Overview

This strategic project focused on modernizing how UXD teams work through agentic AI, workflow orchestration, and emerging technology. As product and engineering environments became more complex, teams needed a faster and more connected way to synthesize inputs, structure work, accelerate execution, and track progress.

The opportunity was to create a more intelligent operating model for the design organization, one that helped designers make insight-informed decisions, document and communicate solutions more clearly to product and engineering partners, and maintain visibility across the work while keeping human judgment, quality, accessibility, and accountability at the center.

A key part of the solution included specialized AI subagents and custom commands to codify repeated, complex workflows. These subagents could be chained to handle sequential tasks, such as moving from product specification to architecture to implementation or testing. This modular approach improved reproducibility, supported clearer separation of concerns, and significantly increased throughput.

The work evolved into a new design operations model supported by a 30/60/90-day adoption framework. Within six months, 80% of team members were actively using AI-enabled prototypes and agentic workflows, operational efficiency and AI engagement increased by 30%, and teams reduced time across ticket creation, synthesis, design acceleration, and concept-to-prototype cycles.

Challenge

At Red Hat, UXD teams had valuable insights, experiments, and ways of working emerging across the organization, but they were scattered, inconsistent, and difficult to scale.

As delivery demands increased, these fragmented workflows created operational drag. Teams were spending too much time manually organizing work, synthesizing research insights, and translating conversations into 3IB actions instead of focusing on higher-value design decisions.

Key friction points included:

• PRDs, Jira epic and ticket creation required repeated manual setup
• Daily synthesis across Research, Slack, email, calendar activity, and related signals took hours
• Fragmented cross-functional conversations were slow to convert into actionable workstreams
• Concept development involved unnecessary repetition before reaching usable fidelity
• Transitions between design exploration and prototype delivery lacked speed and continuity

My role

I was part of an experimental tiger team exploring how agentic AI could support real UXD workflows across Red Hat. I helped move the work beyond experimentation by shaping a practical orchestration model that design, engineering, product, and quality partners could understand and adopt.

My role included testing AI-enabled workflows, identifying where AI could reduce operational friction, and evangelizing the approach across teams responsible for quality, execution, and upstream product delivery.

The value was not automation alone. It was creating a human-centered model that helped teams interpret inputs, translate ambiguity into action, and move work forward with greater consistency, clarity, and confidence.

Strategy

We approached this work as both an operational design problem and a cultural opportunity. as this work was experimental, a major part of the value came from testing, learning, and sharing what worked. I used the tiger team environment not just to build solutions, but to demonstrate new possibilities for design and product teams.

PROCESS

DISCOVERY

Through survey feedback, internal user interviews, and ongoing communication with the design and engineering organization, we learned that 80% of internal teams were struggling to adopt AI in a consistent, confident, and scalable way.

The challenge was not a lack of interest. Designers were already experimenting with AI, but many were still skeptical about how deeply AI should be involved in the design process. The organization needed more structure, guidance, and confidence around how to use AI effectively without compromising quality, trust, or design standards.

Key themes emerged:

1. Adoption was moving fast, but unevenly
   Designers were exploring AI tools at different speeds, creating inconsistency in methods, outputs, and confidence.
2. Quality needed to be protected
   Teams wanted to move faster, but not at the cost of design judgment, accessibility, usability, or product quality.
3. Stakeholder communication was becoming more important
   Designers needed better ways to explain AI-assisted work, socialize insights, and bring product and engineering partners along.
4. Ways of working needed more transparency
   Teams needed centralized sources, shared guidance, reusable practices, and clearer documentation so AI-enabled workflows could scale across the organization.

Define

We defined the core problem, identified the foundational steps required for responsible AI adoption, and shaped a human-in-the-lead, AI-in-action framework to help teams apply AI with clarity, accountability, and confidence across design workflows.

The framework turns fragmented requests into a structured operating model where AI helps collect signals, surface patterns, suggest experience directions, and accelerate handoffs, while humans remain responsible for verification, judgment, decision-making, and final quality.

Core components included:

• MCP servers for tool coordination and workflow context
• Cursor and Ask for synthesis, planning, and task generation
• Jira for epic and ticket creation
• Slack, email, calendar events, and relevant news for daily summaries and priority signals
• Figma plugins for faster low- to high-fidelity wireframing and design system outputs
• GitLab for faster prototyping and tighter design-engineering collaboration
• Momentum app for tracking and reviewing outcomes

ADOPTION

This phase was treated as a structured enablement effort rather than an informal rollout.

Step-by-step integrations, MCP server access, dedicated Slack channels, and centralized Confluence documentation gave teams clear ways to learn, ask questions, share experiments, and refine their workflows. This created stronger visibility into how the system worked and lowered the barrier to experimentation across the team.

It also helped shift behavior from waiting for direction to taking action, learning in public, and improving through shared practice. That cultural shift was as important as the tooling itself, because it encouraged teams to engage with AI as an active part of their workflow rather than as a separate experiment.

Within six months, 80% of team members were actively operating with AI-enabled prototypes and agentic workflows.

Within design teams, each designer had a personalized interface tailored to their workflow, helping them track progress, communicate faster, and centralize platforms for all communications and initiatives.

view agentic ai outcomes 

This system enabled UXD teams to:

• Automated Jira epics and tickets from meeting inputs and discussions
• Summarized daily communication and surfaced what mattered most
• Turned fragmented conversations into structured action
• Accelerated wireframe creation and reusable design system patterns
• Shortened the path from design concept to prototype

More broadly, it showed how agentic AI could support both team operations and design execution when applied with clear intent and workflow structure.

Impact

The initiative delivered measurable improvements in both efficiency and execution speed:

• 30% increase in operational efficiency through AI-driven workflow orchestration
• Jira epic and ticket setup reduced from 1–2 hours to 10–15 minutes
• daily synthesis across Slack, email, calendar activity, and news reduced from hours to minutes
• low-fidelity to higher-fidelity design workflows accelerated from days to hours
• concept-to-prototype cycles shortened from weeks to days through tighter Figma and GitLab integration

Beyond the measurable gains, the work also created momentum around experimentation and practical adoption. It showed how agentic AI could be applied in a repeatable and team-centered way to improve operations, strengthen execution, and support broader organizational learning.

A meaningful outcome was how the team approached transparency. Several designers began introducing “co-authored by AI” disclosures in Jira tickets and generated content to make AI involvement visible to stakeholders. Instead of positioning AI as an author, the team standardized on “AI-assisted” attribution, paired with a brief explanation of how AI contributed (e.g., synthesis, structuring, or generation). This aligned with industry guidance that maintains human ownership while ensuring transparency and accountability.