Case study — Cobria / JAK Solutions
27%
reduction in operational task time — two legacy systems dissolved into one coherent platform
02 — The problem worth inheriting
JAK Solutions operated two separate platforms serving construction professionals in Belgium. Hello Jak handled quotes, invoices, and operational job management. SoS Roof handled roofing-specific workflows. Both had paying customers. Both had established data models. Both had accumulated years of operational dependency. The brief was not to build something new — it was to make two incompatible systems behave as one, without breaking what either group relied on.
The three constraints that made this hard
Live user base
Existing customers could not be disrupted. Workflows already embedded in daily operations required a migration path, not a replacement.
Unstructured business rules
Product logic lived in Notion — unformatted, cross-referenced, and contradictory in places. No formal IA existed before this engagement.
High data density
Construction admin is inherently complex — quotes with dozens of line items, invoices with VAT variants, multi-user planning. The UI had to handle volume without fatigue.
03 — My role and mandate
My engagement at JAK Solutions was defined by a specific problem: the product team had deep domain knowledge and years of product iteration, but no formal UX infrastructure. Business logic was documented in Notion. There was no Information Architecture. There was no module hierarchy. My role was to impose that structure — fast, precisely, and without disrupting ongoing development.
Three strategic execution pillars
Requirements to logic
Translated unstructured business rules from Notion into a formal Information Architecture — module definitions, trigram identifiers, color coding, dependency mapping, and data relationships. Created the structural foundation the engineering team could build against.
Eliminating data silos
Designed a Single Source of Truth architecture bridging field-site data and office management. The Contact module was architected as the entry point for every Quote, Invoice, and Email — a dependency decision that determines how the entire platform scales.
Low-fidelity foundations
Prioritised functional integrity and system logic over visual polish. Construction professionals operate under real-time pressure — the UI had to handle 200+ invoice records, complex VAT scenarios, and multi-user planning without cognitive overhead. Reliability before aesthetics.
My twelve years as a practicing architect — managing procurement, subcontractor contracts, and project documentation on major construction sites — meant I could read a construction invoice workflow the way a developer reads code. This domain fluency compressed the research phase and eliminated translation overhead between client and design.
04 — The architecture I built
Before a single screen was designed, the module system was defined. Each module received a formal specification: a trigram identifier, a colour code, a stated purpose, dependency relationships, and data contracts. This was not documentation after the fact — it was the prerequisite for coherent design. Click any module to inspect its specification.
The critical architectural decision — contact as system entry point
05 — Outcomes and delivery
The system shipped. Cobria is live at app.cobria.com, serving construction professionals across Belgium in five languages. The consolidation delivered quantifiable reductions in operational overhead — and the AI-driven UX work attracted recognition beyond the product's user base.
Belgium Startup Awards 2026 — AI category finalist
Cobria was selected from over 100 applicants as one of 51 national finalists — and one of 9 AI category nominees. The jury evaluated the product's AI integration, innovation, and scalability. The nomination is a direct consequence of the Max AI UX architecture designed during this engagement. Ceremony: June 2, 2026, Brussels.
What was delivered
06 — What I learned
The lessons from Cobria are not about UX patterns or component libraries. They are about what happens when you treat information architecture as engineering work — precise, testable, consequential — rather than as a planning phase you rush through to get to screens.
The entry point decision determines everything downstream
Establishing Contacts as the system entry point for Quotes, Invoices, and Email was a data architecture decision, not a UX convenience. Every module dependency cascades. Get the entry point wrong and you build technical debt into every screen that follows.
Unstructured product knowledge is the real legacy problem
The bigger consolidation challenge was not merging two UI systems — it was reconciling two years of undocumented product decisions living in Notion. The first month was almost entirely analytical: mapping rules, identifying contradictions, building the formal specification the team had never had.
High-density data UX requires a different design philosophy
Consumer UX heuristics — whitespace, progressive disclosure, simplified flows — break down when users are processing 200-row invoice tables under time pressure. Designing for cognitive load in a data-heavy environment is a different discipline than designing for delight.
Making AI cost visible builds trust faster than hiding it
The Bricks credit system metered AI usage transparently — making the cost of intelligence legible and controllable for users skeptical of automation. This converted skeptics faster than any amount of onboarding copy would have.
What I would do differently
Earlier investment in a shared component library would have accelerated module delivery significantly. The first three modules were designed without a shared system — every iteration required manual consistency checks. The fixed cost of building the design system up front pays compound returns.
The contrast with HP — and why both are needed
HP Smart Digitization was about hypothesis testing — what could exist. Cobria was about system integrity — what must work reliably. A portfolio that only shows exploratory work signals that you cannot operate under engineering constraints. A portfolio that only shows consolidation work signals that you cannot generate new ideas. These two projects are not just different case studies — they are evidence of two different modes of professional thinking in the same person.
07 — Enterprise authority
Cobria was not a greenfield product. It was a live system used by real contractors managing real money — where a misaligned data model means a lost invoice, and a confusing UI means a missed payment. This is the environment where enterprise design authority is either demonstrated or it isn't.
HP Smart Digitization
Innovation & Discovery
Cobria
Enterprise Authority
Capabilities demonstrated