Designing software to help criminal investigators connect the dots
Summary
Project Description
Slate was the minimum viable product (MVP) for in8development, a startup I co-founded to address gaps in investigative and case-management tools. Built as a native Windows application, Slate enabled investigators to import evidence, visually organize complex cases, and generate reports within a single, cohesive workflow. The core interaction model and system design were granted a patent, on which I am a co-inventor. The product focused on clarity, flexibility, and supporting real investigative processes rather than forcing users into rigid structures.
My Role
I led all aspects of product design, including user research, visual, and interaction design. I also served as product manager, defining product strategy, prioritizing features, and writing development specifications. This role required balancing user needs, technical constraints, and business goals while shaping the product from concept through MVP.
Users & Environment
The product was designed in close collaboration with four federal special agents working complex financial crime investigations. I worked directly with these subject matter experts to understand their workflows, pain points, and limitations of existing tools. To broaden perspective, I also interviewed certified fraud examiners, who represented a key segment of the initial target market and helped validate alternative use cases.
“Aryk is an accountable and hardworking teammate, demonstrating a willingness to own the design from user research and low-fidelity mock-ups to prototyping, visual design, usability testing, and iteration. ”
Research
Note: For much more detail on my research and design process, check out: Select an Airplane Case Study
Canvas & Table
To inform the design of the Canvas, I facilitated working sessions with federal special agents to observe their workflows and better understand the challenges they faced and how investigations actually unfold. These sessions focused on how evidence is collected, compared, grouped, and revisited over time. The research led to multiple proposed approaches for canvas layout, interaction models, and card styles, all grounded in how investigators naturally externalize and organize information.
In parallel, the Table view existed as a conceptual placeholder throughout early development, which created space for discussion and exploration. Research revealed that investigations often involve many different object types, making a single, static, Excel-style table impractical. Instead, the solution evolved into repositionable and collapsible table cards grouped by object type, allowing users to switch between spatial and structured views of the same data depending on the task at hand..
Notes from user interviews on building an interactive workspace
Notes from user interviews on building a table viewer
Rough sketches of canvas concepts during initial interviews
Initial sketch of a table concept
Canvas Mockups
Description
The Canvas was the primary workspace within Slate, functioning as an open, spatial environment where investigators could visually organize and reason through complex cases. Similar to collaborative canvases used for brainstorming or diagramming, the Canvas allowed users to freely arrange information. Unlike those tools, it was designed specifically to support investigative work by preserving context, rationale, and evolving understanding behind relationships.
Problem
The challenge was to replace a mix of physical and digital tools, including whiteboards, string boards, and complex Visio-style diagrams, with a single, flexible workspace. Investigators needed to spatially organize large volumes of evidence while also qualifying and validating the relationships between items. Given that investigations could involve thousands of objects, each element needed to convey meaning beyond a simple icon and remain usable at scale.
Solution
Through user interviews and A/B testing, I determined that a card-based representation provided the right balance of flexibility and information density. Each object appeared as a movable card on the canvas, allowing users to cluster, group, and relate items spatially. Relationships were created directly through interaction, and when additional context or metadata was required, the system guided users through a focused workflow.
The Canvas supported zooming and panning to let users move fluidly between high-level structure and detailed analysis, similar to modern collaborative canvases. This approach allowed investigators to externalize their thinking, explore connections, and maintain confidence in the integrity of their work as cases evolved.
Table Mockups
Description
The Table provided a structured view of the same objects shown on the Canvas, offering a complementary way to explore and manage investigative data. It supported the Canvas by exposing full object properties, provided a familiar interaction model for Excel power users, and enabled data export for reporting and analysis.
Problem
While most target users were highly proficient with Excel, a traditional spreadsheet model did not scale to the complexity of real investigations. Cases could include hundreds of object types, each with distinct properties, resulting in tables that were thousands of columns wide and difficult to navigate. A single, monolithic spreadsheet would have hindered usability rather than improved it.
Solution
User research showed that a card-based table approach better matched investigative workflows. Each object type was represented as its own table card, allowing users to focus on relevant data without being overwhelmed. Table cards could be repositioned and collapsed as needed, while retaining familiar spreadsheet capabilities such as sorting and filtering.
Edits made within the Table automatically updated the underlying data and were reflected across all other views, including the Canvas. This ensured consistency while allowing users to move fluidly between spatial and structured representations of the same information.
Fuzzydate
Description
The Fuzzydate was a custom date and time input designed specifically for Slate to support how investigators actually describe events. Rather than forcing artificial precision, the component allowed users to capture approximate, uncertain, or evolving temporal information. This interaction model was novel enough that it became part of the product’s granted patent, reflecting its role as a core innovation rather than a standard UI control.
Problem
Investigative work frequently relies on imprecise temporal language. Users needed to record information such as “on or about noon,” “sometime in June,” or “before the end of 2017,” without losing meaning or introducing false accuracy. Traditional date and time pickers assume certainty and precision, making them poorly suited for law enforcement workflows where details are often incomplete and/or refined over time.
Solution
I explored variations on traditional date and time pickers and extended them to support qualifiers and degrees of certainty. The final design allowed users to apply contextual modifiers such as on or about, before, or unknown, while still capturing as much structure as possible. When specific values were not known, the interface supported intentionally vague or placeholder inputs rather than forcing guesses.
A persistent, human-readable readout provided immediate feedback, reflecting the combined date, time, qualifiers, and time zone in plain language. This helped users understand exactly what had been recorded, reinforcing confidence and accuracy while preserving the flexibility required for investigative work.
Timeline
Description
The Timeline view combined the spatial flexibility of the Canvas with a chronological representation of events. While many tools exist for building timelines, they typically focus on isolated events rather than the broader investigative context. This view was designed to help agents understand when things happened, how they were connected, and what else was happening at the same time.
Problem
Existing timeline tools used by agents were largely static and manual. Solutions like Visio required investigators to create and maintain timelines by hand, duplicating effort already captured elsewhere in the investigation. As cases evolved, timelines quickly became outdated, disconnected from underlying evidence, and difficult to compare across related events or suspects.
Solution
Through iterative design exploration, I focused on two capabilities that significantly expanded the value of the timeline. First, the timeline displayed related objects alongside primary events, allowing agents to see people, assets, and transactions in context rather than as isolated points in time. Second, the interface supported multiple timelines within a single view, enabling comparison across suspects, accounts, or parallel investigative threads.
Crucially, the Timeline was fully synchronized with the Canvas. Any updates to evidence or relationships were immediately reflected across views, eliminating redundant work and ensuring consistency. The result was a dynamic, context-rich timeline that supported pattern detection, and comparison as investigations evolved.
Swimlanes: The ability to compare two or more timelines to each other