Sana Coach

Designing an AI-first model for Continuous Performance Management

Overview

At Workday, I helped define and prototype a new AI-first product concept called Sana Coach. The goal was to rethink performance management as something continuous and useful throughout the year, rather than a periodic administrative task centered around annual reviews.

Coach was designed to support managers, individual contributors, and people leaders across a range of connected activities, including feedback, goal tracking, career development, coaching, performance conversations, and employee growth.

My role sat at the intersection of product design, design engineering, systems thinking, and AI interaction design. I was responsible for shaping the experience, defining interaction models, and building working React prototypes that allowed teams to experience the concept as software rather than static design.

The challenge

Performance management is often fragmented across multiple tools, workflows, and moments in time. Managers may need to gather feedback, track goals, remember previous conversations, prepare performance reviews, coach employees, identify development opportunities, and navigate career conversations—often across different systems and with limited continuity between them.

The opportunity was not simply to make one of those workflows faster.

The larger question was:

That required thinking beyond traditional page-based UX.

We needed to explore how an AI-driven experience could:

• understand context across multiple talent workflows

• surface relevant information at the right time

• help users prepare for important conversations

• support ongoing feedback and coaching

• connect goals, performance, skills, and career development

• reduce repetitive administrative work

• maintain user trust and control

  
  

My role

I served as the senior UX design engineer on the effort, with responsibilities spanning both product design and technical prototyping.

My work included:

• helping define the product experience for Sana Coach

• translating broad AI-first concepts into concrete interaction models

• designing experiences across Performance Management, 360 Feedback, Learning, coaching, and career development

• building high-fidelity React prototypes

• using Workday Canvas and Sana design-system components

• collaborating with product, engineering, and design leadership

• helping teams evaluate agentic workflows through working software

• identifying reusable interaction patterns that could extend across multiple product areas

Because the product space was still emerging, a large part of the job involved turning ambiguity into something teams could react to.

Reframing performance management

One of the core design shifts was moving from episodic performance management to continuous performance management.

Traditional systems often organize the experience around events:

• Annual review

• Goal-setting period

• Feedback request

• Performance conversation

Coach allowed us to explore a different model—one in which the system could maintain context throughout the year.

Instead of treating each interaction as a separate transaction, the experience could potentially understand:

• current goals

• previous feedback

• upcoming milestones

• development priorities

• skills

• career interests

• prior conversations

• team dynamics

• organizational expectations

That context could then inform the assistance provided to the user. The design challenge became less about creating isolated screens and more about designing an ongoing relationship between the user and the product.

Designing for three different users

Coach needed to support users with very different needs.

Managers

Managers needed help preparing for conversations, providing meaningful feedback, supporting employee development, tracking goals, and completing performance-related work. The design needed to reduce administrative effort without removing manager judgment.

Individual contributors

Employees needed a way to understand their progress, receive useful coaching, reflect on feedback, identify growth opportunities, and prepare for career conversations. The experience needed to feel supportive rather than evaluative or intrusive.

People leaders

More senior leaders needed broader context around teams, talent, development, and performance while still respecting the boundaries between organizational insight and individual employee experiences. Designing across these perspectives helped us avoid treating Coach as a single-purpose chatbot.

It needed to operate as part of a larger talent ecosystem.

From assistant to agent

Another important design question was determining how proactive the product should be. A traditional assistant waits for the user to ask a question. An agentic system may recognize that something needs attention and offer help before the user explicitly requests it.

For example, Coach might recognize that:

• a performance conversation is approaching

• an employee has not received feedback recently

• a goal has stalled

• development activity connects to a career aspiration

• previous feedback may be relevant to an upcoming conversation

This introduced a new set of UX questions:

• When should the system intervene?

• When should it stay quiet?

• How much context should it expose?

• How should users verify or correct what the AI understands?

• When should AI recommend versus act?

• How do we preserve human judgment in sensitive talent decisions?

These questions became part of the product model itself.

Prototyping as a design tool

Static mockups were not enough to explore these interactions.

Many of the most important questions only became visible when the experience was interactive:

• How does the conversation evolve?

• How does context persist?

• What happens when the user changes direction?

• How does the product surface supporting information?

• How do traditional UI elements work alongside conversational interaction?

• How does the experience feel over time?

  
  

I built high-fidelity prototypes in React, using existing design-system components and shared code.

This allowed us to create product experiences that behaved much more like real software.

The prototypes became a shared artifact for design, product, engineering, and leadership.

Instead of debating what a concept might feel like, teams could use it.

Working at prototype speed

This work also built on a broader design-engineering practice I had been developing at Workday.

By using React, GitHub, shared component libraries, and existing design systems, we were able to create sophisticated prototypes much faster than traditional prototype-development workflows.

Experiences that might previously have required weeks of handoff and development could often be explored in hours or days.

That speed mattered particularly for AI product work, where ideas evolved rapidly and teams needed to test multiple interaction models before committing to a direction.

The prototype became less of a final presentation artifact and more of an active design environment.

Designing a connected experience

One of the most important principles behind Coach was that performance management should not exist in isolation.

Performance conversations are connected to:

• Goals

• Feedback

• Skills

• Learning

• Career development

• Coaching

• Talent Mobility

Because I had previously worked extensively on Workday’s skills-based products, I was able to approach Coach with a broader systems perspective. The challenge was to design experiences that felt simple to the user while drawing from a much more complex underlying talent ecosystem.

That meant focusing heavily on context, continuity, and progressive disclosure. The user should not need to understand the structure of the system in order to benefit from it.

Collaboration

The work required close collaboration across product, design, engineering, and leadership. My role often involved translating between disciplines. Product teams needed to understand what was possible. Designers needed to understand how AI behavior affected interaction models. Engineers needed a clearer picture of the intended experience. Leadership needed something concrete enough to evaluate strategically.

Working prototypes helped bridge those conversations. They created a common reference point that allowed teams to make decisions faster and with greater confidence.

Outcome

The work contributed to Workday’s broader vision for continuous performance management powered by AI.

The concept was positioned around helping organizations improve manager efficiency while supporting ongoing feedback, employee growth, and talent retention.

The most important outcome from a design perspective was the creation of a coherent model for how an AI coach could connect traditionally fragmented talent experiences into an ongoing relationship with the user.

For me, the project also reinforced a broader principle:

What I learned

Sana Coach changed how I think about product design in several ways.

AI products require behavioral design

The interface is only part of the experience. Designers also need to define when a system acts, what context it uses, how it explains itself, and where human judgment remains essential.

Prototypes are increasingly strategic

As product behavior becomes more dynamic, working prototypes become one of the fastest ways to align teams around an idea.

Context is a product capability

The value of an intelligent system increases when it can connect information across time and across workflows.

Designing that continuity is as important as designing the individual interactions.

Simplicity often sits on top of complexity

The best experience may look conversational and simple while depending on a sophisticated underlying system.

My job is often to absorb that complexity and expose only what the user actually needs.

My contribution

Product Design

AI-first interaction models, continuous performance management, coaching experiences, connected talent workflows

Design Engineering

React prototyping, design-system implementation, GitHub collaboration, reusable components

Systems Thinking

Performance Management, 360 Feedback, Learning, Skills, career development

Leadership

Cross-functional alignment, prototyping strategy, mentoring, design-engineering practice