A barebones notes viewer became the research management system that hedge funds are migrating to — displacing tools that have been shipping for a decade. 100+ screens. 9 feature areas. 33K+ notes trusted to the platform by real clients.
Arcana is a financial analytics platform used by hedge funds and asset managers. The platform handled portfolio analytics, risk decomposition, and factor modeling. But clients kept asking the same question: “Where do we put our research?”
The problem was not a missing feature. It was fragmented workflows across competing platforms. Analysts maintained research in Outlook inboxes, OneNote notebooks, Word docs, Slack threads, and dedicated RMS tools like Bipsync and Verity. None of these systems talked to each other, and none were connected to the portfolio data Arcana already provided. Notes weren’t linked to tickers or positions. Search failed at scale. Emails with critical sell-side research sat untagged and unsearchable. Meeting notes couldn’t be tied to calendar events. There was no pipeline to track ideas from thesis to execution, and no intelligence layer to surface patterns across thousands of documents.
The design challenge was to pull these scattered workflows into one research operating system: notes, templates, search, email ingestion, calendar, collaboration, AI intelligence, and CRM-like workflow tracking for contacts, meetings, and pipeline. The product had to be good enough that firms would migrate tens of thousands of notes away from tools they’d used for a decade.
Hedge funds arrived at Arcana for portfolio analytics. The research workflow question kept surfacing in calls because incumbents — Bipsync, Verity, EDS — were standalone tools with no connection to live portfolio data. Firms were paying for two separate systems and manually bridging them.
The opportunity was to use Arcana’s native portfolio and market data layer as a competitive advantage. Because RMS was being built on the same platform, notes could be linked directly to positions, tickers, and risk data that no standalone competitor could match. The displacement opportunity existed because the connection between research workflows and the broader investment platform had never been solved.
The chosen direction treated RMS as an operating layer beneath the whole platform rather than a pile of adjacent tools. Each of the 9 feature areas had to stand on its own: a firm could adopt just email ingestion, or just the calendar, or just search. The system design made sure that, used together, they reinforced each other.
The risk of the alternative approach — shipping disconnected modules incrementally — was a product that felt bolted on rather than built in. The design decisions across notes architecture, field taxonomy, sharing model, and search were made with the full system model in mind, because the cost of re-architecting core structures after client migrations would have been prohibitive.
Each phase built on the last. The product got more complex, but the design system and component library established in phase one meant each new feature started further ahead. By month 6, entire feature areas that would have taken a month at the start were shipping in a week.
Each feature area below was covered end-to-end: competitive analysis, information architecture, interaction design, prototyping, and production specs for engineering handoff. The design challenge across each area was not just what to surface, but how it connected to the rest of the system — because decisions made in the notes architecture affected how search, email, AI, and calendar would function downstream.
Every feature area began with a structured product audit. Bipsync, Verity, EDS, AlphaSense, and FactSet were analyzed across email ingestion, calendar integration, search, templates, AI, mobile, and portfolio connectivity. The analysis served two purposes: identifying gaps where incumbents were weak, and recognizing patterns familiar enough to users that deviating from them would create friction without benefit.
The decision framework was not about matching feature for feature. It was about choosing where familiarity would accelerate adoption (note types, pipeline stages, shared templates) and where differentiation would create durable competitive advantage (native portfolio integration, zero-config email sync, AI over the full research corpus). Each of those choices is traceable in the feature architecture.
| Dimension | Bipsync | Verity | EDS | Arcana |
|---|---|---|---|---|
| Email ingestion | Outlook plugin + AI Mail Classifier | Outlook add-in (manual deposit) | Outlook add-in | Full mailbox sync, zero config |
| Custom field filters | Admin-defined filter shortcuts | Basic type/tag filtering | AI-driven classification | Rule builder, AND/OR, 6 field types |
| Calendar | Day/Week/Month + Rules Engine | Wall Street Horizon + Outlook sync | None documented | Earnings + macro + shared + personal |
| Portfolio integration | None | Limited | Quantitative models | Native — notes linked to positions |
| AI intelligence layer | None | None | LinqAlpha partnership (limited) | Native chat, summaries, citations |
| Teams / Slack integration | None documented | None documented | None documented | @mention creates note from thread |
| Migration from competitors | — | — | — | Documented playbook, 3 migrations done |
Platform-scale product design is a sequence of tradeoffs. These were the recurring tensions across the 9-month build.
The solo designer role meant owning the full design surface — from competitive research and information architecture to interaction design, prototyping, and engineering handoff. There was no design team to delegate to, no design lead to escalate to. Every decision at the design layer was a direct output of this role.
The collaboration model was tightly integrated with the PM. Product direction, PRD development, and competitive analysis were joint outputs. The design thinking was embedded in the product specs before Figma was opened, because the decisions that mattered most were made at the requirements layer, well before the visual one. Engineering handoff went past screens to interaction specs, edge-case matrices, and state documentation.
Stakeholder alignment happened through working prototypes rather than static presentations. Because the product had no prior design precedent, HTML prototypes at key decision points were more effective at building shared understanding than slides. The calendar feature went from concept to production handoff in 7 days using this approach.
The design role extended to product specification. Each PRD below incorporates competitive teardowns, user workflows, interaction specs, edge case matrices, and design rationale. These are not design-only documents — they are the full product specs engineering builds from, with product direction and design thinking embedded throughout.
Each card below is a standalone case study covering a specific design challenge. Each one traces the problem, the decision framework, and the outcome.
The hard part was never a single screen. It was turning the scattered research habits of eight different desks into one workflow they'd actually adopt — without flattening what made each of them distinct.
Nine months. Nine feature areas. One designer embedded in every decision from customer call to engineering handoff. The product is not a set of screens. It is a system of decisions that compounds at enterprise scale.
In a portfolio presentation, the focus would be on three decisions: how the system model was defined across 9 feature areas, how competitive gaps informed product differentiation rather than just feature parity, and how AI-assisted workflows were integrated without fragmenting the research experience or compromising analyst trust in sourced information.