Skip to content
Behind the Release Notes: Rasa Pro 3.6

June 30th, 2023

Behind the Release Notes: Rasa Pro 3.6

  • portrait of Lauren Goerz

    Lauren Goerz

In Rasa Pro 3.6, we introduce new capabilities to help conversational AI teams better secure and track the performance of Rasa AI Assistants with PII Data Management and Real-Time Markers. This release also includes Load Testing Guidelines to provide resource recommendations for Rasa Assistants operating at scale, and we have included an alpha release of Spaces. Spaces is a capability that allows for the modularization of Rasa assistants so that conversational AI teams can work on different parts of an AI Assistant at the same time.

PII Data Management

We know that data privacy is essential. When customers place their trust in your brand, they expect that their data and personal information is handled securely. In Rasa Pro 3.6, we introduce a comprehensive solution that will allow enterprise AI teams to safely process PII (Personal Identifiable Information) such as credit card numbers, addresses, and social security numbers when necessary. This approach gives enterprise teams the ability to confidently anonymize PII data in logs and events streamed via the Kafka event broker.

With our PII solution you will be able to do the following:

  • One-way PII anonymization powered by Presidio, an extensible, customizable approach to PII handling that leverages natural language entity extraction to efficiently enable consistent PII recognition.
  • Anonymization support for 100+ languages. A language is considered supported if there is a trained-pipeline from any of the three recommended model providers: spaCy, Stanza, or Hugging Face. Keep in mind, PII data can be anonymized in only one language per assistant.
  • With our solution you can pseudonymize, mask, or substitute PII data:

To learn more about how to apply our PII Data Management solution in your product, check out our code along video with Rasa Software Engineer Tawakalt Olaniyi.

Real-Time Markers

Real-Time Markers gives conversational AI teams and analysts the ability to insert metadata tags within Rasa Assistants to support the targeted analysis of key moments in the user journey for augmented Conversation-Driven Development. When implemented, the metadata collected by Real-Time Markers can help your team answer these types of questions:

  • Why did our solution rate drop last week?
  • How many people used our new user journey since it was released?
  • Show me all conversations where users drop out of our appointment booking flow.

This can be done by inserting individual markers to capture specific moments in the user journey, or you can also string together a pattern of sequential markers to track a full user journey with your AI Assistant.

These markers are also real-time, which means that the captured metadata is immediately processed at the end of each user session so there is a continuous flow of information giving AI Analysts the most up to date information. To learn more about Real-Time Markers, check out our Docs.

Load Testing Guidelines

In Rasa Pro 3.6, we introduced load testing guidelines for enterprise AI Assistants at scale to help better estimate the resources required for operation in production. We have also provided some guidance detailing how to run load-tests on your Rasa Assistants.

Spaces (Alpha Release)

We would also like to highlight the alpha release of the Spaces feature in Rasa Pro. With Spaces, you can modularize your AI assistant by creating different spaces that can be built and tested separately, and then integrated into one single assistant. This alpha release increases isolation between subparts and improves classification performance for intents and entities that are only relevant in specific contexts.

To try out Spaces, check the documentation, and reach out to your Customer Success representative if you would like to learn more.

Additional Enhancements

  • Slot Mapping: Slot mapping conditions now accept active_loop: null in those cases when slots with this mapping condition should be filled only outside form contexts.
  • Python 3.7 Support: We have also Removed Python 3.7 support as it reaches its end of life in June 2023.
  • ARM Support: Added capability to deploy on arm64 processor architecture (e.g. to allow running Rasa Pro inside Docker on M1/M2 chipsets) available in Rasa 3.6.2 or later.
  • Check out the rest of the enhancements here.

Next Steps

These new capabilities are all a part of Rasa Pro, our enterprise infrastructure suite that can help you scale, test, secure, and monitor your AI Assistants. Rasa Pro enables a pro-code approach to the build of AI Assistants, and allows you to automate critical components of your conversational AI Assistant’s lifecycle.

Want to learn more about Rasa Pro and how you can get access to these new capabilities? Connect with an expert, or try it out for free.