Documentation Index

Fetch the complete documentation index at: https://docs.knovvu.com/llms.txt

Use this file to discover all available pages before exploring further.

Version 5.3

  • Published on Jul 2, 2026
  • 3 minute(s) read
  • Suara Atay
 Prev Next

Version 5.3

2 July, 2026

AI Agents

  • Data Tables:
    A new "Data Tables" module has been introduced, enabling structured data such as customer records, order details, or FAQ entries to be created, queried, and managed directly within the platform. Tables can be defined from scratch with custom columns, while system columns (Id, DateCreated, DateModified) are managed automatically. The feature is controlled via a tenant-level flag with configurable storage and row limits per tenant, along with a live storage usage widget on the Data Tables screen.

  • Data Table Node:
    Data Table node has been added to the Flow Designer palette under the Service Request category, allowing flows to query records from Data Tables based on configurable matching conditions with AND/OR logic. Query results are written to a session variable for use in subsequent nodes such as Say, Ask, or Condition, with dedicated found and failed output ports to handle successful queries and error scenarios.

  • Data Table String Column Type Improvements:
    The String column type in Data Tables has been split into "Text" and "Long Text" to better balance flexibility with performance. Text supports up to 200 characters and can be indexed for faster equality and prefix lookups, while Long Text allows unlimited length but cannot be indexed. Both types are fully supported by the Data Table node's existing string operators.

  • AI Testing with Historical Data:
    The "Test with Historical Data" option is now active when creating a scenario, allowing projects to be evaluated using real production conversations instead of simulated ones. Users define a test name, select a dataset, and configure at least one success criterion before running the test via Save & Run, with results routed to the Evaluations tab. Each evaluation run pulls a fresh random sample from the selected dataset, and test cards are now tagged and filterable at AI Testing page to distinguish test types at a glance.

  • Datasets for AI Testing:
    Users can now define reusable samples of production conversations for evaluation purposes through the new "Datasets" page under AI Testing. Each dataset specifies the number of conversations to pull, the desired live agent handover ratio, a fixed or relative date range, channel, and an optional minimum session duration. Upon saving, a confirmation popup displays the requested vs. found conversation counts along with the live agent handover breakdown, giving users visibility into the actual sample composition.

  • Change LLM Configurations in Bulk:
    A new "Model Assignments" tab has been added to Project Settings, grouping all AI Agents, RAG Agent nodes, Ask LLM nodes and Virtual Agent nodes in a project by the LLM model they currently use. Users can update a model everywhere it's used at once, rename components inline, and apply changes across the project via the "Save & Prepare Agents" button, removing the need to update each agent individually.

  • Setting Current Agent in Tool Flows:
    A "Set the next agent manually" toggle has been added to the End with Result node, allowing scenario designers to route directly to a known agent after a tool call instead of relying on the supervisor's LLM-based routing. Once enabled, the target agent can be selected from a Next Agent dropdown listing the relevant supervisor and its sub-agents, making routing faster, deterministic, and free of an extra LLM call.

Improvements

AI Agents

  • Exporting Conversations List with Rating Comment:
    The Conversations List page and its export now include a "Rating Comment" column, making it possible to review and extract customer feedback comments alongside historical conversation data.

  • Ext API for AI Insights:
    An external API endpoint is now available to retrieve AI Insight data for AI Agent conversations, returning summary, topic, average sentiment, churn risk, non-resolution risk, conversation results, actions, and improvement areas for a given external conversation ID. The endpoint is secured via the dataservice_external_session_metrics scope.

  • Minor UI Improvements:
    Several minor improvements have been made to the user interface to
    enhance user experience and usability.

  • Minor Bug Fixes:
    Various minor bugs have been addressed to improve overall system
    stability and performance.