Chatbot: Q3 Product Design Requests

[kaylachumley]

Problem statement

As PF expands its adoption of AI-powered chat experiences across products and platforms, the AI Chatbot Working Group is increasingly receiving product requests from multiple teams. High-priority products like RHOAI have started to approach the ChatBot extension with new requests for features and design guidelines. We need to be able to accommodate requests for new features in order to keep the Chatbot as an effective source of shared design and code.

Outcome

By the end of Q3 2025, we will have completed and passed off all approved designs requests to development and complete ready-for-development issues that are requirements for RHOAI, the AI Installer project, or other high-priority products.

• Designs reviewed
• All issues will be closed, with designs updated in the Figma PF6 Patterns and Extensions library.
• Designs handed off to development to begin implementing. (usually Rebecca Alpert)

Impact

The AI Chatbot Working Group's ability to efficiently integrate new features and design guidelines directly impacts the scalability and effectiveness of AI-powered chat experiences across all Red Hat products and platforms.

Execution details

This epic's success hinges on a streamlined and collaborative approach to feature integration. By adhering to these execution details, we aim to not only accommodate new feature requests efficiently but also to strengthen the Chatbot's role as a source of truth for AI-powered experiences across UXD.

Process

• Weekly sync review: During the established weekly Chatbot syncs, the AI Chatbot Working Group will review all incoming requests. Each request will be assessed for its alignment with strategic objectives, potential impact on user experience, technical feasibility, and resource requirements.

• High-Value/High Impact - Action item designation: Requests deemed high-value will be designated as such and will receive prioritized discussion and approval from the team. This designation will be based on factors such as product criticality, user impact, and alignment with PatternFly components.

• Request submissions: All features and design guideline requests will be formally tracked and submitted through github and Jira. This ensures a consistent intake process and captures all necessary details from the requesting teams.

• Cross-Functional collaboration: For complex features or those impacting multiple products, we'll establish direct lines of communication and collaboration with the requesting product teams (e.g., Lightspeed, RHOAI, Docs.rh). This includes regular check-ins and opportunities for feedback throughout the development lifecycle.

• General status updates: The status of all incoming and in-progress requests will be regularly communicated to stakeholders through Jira and during the weekly syncs.

• Wireframing and iterating: Design will develop wireframes or mockups to visualize the proposed features and design guidelines.

• Design Review: Designs will be informally reviewed by Andrew Ronaldson or discussed with the larger team to ensure we do not have a breaking change.

• Figma design kit addition: Post design review, the final pattern/component will be added into the figma kit before development begins. The design request will be marked as complete and passed off to Rebecca to implement. The working figma branch should be ready to merge once development is complete.

Challenges

Challenge 1: Ensuring effective cross team collaboration and communication with Lightspeed and the product area involved with the request. This could be potentially challenging as we want to avoid duplicate feature work , misaligned expectations and slow feedback loops.

• Solution 1: The Chatbot Working Group should clearly communicate with product liaisons and proactively provide regular status updates on in-flight features, rather than waiting for requesting teams to inquire.

Challenge 2: As the demand for AI-powered chat experiences grows, the Chatbot working group may face challenges in having sufficient resources (people, tools and contributions. Design can quickly become overrun with incoming requests

• Solution 2: We should try to regularly forecast incoming demand for features based on product roadmaps, strategic initiatives and evaluate other AI competitors.

Lessons

Clearly defining the problem statement, impact, and a robust intake/prioritization process upfront is crucial for managing expectations and resource allocation across sprints.
Anticipating communication challenges, planning with designated liaisons, and providing proactive updates is critical for successful collaboration between the chatbot working group and various product teams.

Next Steps:

• Development: Rebecca(or Nicole) is responsible for coding the feature based on the approved designs and technical specifications.
• Documentation: Each new feature and design guideline will be documented on patternfly.org by Erin. The documentation should include specifications, demos, usage guidelines, and any other relevant examples

Resources

Links:

• Development Github epic
• Development Jira Issue
• Design Github Epic
• Footnotes
• Response Actions
• Deep think sections

Notes:

Will need design guidelines from Erin Donehoo for the footnotes usage/application and response actions hover/clicked state expectations(filled and unfilled icons)

[rebeccaalpert]

Kayla and I talked on Slack and at a minimum it sounds like we want to include these:

• Demo for sources drawer with footnotes: Add demo for how to use footnotes to link to sources chatbot#573
• "Thinking" or "reasoning" messages: Support expandable <think> sections for reasoning models chatbot#467