Skip to main content

Release 5.0 2026.07.28.1

· 3 min read

New Features 🚀​

Media Monitoring App​

The Media Monitoring App is a new addition to the Serenity* Apps catalog. An agent carries out daily media surveillance on your behalf and presents the results in a detailed dashboard covering relevant news, media updates and more.

With a purpose-built interface, competitive analysis and media trend exploration are easier than ever, with no prompt engineering required.

alt text

OCR Service​

A new dedicated AI Service brings AI-powered OCR to Serenity* Star. Extract text, tables, formulas and detailed information from image files, and feed the results straight into your AI Workflows.

alt text

Onboarding​

Setting up your organisation in Serenity* Star is now much simpler. Configure your company information, AI guidelines and more from a single guided flow, so your team starts from a well-configured workspace and gets more out of the platform's AI tools from day one.

alt text

alt text

Enhancements ✨​

Agent Design Studio Performance Improvement​

The Skills section of the Agent Design Studio has been completely redesigned for optimal performance. No more sluggish UI when working with large numbers of skills and data.

alt text

Anthropic via Azure​

Anthropic models can now be used through Azure AI Foundry, either as your main model, as a fallback, or with your own API key.

alt text

Chat Widget​

The chat widget's floating action button can now be customised with your own image or GIF, so the agent on your website matches your branding.

alt text

alt text​

Bug Fixes & Minor Improvements 🛠️​

  • Improved the JSON extractor used on agent responses.
  • Improved the JSON extractor's handling of Markdown links inside JSON content responses.
  • Fixed an issue with Knowledge on Demand skills when the agent did not complete the relevant tags.
  • Fixed access to the code snippets in the API Integration section for Activity and Chat agents.
  • Fixed an encoding issue in Agentic RAG when the agent generated accented characters.
  • Improved the preview feedback shown when a very large volatile knowledge file is uploaded and consumes the entire context window.
  • Unsupported parameter warnings are now cleared when the model of an agent is changed.
  • Improved the mapping of Anthropic vendor errors.
  • Fixed a warning notification in the Copy to Clipboard in the agent instance details.
  • Improved dataset schema validation when replacing and appending tables.
  • Fixed an issue that prevented conversation starters from being used in the Agent Design Studio after changing the model.
  • An email is now sent to the tenant whenever a renewal attempt fails.
  • Fixed an issue where a green tick was shown in the agent's View More even when a skill had failed.
  • UI improvements to the API key components for obfuscation and display.
  • Improvements to the extra body implementation on the OpenAI-compatible endpoints.
  • Fixed an issue when duplicating agents that had the AI Chat channel enabled.
  • Fixed an issue in proxy agents that took max_completion_tokens as 0 when it was not explicitly sent in the request.