Role: UX/UI Lead, Systems Architect, Ethical AI Integrator
Tools: Figma, React, Mistral Small 3.1 API, Custom Supply-Chain API
Duration: 8 weeks (Closed Beta)
A global leader in logistics intelligence approached me with an ambitious challenge. They wanted to reimagine their enterprise-grade tracking and AI systems into a modular, mobile-first tool designed specifically for small and medium-sized businesses. Their core platform caters to Fortune 500 clients with complex integrations and high-volume orchestration, but this new initiative aimed to make their technology more accessible.
The goal was to provide lean teams with the clarity, support, and responsible AI tools they need to succeed.
I worked on this project under a non-disclosure agreement, but I've been given permission to share the design work publicly without revealing the client's identity. Throughout the project, I focused on creating user research, persona development, and interface logic that could be applied to a wide range of SME workflows. I prioritized transparency, cognitive clarity, and cost-efficient deployment. The result was Strata, a lightweight tool that's intuitive, adaptable, and genuinely helpful.
The goal was to provide lean teams with the clarity, support, and responsible AI tools they need to succeed.
I worked on this project under a non-disclosure agreement, but I've been given permission to share the design work publicly without revealing the client's identity. Throughout the project, I focused on creating user research, persona development, and interface logic that could be applied to a wide range of SME workflows. I prioritized transparency, cognitive clarity, and cost-efficient deployment. The result was Strata, a lightweight tool that's intuitive, adaptable, and genuinely helpful.
Research and foundations.
Challenge
Small and medium-sized enterprises wrestle with unpredictable per-shipment fees, fragmented APIs, and black-box logistics platforms. Strata needed to reshape this complexity into a six-shipment companion under a fixed monthly plan, delivering real-time risk alerts and human-centric AI guidance without overwhelming solo operators.
Problem Definition
Logistics teams face significant challenges:
Fragmented data sources require constant dashboard switching, disrupting workflow.
AI recommendations lack transparency, eroding trust and slowing decision-making.
Unpredictable billing models make it difficult for small and medium-sized enterprises (SMEs) to scale sustainably.
The Design Challenge
How can we design a lean, mobile-first tracking solution that:
Provides critical insights at a glance
Offers one-tap solutions to common issues
Transparently reveals AI logic to build trust and prevent errors?
Design Principles & System Logic
As I designed Strata, I guided my work with three key principles:
Cognitive Clarity: I prioritized simplicity and transparency, ensuring each screen provides the right information at the right time. By using progressive disclosure and visual clarity, I aimed to minimize decision fatigue and make it easier for users to focus on what's important.
Modular Extensibility: I designed components to be flexible and adaptable, fitting seamlessly into different SME workflows without requiring extensive customization. Whether it's shipment tracking or risk alerts, every element is reusable and scalable.
Ethical AI Augmentation: I believe AI should be transparent and trustworthy. That's why I ensured each recommendation includes clear explanations, empowering users to understand the reasoning behind the suggestion and override decisions when necessary. This approach builds trust and ensures users remain in control.
User Journey Visualization: From Enterprise Complexity to SME Clarity
Before designing Strata’s interface, we needed to understand how real users navigate logistics risks—not just in theory, but in practice.
We began with anonymized journey data from the parent large enterprise platform to identify common pain points. From there, we iterated through eight versions of the SME journey map during closed beta, refining each step based on user feedback.
Early maps were text-heavy and focused on simulation logic but lacked emotional clarity. Over time, we introduced key elements like the “Why?” button, confidence bar, and shared decision checkpoints.
This final visualization—approved for public release—captures the emotional and cognitive flow of two core personas: Maria, a solo logistics owner, and David, a compliance manager.
Design Process.
Desktop Dashboard
Mobile Dashboard
Why button Reveals AI reasoning with source links and confidence scores—guardrails against hallucinations.
Shipment Card Displays status, ETA, and timeline dot. Actions only surface on risk states to reduce noise.
Parses legal clauses and updates routing logic automatically.
Simulate Reroute screen
When designing Strata's interface, I focused on creating simple, powerful components that are easy to use. Each component is optimized for clarity and reusability.
The Shipment Card shows status, ETA, and timeline, with actions only appearing in risk states.
The Reroute Button is context-aware and mobile-optimized, providing relevant options for adjusting plans.
The "Why?" Insight Pane provides transparency into AI decision-making, showing users the reasoning behind each recommendation.
The Port Data Agent fetches live port status, while the Social Disruption Feed flags potential issues like weather delays and port risks.
I prioritized extensibility, accessibility, and consistency across devices and documentation, creating a system that's both powerful and intuitive. By focusing on a tightly scoped set of components, I was able to design a tool that's easy to use and adaptable to different needs.
The Reroute Button is context-aware and mobile-optimized, providing relevant options for adjusting plans.
The "Why?" Insight Pane provides transparency into AI decision-making, showing users the reasoning behind each recommendation.
The Port Data Agent fetches live port status, while the Social Disruption Feed flags potential issues like weather delays and port risks.
I prioritized extensibility, accessibility, and consistency across devices and documentation, creating a system that's both powerful and intuitive. By focusing on a tightly scoped set of components, I was able to design a tool that's easy to use and adaptable to different needs.
Strata’s final design followed a lean, iterative path:
Mobile-First
Six closed cards show status, ETA, tracking ID, and timeline dot.
Actions appear only on risk states to minimize noise.
“Why?” button slides in a logic pane with source links.
Port Data Agent actions animate API queries and display real-time port metrics.
Displays status, ETA, and timeline dot. Actions only surface on risk states to reduce noise.
“Why?” Insight Pane - Reveals AI reasoning with source links
“Why?” Insight Pane - Reveals AI reasoning with source links
Reroute options and simulation
Reroute Simulation on Tablet/Desktop
Customs information at glance
Customs information at glance on Tablet/Desktop
Downloadable Logic with links
Port Data Agent Fetches live port status via API, bypassing OCR and manual entry.
Contact Customs agents directly
Strata’s AI layer offers transparent reporting through detailed, source-attributed outputs that clearly explain the rationale behind each decision. Every action is logged in a comprehensive decision audit trail, providing full visibility into the system’s reasoning and supporting data—empowering management with oversight, accountability, and traceable intelligence.
Scalable System Tokens
Material 3 Expressive color and typography scale from mobile through desktop.
The prototype was tested with real SME users, refining clarity, responsiveness, and trust signals at every step.
AI Implementation
Augmenting Human Decision-Making
Strata's AI supports, rather than replaces, human judgment. Every recommendation is transparent, customizable, and based on public data.
Key Principles
Augmentation over Automation: AI suggests alternatives, but users confirm and override decisions.
Transparency: "Why?" insights reveal decision paths and data endpoints.
Data Boundaries: Recommendations are powered by public port feeds and client contracts, with no external datasets or profiling.
Affordability: A capped monthly plan ensures access without scale penalties.
Ethics by Design
Strata's architecture prioritizes clarity, control, and human-centered design. Every component is built with the user in mind, emphasizing transparency, accountability, and user trust.
Why We Chose Mistral 3.1: A Balance of Performance and Practicality
When building Strata's AI companion, we needed a solution that was fast, transparent, and cost-effective - especially for small and medium-sized businesses with limited resources. After careful consideration, we chose Mistral Small 3.1 for its unique combination of benefits.
Here are three key reasons why Mistral 3.1 stood out:
Cost-Efficient Reasoning: Mistral's API provides high-quality insights at a lower token cost than more complex models. This allows us to offer real-time support without sacrificing scalability or reliability.
Lightweight and Fast: With fewer parameters and faster response times, Mistral 3.1 enables us to filter and prioritize information effectively without introducing latency or overfitting risks.
Streamlined Data Processing: By using Mistral to parse structured data directly, we can skip expensive OCR pipelines and reduce infrastructure complexity. This makes our system more accessible and user-friendly, especially for mobile users.
Choosing Mistral 3.1 was both a technical and strategic decision. It aligns with our commitment to creating a system that's not only powerful but also transparent, scalable, and cost-effective.
AI Transparency & System Logic
Understanding how "Why?" insights are generated through our transparent AI pipeline
From Monolith to Modular: How We Cut Vision Costs by 99% Without Sacrificing Trust
Discover how Strata redefined vision processing—cutting costs by 99% while enhancing transparency and control. Read the full breakdown below.
Beta & Iteration
To ensure Strata met the needs of small and medium-sized logistics businesses, we conducted a six-week closed beta testing phase with 10 SMEs. This rigorous testing process allowed us to evaluate Strata's performance in real-world scenarios, focusing on three key areas:
Clarity: How intuitive and easy to use is Strata's interface? Do users quickly understand the information presented and the actions they need to take?
Trustworthiness: Do users trust the insights and recommendations provided by Strata's AI companion? Are they confident in the system's ability to support their decision-making?
Effectiveness: How well does Strata support task success across a variety of logistics workflows? Can users achieve their goals efficiently and accurately using the system?
Through this beta testing phase, we gathered valuable feedback and insights from our SME partners. This feedback informed further refinements and improvements to Strata, ensuring that the system meets the needs of its target users.
Reflection & Next Steps
Current Status and Key Takeaways
We're currently testing Strata with a select group of partners in a closed beta. While performance data remains confidential, early feedback has been encouraging, validating our approach to clarity, modularity, and responsible AI.
This project has reinforced several key lessons:
Modular design enables adaptability and scalability across diverse workflows.
Storytelling plays a crucial role in building trust in AI-driven systems.
Ethical design it's an opportunity to create a better experience for everyone.
Strata serves as a test case for human-centered AI, prioritizing transparency, flexibility, and usability. As we refine the system, our focus remains on putting people first.
This case study outlines our design process and intentions, rather than the final outcome. If you'd like to learn more or discuss potential collaborations, I'm open to chatting further.
