Assistant
Guide on using the Ketryx Assistant to accelerate safety-critical work
Last updated
Was this helpful?
Guide on using the Ketryx Assistant to accelerate safety-critical work
Last updated
Was this helpful?
Ketryx is designed to help users interact with the platform more efficiently by providing natural language understanding of queries and automating common tasks. Powered by Large Language Models (LLMs), Ketryx can understand context, interpret complex questions, and provide relevant answers. As with any AI system, while these features are designed to be helpful and accurate, users should always carefully review and verify all suggestions, results, and generated content before taking action. Whether you're searching for items, analyzing requirements, or creating new configuration items, Ketryx can help streamline your workflow while you maintain full control and oversight of the process.
Ketryx prioritizes the security and privacy of your data. When interacting with AI features, your project data is processed using a zero data retention LLM endpoint. This means your data exists temporarily in memory only to serve your immediate request and is not persisted, stored, or used for training purposes.
Throughout the platform, the magic wand icon (🪄) indicates the availability of assistant support.
The Assistant provides different response modes to match your workflow needs:
Fast Mode: Delivers shorter responses optimized for speed when you need quick answers.
Standard Mode: Provides more detailed responses with deeper analysis for comprehensive understanding.
Extended Mode (coming soon): Offers the most comprehensive responses with in-depth reasoning for complex queries.
Ketryx can help create new items and suggest improvements to existing ones. Rather than spending manual time creating various configuration items, slowing time to market, Ketryx will generate drafts that you can review and edit before automatically syncing with systems like Jira.
Tip: For optimal performance, consider that requests for multiple items (e.g., "create five test cases") will take longer to process than requests for single items – the response time increases with the number of items requested.
Examples:
"Please create a new requirement for bluetooth connectivity"
"I want to create a software spec for Android integration"
"Identify risks for requirement KX-371 and create two risk controls for them"
"Create a risk that is aligned with ISO 14971 for a web application"
"Create a detailed test case for KP-2040"
Ketryx can translate natural language queries into Ketryx Query Language (KQL) to help you find and filter items based on specific criteria. Using AI-based natural language understanding, it converts your natural questions into precise database queries.
Examples:
"Show me all requirements that are not yet approved"
"Find software items that fulfill requirements"
"Show me test cases in the test plan for version 1.0"
"Get items approved by R&D Leads"
"Find requirements containing 'user interface' in the description"
Ketryx uses Retrieval Augmented Generation (RAG) technology to analyze your project's content and find relevant matches, even when exact criteria aren't known. This means it can understand context and find items based on concepts and meaning rather than just matching keywords or fields.
Examples:
"Do we have requirements that talk about bluetooth?"
"Show me all specs relating to cybersecurity"
"Someone at the meeting mentioned depth for water resistance. Can you find the relevant requirement?"
"Is there an existing anomaly for this case where a title is one long word without spaces?"
"I would like to add a requirement that dictates that the SaMD has to connect to a glucose monitor sensor and have continuous synchronization. Are there any requirement configuration items that relate to that idea that I could adapt instead of authoring a new requirement?"
In complex systems with hundreds or thousands of interconnected items, it can be challenging to keep track of all relationships and potential conflicts. Ketryx helps you navigate this complexity by performing sophisticated analysis tasks:
Analyze relationships between items
Identify potential conflicts
Summarize groups of related items
Evaluate the impact of potential changes
Identify risks
Examples:
"Can you look at requirements that have to do with the mobile app and summarize them for me?"
"Are these two requirements consistent with one another?"
"What are the two most similar requirements in the project?"
"Which requirements might be fulfilled by a new spec for Bluetooth connectivity?"
"Today at the meeting someone suggested adding a requirement to disallow Bluetooth connection. Will this conflict with any existing requirements?"
Ketryx maintains context throughout the conversation, allowing for natural follow-up queries (e.g., "Create tests for those" or "Would any of those be fulfilled by a new spec for bluetooth connection?"). Using its memory of previous interactions and referenced items, it can provide more relevant and specific responses to follow-up queries.
Ketryx leverages RAG technology to provide accurate answers about the platform, drawing from the official documentation. This means it can reference specific documents and support articles while providing contextually relevant answers.
Examples:
"How do you add your repository to Ketryx for a new project?"
"How do I create a Personal Access Token for GitLab?"
"What languages does Ketryx built-in dependency scanning support?"
"How can I use 'Release ref pattern' for connected repositories in the Ketryx project settings?"
"What is the difference between the release vulnerability report and the 'on-demand' vulnerability report data grid option on the Vulnerabilities page?"
"How do I get automated tests to show up in Ketryx?"
Ketryx can help suggest traceability relationships between items by using the 🪄 Suggest button next to relation fields. This feature can help with:
Identifying affected items for Change Requests
Finding parent requirements for new Requirements
Suggesting related test cases
Identifying potential risks and controls
Ketryx can help analyze existing documents and provide guidance for creating templates that match your documentation needs.
Ketryx helps with:
Analyzing document structure and suggesting appropriate template organization
Providing relevant KQL queries for data access
Recommending template sections and their mapping to Ketryx items
Offering step-by-step guidance for template implementation
Examples:
"How can I create a template that shows requirements grouped by priority?"
"Help me structure a template based on this existing specification document"
"What's the best way to show test coverage in my requirements template?"
"Can you suggest a template structure for my risk management document?"
"How do I create a template that includes both requirements and their linked test cases?"
If you would like to set default prompts for your organization or project, use the .
Ketryx is continuously improving. If you have any feedback or encounter any issues, please use the thumbs up/down buttons in the chat or and let us know!
