Valence

Overview

Nadia is an AI leadership coach used by managers and organizations to prepare for performance reviews, difficult conversations, and ongoing team development. I was brought in on a trial basis as Head of Design to reimagine how the product could evolve from a conversational chatbot into an organizational intelligence platform—what we called the "org nervous system."

The Profile Insights Dashboard became the first concrete artifact of this vision: a working Vercel prototype that demonstrates how managers could access relationship intelligence, team dynamics, and critical context in one unified workspace.

Problem

Nadia had strong coaching content but weak infrastructure for organizing people, relationships, and context, which made it hard for leaders to see patterns across their teams. Through user and stakeholder conversations, three core problems emerged:

Cognitive Load Mismatch The same UI tried to serve calm planning moments and high-stress "meeting in 5 minutes" scenarios, overwhelming managers when they most needed simple guidance.

Context Blindness Nadia didn't track relationships, history, or org dynamics in a reusable way, so every coaching interaction started cold compared to tools like Copilot that sit on top of documents and comms.

Fragmented Experience Web app, Teams app, and email flows were disconnected, with no single, opinionated view of people and teams. The existing design system, Teams integration, and coaching UI all felt like separate products rather than parts of a cohesive platform.

Goals

The Profile Insights Dashboard was conceived as a small, opinionated slice of a larger "Nadia Intelligence Platform" that could:

• Prove the value of a context-rich org nervous system before overhauling the full product
• Define Nadia's differentiation as "the org's nervous system—sensing stress, distributing context, and strengthening connections"
• Create a unified view of people, relationships, and "moments that matter"
• Establish foundational frameworks (Heart Rate, Context Blocks, The Trinity) that could guide future development
• Give leadership a concrete prototype to test assumptions and guide roadmap prioritization

Process

Vision: Nadia as an Org Nervous System The strategy work defined Nadia's differentiation in contrast to BetterUp (live human coaching at high price, low organizational memory), Copilot (great at documents but not relationships), and generic LLM tools (powerful but context-free and forgetful).

Three pillars guided the redesign:

1. Heart Rate-Based Design: Match interface complexity to emotional state (crisis vs daily work vs long-term growth)
2. Context Blocks: Atomic, reusable units of professional context (relationships, feedback, history) that can surface anywhere in Nadia
3. The Trinity: Profiles (people), Journeys (timelines and rituals), and The Stack (systems and tooling layer) as the core data and interaction model

Core User Flow The app's primary flow mirrors how managers think in practice:

1. Scan the landscape: Land on the Profile Insights dashboard to see a unified view of key people, relationships, and profile/completion signals
2. Drill into context: Open a specific person or team to inspect their profile, relationship intelligence, and key context blocks, identifying gaps or moments that need attention
3. Decide and act: Use those insights to decide what to do next—prep a conversation, trigger a journey, or adjust a team plan—then return to the dashboard to monitor changes

Building the Prototype With no production engineering support, I built a realistic prototype in Vercel using v0-style AI-assisted tooling while staying close to Nadia's actual data model and constraints. The prototype had to:

• Work as a credible demo for customers, investors, and internal stakeholders
• Codify interaction models and AI behaviors that engineering could reference
• Prove the intelligence backbone concept before committing to a full rebuild

Solution

Heart Rate Navigation Using the Heart Rate framework, navigation and layout were rethought so managers could find what they need based on emotional state:

• NOW (High HR): Quick prep, crisis coaching, review assistants where cognitive load must be minimal
• TODAY (Medium HR): My Team, Colleagues, Calendar prep, Context Library—spaces for daily relationship work
• DEVELOP (Low HR): Goals, learning paths, and team analytics for deeper reflection

The Profile Insights workspace sits primarily in the TODAY/DEVELOP band but is reachable from high-heart-rate alerts.

Context Blocks and Colleague Cards To reduce duplication and fragmentation, the dashboard is built on Context Blocks and colleague cards:

• Context Blocks encapsulate feedback, history, survey results, and milestones, and can surface on profiles, in meeting prep, and in team views
• Colleague cards act as compact, expandable previews that can be invoked from chat, email, or the dashboard

This allowed the Profile Insights dashboard to act as a "home" for these primitives rather than a bespoke one-off view.

States, Privacy, and Edge Cases Because leadership data is sensitive, robust privacy and state handling were designed even at prototype stage:

• Profile completeness thresholds: Profiles stay hidden below 30% completion; manager features unlock at 50%
• Privacy tiers and overrides: Three visibility levels (Private, Team, Public) with temporary overrides and audit logs
• Data-availability fallbacks: Explicit flows for missing SSO/HRIS data and manual profile creation
• Relationship conflicts: Users can reset relationship data and override compatibility scores when AI misreads dynamics

Phased Delivery Given limited engineering bandwidth, delivery was split into phases that could each be used standalone:

• Phase 1 – Foundation: Basic profile page, privacy controls, manual editing, SSO integration
• Phase 2 – Insights: Relationship tab, interaction timelines, compatibility analysis
• Phase 3 – Integration: Colleague cards in chat, dynamic sidebar, context block display, calendar hooks
• Phase 4 – Scale: Manager team view, bulk profile management, analytics dashboard
• Phase 5 – Optimization: Performance, mobile, multi-player profiles, advanced insights/APIs

Architecture-level decisions were documented but right-sized for the prototype so it could run credibly on Vercel while guiding future implementation.

Result

Even as a prototype, the Profile Insights Dashboard and the underlying strategy doc shifted how Nadia's leadership thought about the product:

Strategic Impact

• Gave teams a shared language ("Heart Rate", "Context Blocks", "Trinity") to prioritize features and cut noise
• Turned an abstract "org nervous system" vision into a clickable, testable artifact that could be shown to customers, investors, and internal stakeholders
• Provided a reusable spine for future AI behaviors—agents that surface the right blocks at the right moments—without committing to a full conversational rebuild on day one

Design Direction The prototype defined and visualized foundational pillars that converted into a prioritized implementation plan (phased rollout from foundation to scale) instead of 13 competing P0s. By codifying navigation, states, and AI patterns in a unified workspace, it reduced ambiguity for engineering.

Next Steps The work established a foundation for:

• Running structured tests with managers and talent leaders to validate which insights and views are most actionable
• Extending the same primitives into meeting prep, journeys, and team analytics to create a coherent Nadia Intelligence platform
• Building agent behaviors that know when to surface context blocks based on manager state and needs