My role
Sole product designer, full-cycle ownership
Discovery & Research · information architecture · UI · Design system · design–engineering handoff · Design QA
Product
0-to-1 B2B SaaS platform that connects fields, machinery, tasks, staff and seasonal planning into a single map-first interface for farm managers and agronomists across the Russia/CIS market.
Context
Farm operations were managed across disconnected systems — paper-based task logs, separate GPS tracking, and manual spreadsheets for field records. Managers had no unified view of where machinery was, what tasks were active, or whether field workers were following plans. The business needed a single operational platform that worked for both desk-based managers and field supervisors with limited tech familiarity.
Main Challenges
Technical constraints - external third-party system which was not intended for a complex B2B product
3 type of non-technical users- managers, agronomists, Field supervisors
Problems
No real-time visibility
Managers had no live view of their fleet
Machine locations were tracked through phone calls and end-of-day reports. Violations — speeding, zone exits, unauthorized stops — were discovered only in retrospect, often weeks later during GPS export review.
Fragmented data
Fields, tasks, and machines lived in separate tools
Agronomists maintained field maps on paper, crop history in spreadsheets, and machine availability by calling the fleet manager. Producing a single coherent plan required cross-referencing three sources that never agreed.
Paper-based workflows
Task assignment and completion were manual processes
Task logs were paper forms collected every two days. Completion was verified by agronomists physically visiting fields. Data was regularly lost, duplicated, or entered late — making historical analysis unreliable.
Manual compliance
Violations couldn't be linked to operators or tasks
Speed violations and zone exits had no accountability trail. There was no mechanism to attach a violation to a specific operator, task, or machine in real time — only in a monthly export that nobody had time to review fully.
"I call the foreman. Sometimes he knows, sometimes he has to go check. By the time I get an answer it's been 20 minutes."
— Farm manager, discovery interview
User Research & Insights
I worked with farm managers, agronomists and field supervisors through a combination of interviews, surveys, workflow observation, and PM/client collaboration.
Stakeholder and user interviews
Conducted discovery sessions across all three target roles — farm managers, agronomists, and field supervisors. Interviewed participants in their working context where possible, with questions focused on existing workflows, tool usage, daily frustrations, and ideal future states. PM and client input supplemented direct sessions, providing operational context that users often couldn't articulate formally.
Competitor and market benchmarking
Analysed comparable agricultural management and fleet monitoring platforms in the CIS market. Identified established conventions (map-first layouts, color-coded field polygons, timeline playback) and critical gaps — particularly around multi-role support and task-to-field linkage, which competitors handled poorly or not at all. Previous tool adoption failures in this sector were consistently traced to complexity and software-generic terminology.
Existing workflow observation and analysis
Reviewed paper-based task logs, spreadsheet field records, and the business requirements document before designing anything. This was essential to prevent the 0-to-1 failure pattern of designing for an idealized workflow rather than the real one — including the handwritten crop history notebooks that agronomists had maintained for up to eight years.
Farm manager
Strategic oversight
Needs real-time visibility across all fields and machines without switching between tools. Primary concern: violations, task completion rates, and fleet utilization. Often mobile, prefers map over dashboard.
Agronomist
Crop planning and field assignment
Works with field boundaries, crop rotation history, and seasonal task planning. Needs to assign machinery to specific geozones and track work completion against plan across the growing season.
Field supervisor
On-ground task execution
Operates in the field with limited connectivity. Needs to see machine and operator assignments quickly without navigating complex menus. Limited digital literacy — previous tools failed here.
Key design decisions
Every major decision was driven by a specific research finding — not a visual preference. The before/after framing shows the problem each decision solved.
Map as the primary interface, not a feature
Problem
existing tools buried the map behind menus. Users with low digital literacy abandoned them within days.
Decision
monitoring is the product's landing view. The map occupies 60% of the viewport at all times. All other data — machine details, alerts, field info — surfaces contextually on selection without navigating away from the map.
Sidebar status hierarchy replaces a flat machine list
Problem
Machines in a flat list with no status differentiation. Managers had to scroll through everything to find problems.
Decision
Machines are grouped by operational state — working, alerts, offline. Alert machines appear first, red, with the specific violation (speed, zone exit, idle) visible without clicking. Managers see the problems before they see anything else.
Task creation consolidates three separate workflows
Problem
Agronomists maintained fields in one tool, machines in another, and personnel in a third. Planning a single task required 2 hours of cross-referencing daily.
Decision
Task creation is a single three-section accordion — basic parameters, geozones with area totals, and machinery with per-day personnel scheduling. Each section shows a summary count when collapsed so the user sees task scope at a glance. A system warning prevents the most common data-loss error: changing key parameters clears assignments.
History playback replaces manual GPS export review
Problem
Compliance review was a weekly manual process — export GPS data, review in a spreadsheet, flag violations by hand. Most violations went unreviewed.
Decision
The history module provides a timeline scrubber with event markers on the map and a color-coded action table. Managers can reconstruct any machine's full shift without leaving the platform. Violation tooltips show actual speed vs. limit directly on the map at the point of violation.
Unified reference library links all entities
Problem
Fields, machinery, personnel, and crops existed as separate records across different tools. An agronomist assigning a machine to a field had to check three sources before making a decision.
Decision
Reference books is a single library of 15+ entity types — geozones, machinery, aggregates, personnel, crops, work types, shifts, and more. All entities are defined once and referenced across monitoring, tasks, and documents. Crop rotation history is attached to every geozone and visible on the monitoring map on selection.
Key screens
Monitoring
Real-time fleet and field view
History
Shift playback with violation log
Tasks
Task creation — the most complex flow
Reference library
15+ entity types, one unified system
Fields data
import from file or edit by drawing areas
Results and Reflection
A mix of delivered outputs, verified process decisions, and projected impacts. Estimated impacts are framed honestly — this is a live production system without post-launch analytics access.
Impact:
Better agronomic planning
Operational visibility
Faster decision-making
Trust in monitoring data
Audit-ready operational history
Outcomes
Task planning time reduced significantly
The task creation flow consolidates what agronomists described as a 2-hour daily cross-referencing process into a single interface. Projected time savings are substantial — though exact figures are not available without post-launch measurement. The structural change from 3 tools to 1 flow is the design's primary claim.
Compliance violations now visible in real time
The shift from weekly retrospective review to real-time in-monitoring alerts is a category change, not an incremental improvement. Managers who previously discovered violations "at the end of the month" now see them as they occur. Behavioral impact on driver conduct is expected but not yet measured.
What I learned:
What does 0-to-1 design actually require?
The hardest part of a 0-to-1 build is not designing screens — it's resisting the urge to design before you understand the real workflow. The paper task logs and handwritten crop notebooks I reviewed before designing anything changed every major architectural decision. A product built on top of the real process, not an idealized one, earns trust from users who have been burned by tools that didn't fit how they actually work.
How do you design for users with low digital literacy?
You design for their mental model, not yours. Farm managers think spatially — they understand a map immediately and abandon menus within days. Building around the map rather than making it a feature was not a visual preference; it was the only architecture that had a chance of adoption. Every menu depth decision, every label choice, every default state was filtered through "would someone who hasn't used software in years understand this without being taught?"