Unified AI Product Vision
Establishing Northstar vision, design governance, and scalable ecosystem for an AI-first platform.
About the project
Company
Level Access
Timeline
2026
My role
Product strategy, Product designer, AI UX
Challenge
Context
As AI capabilities were rapidly integrated into the platform, teams were building in silos without a defined "AI-first" standard. There was no unified strategy dictating how these features should coexist, scale, or look, creating deep ambiguity.
Design Governance
Establishing the rules of engagement for cross-functional pods building AI features.
Ecosystem Vision
Defining the Northstar strategy for how distinct AI agents interact.
Ambiguity Management
Validating theoretical strategy through tangible, interactive prototypes.
Long-Term AI Design Governance Foundation
To transform isolated AI initiatives into a unified platform layer, I designed a scalable long-term AI governance framework. My focus was establishing a shared cross-functional foundation to align product and engineering teams on what, how, and why we build AI.
Aligning Product Concept
Developed an extensive experience vision defining our "Identity and Positioning" as the Guided Intelligent Workflow for accessibility.
The AI Product Glossary
Standardized naming conventions for all agentic interaction modes (Ask, Guide, Draft, Execute, Learn) so teams could communicate without friction.
A Shared Design System for AI
Created reusable component guidelines and explicit behavior rules (such as Transparent context use, Human-in-the-loop triggers, and Output anatomy) to eliminate fragmented experiences.
The strategy and execution
Due to confidentiality, I can only share a limited view of the project. Here is how I structured the ecosystem.
A Cohesive Platform Experience
Rather than adding disconnected AI features, we visualised Level AI as a consistent layer embedded natively into existing platform surfaces. Focus on the spatial integration.
Platform-Wide Search
Platform-Wide Integrated Experience
Powerful Context Capabilities
Platform-Wide Context Grounding
Native AI UX Patterns Architecture
Moving from fragmented feature-building to a platform-wide cohesive agentic experience meant embedding AI natively into users' existing workflows; incorporating natural, context-aware visual patterns across every workspace.
What I did
Architected the AI integration dashboards, contextual side panels, and in-line workspaces, including search bars and auto-agent triggers. To sustain user trust, I designed real-time screen-reader friendly "thinking" loading UI states, progressive text "streaming," and inline inline artifact generation. This allowed users to generate tasks, audits, and code directly inside active workspaces using grounding platform data as natural context.
Impact
Unified fragmented capabilities into a single interactive environment, successfully positioning Level Access as an intelligent ecosystem rather than a tool destination.
Secured executive approval to integrate this unified Northstar vision into the upcoming yearly roadmap.
AI-Driven Onboarding and Setup
An intelligent platform is only as good as its activation rate. Moving legacy enterprise users into an AI-driven, customized workflow required dismantling "blank canvas" intimidation through progressive setup patterns.
What I did
Prototyped a guided, personality-driven AI setup flow that collects role-based focus areas and organization data during initialization, giving users immediate onboarding context.
Built interactive prototypes of a dynamic, AI-driven onboarding workflow to demonstrate a seamless transition to the new UI.
Impact
Reduced estimated user setup time by 50% during usability testing, using personalized progressive disclosure to validate and launch high-impact product experiences with ease.
Considerations
Mitigating 'Black Box' Anxiety with UI Scaffolding
Users mistrust AI when outcomes feel unpredictable. I created a strict design principle of Understand - Recommend - Preview - Act - Learn across the platform, forcing the system to explicitly preview expected changes, assume permissions, and require explicit confirmation for consequential actions.
Navigating Ambiguity & Stakeholder Alignment
To eliminate ambiguity and fragmented UX when building AI products, I authored a centralized AI glossary and governance framework. By defining strict design standards and a shared vocabulary, this core hub provided the necessary guardrails for Engineering, Product, and Design to execute fast, independently, and cohesively.
Results
Metrics
Roadmap approval
Successfully translated abstract platform architecture into a tangible business case, securing executive buy-in for the upcoming product cycle.
Cross-functional alignment
Consolidated siloed development efforts into a single, standardized AI product glossary utilized by multiple engineering pods.
Next steps
Global governance rollout
Systematically deploy the authored AI design standards across all remaining product teams.
Deepen multi-agent orchestration
Expand the Northstar strategy to define exactly how specialized AI agents collaborate on complex, cross-document accessibility audits.
Key takeaways
Governance is the bedrock of scale
Authoring clear standards and product glossaries proved just as critical as the UI itself. Without rules of engagement, multiple pods building AI features will inevitably create a fragmented UX.
Systems thinking beats isolated features
Designing for an AI-first platform means orchestrating continuous flows, agent hand-offs, and capabilities across the entire ecosystem, not just polishing single touchpoints.
Prototypes secure executive buy-in
Translating abstract architectural strategies into tangible, interactive onboarding workflows provided the hard usability data needed to validate the Northstar vision with leadership.
Ambiguity requires defined constraints
In a rapidly shifting AI landscape, creating a shared glossary and unified design standards gave cross-functional teams the confidence to innovate quickly without breaking the overarching platform strategy.
