface of the sharable link will also update according to the selected modality settings, ensuring a consistent and tailored user experience.

This enhancement allows for a more tailored integration of AI companions, aligning with specific user needs and interaction preferences.

We also support interactions across different channels within the scope of a single conversation, enhancing the versatility of your AI companions.

This enhancement broadens user access and engagement, providing a seamless way to showcase your AI companion’s capabilities to a wider audience.

We are committed to enhancing the AI Builder Tool to meet the evolving needs of our customers. Stay tuned for more updated!

We now supports Custom Function Calls, allowing you to extend the functionality of your AI companions by integrating with external services and executing specific tasks. This feature enables more dynamic and context-aware interactions.

• Flexible Integration: Easily integrate with various APIs and external services to perform actions such as retrieving data, triggering workflows, or accessing third-party functionalities.
• Dynamic Responses: Enable your AI companions to provide more relevant and actionable responses by executing custom functions based on user inputs.
• Enhanced Interactivity: Improve user engagement by allowing AI companions to perform tasks like booking appointments, updating records, or providing real-time information.

Please check the comprehensive documentation and examples to help you get started with creating and deploying custom function calls in API docs.

Thank you for your continued support!

We now supports Custom Variables, allowing you to define and utilize dynamic variables within your AI companions. This feature enables more personalized and context-specific interactions.

• Dynamic Data Handling: Define test variables that can be used to test enhanced interactions that allow for more complex and tailored responses based on user input.
• Enhanced Personalization: Use custom variables to create more personalized user experiences by retaining user-specific information across interactions.
• Flexible Integration: Easily add custom variables into your existing workflows and prompts, enhancing the capabilities and responsiveness of your AI companions by wrapping the variable in curly brackets.

Please check our detailed documentation to help you implement and make the most of custom variables.

Conversation Logs

The platform now includes Conversation Logs, which allow you to view detailed records of all interactions between users and your AI companions. This feature provides valuable insights into user behavior, interaction patterns, and conversation quality.

• Detailed Interaction Records: Under each companion each log entry includes timestamps, user messages, AI responses, and metadata, giving you a comprehensive view of each conversation.
• Search and Filter: Easily search and filter conversation logs to find specific interactions or analyze trends over time.
• API Capabilities: Receive conversation logs for further analysis or reporting, ensuring you have the data you need to make informed decisions.

This new method combines the robustness of private key cryptography with the convenience and flexibility of access tokens, providing an extra layer of protection for our customers.

Implementation details

CK generates and provides each client with a key in a format of a random hash assigned to an Organization to be used from the backend only.

CK_SECRET_API_KEY = ck_secret_344487fb916678f6157871d53b5301482cfd9c7f;

This key should not be exposed on the front-end.

CK REST API will provide an endpoint for generating an access_token based on the CK_PRIVATE_API_KEY which then can be used on the front-end.

You can find more details in API documentation.

This new authentication method strengthens our commitment to providing secure and reliable services. We encourage all users to implement this method to take full advantage of the enhanced security it offers.

Key Capabilities of Sonnet 3.5

• Enhanced Performance: Sonnet 3.5 delivers superior accuracy in natural language tasks, outperforming previous models such as Sonnet 3.4. It excels in handling complex instructions, humor, and context-sensitive tasks.
• Increased Speed: Operates at twice the speed of previous models, ensuring faster response times and more efficient processing.

Stay tuned for more updates and new features!

Implementation details

We generates and provides each client with a key in a format of a random hash assigned to an Organization, e.g.

CK_PUBLIC_API_KEY = ck_344487fb916678f6157871d53b5301482cfd9c7f;

Customer’s app should provide CK_PUBLIC_API_KEY into the integrated widget as a param on the front-end.

Public key would have read-only permissions only, no access to management operations (ability to modify an assistant, etc.)

By default we consider this method for Organization-level auth, not user level. Please connect with OutRival’s team if you are interested in user-level authentication.

You can find more details in API documentation.

Memory Support for AI Companions

Our AI companions now have memory capabilities, allowing them to retain context and recall previous interactions. This feature enhances user experience by enabling more personalized and context-aware conversations.

Memory is enabled on a user level, for all organizations that opt in to use this add-on service.

We are committed to continuously enhancing the AI Builder Tool to meet the evolving needs of our customers. Stay tuned for more updates and new features.

• Open-Source Model Integration:With Groq, customers can now seamlessly integrate and leverage open-source LLM models. This flexibility allows for customized and highly specific use cases, tapping into the extensive capabilities of the open-source community.
• Enhanced Performance: Expect improved accuracy and efficiency in natural language processing tasks, leading to more precise and reliable outputs.
• Optimized Response Times: Enjoy quicker response times, thanks to the optimized performance of open source models.

Key Features:

1. No-Code Builder:
   • Intuitive Interface: Easily create AI companions using templates or start from scratch in just a few minutes.
   • Customizable Components: Tailor your AI companion with configurable prompts, welcome messages, LLM model usage, and function calling to suit specific use cases.
2. Multi-Modality Configuration:
   • Voice, SMS, Text: Choose from three different modalities to interact with users, providing flexibility in communication channels.
3. In-Built Testing:
   • Real-Time Testing: Test your AI companion directly within the builder, making it easy to iterate and refine based on performance feedback.
4. API Access
   • Comprehensive API: Access your AI companions on production through our API. Coupled with extensive API documentation we are positioned to support seamless integration for your teams.
   • We prioritize security in all aspects of our platform usage, ensuring that your data and interactions remain protected.
5. Seamless Integration Options:
   • Flexible Deployment: Integrate AI companions into web apps via widget, iframe, or subdomain, ensuring quick and efficient production rollout.
6. Demo Capabilities:
   • Showcase Potential: Bring the AI companion to your own customer demos, demonstrating its capabilities and potential to your prospects.

Upcoming Enhancements:

1. Knowledge Base Integration:
   • RAG Workflow: Implement a Retrieval-Augmented Generation (RAG) workflow allowing customers to bring their own knowledge base or create one within the builder. This will enhance AI companions by providing accurate, context-specific responses drawn from proprietary data.
2. Multi-Agent Flow Builder:
   • Sophisticated Workflows: Enable customers to create complex, multi-agent workflows for various tasks such as conducting research, automated email communication, data collection, and task execution. This feature supports intricate processes involving multiple AI agents working collaboratively to achieve specific goals.
3. Comprehensive Analytics:
   • Usage Analytics: Gain detailed insights into user interactions, companion performance, and overall engagement metrics to monitor and optimize use cases.
   • AI Evaluations: Benchmark and evaluate LLM model performance with metrics on accuracy, efficiency, and other key indicators to ensure high standards are met.

We are committed to continuously enhancing the AI Builder Tool to meet the evolving needs of our customers. Stay tuned for more updates and new features in the coming months.

Thank you for your continued support!