Make capture effortless

Meridian is an enterprise construction management platform for complex projects across planning, operations, documentation, and reporting. It had a fundamental adoption problem: the people on the job site weren't documenting in it. They relied on phone cameras, text messages, group chats, WhatsApp, Slack, Teams, and handwritten notes. Field teams didn't lack technology; the enterprise system demanded more effort than the workflow they had already built for themselves, so important project information sat fragmented across devices and conversations.

The instinct in enterprise software is to fix adoption with training, reminders, or new forms. The behavior on site said something else: the capture experience is the problem. The product should adapt to the worker instead of forcing the worker to adapt to the software.

• Capturing one observation in Meridian meant navigating to a project, selecting a location, completing a form, uploading a photo, and classifying it.
• Field teams already had a faster workflow (phone cameras and group chats), so that's what they used.
• The result: project information existed, but fragmented across devices, conversations, and apps.
• Feature depth didn't matter if the input layer was broken; more downstream capabilities would sit on unreliable data.

Product Leader, Product Strategist, and UX Strategist for my Product Leadership capstone. I owned the strategy end to end: the product insight, the layered architecture, the experience principles, and alignment across the executive team.

1. 1.

   Field Capture Engine

   A lightweight ingestion layer between field workers and Meridian. A user sends a photo, message, or voice note through a channel they already use, and the system determines the context automatically: project, location, timestamp, employee, work type, issue category, related task, potential safety concern. Capture now. Organize automatically.

2. 2.

   Context Intelligence Layer

   AI transforms raw field inputs into structured enterprise data. A photo is no longer just a photo; it becomes photo + project + location + time + worker + work activity + potential issue. A text message can become a project event, punch-list item, safety observation, or daily-log entry.

3. 3.

   Enterprise Intelligence

   With field activity structured, it can feed the workflows the rest of the platform runs on: daily logs, punch lists, scheduling, compliance, payroll, safety, reporting, progress tracking, and risk identification. Field Capture stops being a better photo upload and becomes a data acquisition layer for the entire platform.

4. 4.

   Designing for speed

   The experience target: reduce documentation time from roughly 90 seconds to under 20. One-handed capture, minimal input, familiar channels, confirm instead of configure. If AI decides where an event belongs, the user can see and correct it. Automation reduces effort without removing control.

5. 5.

   A roadmap in three rocks

   Rock 1 ships the Field Capture Engine to solve the immediate adoption problem. Rock 2 adds intelligent classification and contextualization. Rock 3 turns the improved data foundation into enterprise intelligence: scheduling, compliance, payroll, safety, forecasting, risk.

• Low adoption is product feedback, not a change-management problem. Users choose the workflow with the least friction, even when the company wants them to use something else. The right question was never how to get field workers into Meridian; it was why they were choosing something else.
• The phone camera wasn't the competition. It was evidence of what the user actually needed. The opportunity was to make Meridian work more like the behavior they had already chosen.
• Adoption before features. Improving one small interaction (the input layer) improves reporting, retention, and the economics of the entire platform.
• AI does the administrative work, not the observation. AI's role is to infer context, classify, and suggest structure. The human still owns what they saw.