0→1 Creating for Jira for farm management

Shipped build-ready cross-platform product flows using AI-accelerated prototyping to compress design feedback loops by 67%.

Property map view — paddocks, mob markers, and the map options panelMobile paddock detail — grazing status, related mobs, and bottom navigation

Funded agtech startup building a multi-platform farm management product for farmers in Australia and the US. End-to-end design on planning tools, dashboards, and data visualisation across mobile app, mobile web, and desktop web.

Goal: ship build-ready flows fast enough to outpace a competing product racing to market.

Role
Lead Product Designer
Period
2025
Stack
FigmaMiroV0.appChatGPTGluestack-uiTailwindCSS
Mob and weight record flow — step 1 of 4Mob and weight record flow — step 2 of 4Mob and weight record flow — step 3 of 4Mob and weight record flow — step 4 of 4
Property map and paddock flow — desktop step 1 of 3Property map and paddock flow — desktop step 2 of 3Property map and paddock flow — desktop step 3 of 3

A design process focused in AI-accelerated loops

With no time for traditional discovery and requirements held mentally by the Product Owner, we leaned on AI UI generation tools (Readily, Replit) to compress the design feedback loop. AI output became our low-fi exploration: what would normally be an hour of wireframing was 20 minutes of generated screens to react against. Gluestack ui as the underlying component scaffolding kept designs build-ready from sprint 1.

DEFINE
AI DESIGN
FEEDBACK
UI DESIGN
FEEDBACK
HANDOVER
accelerating the design feedback-loop
velocity focused MVP handover

The Work

Two highlighted key feature areas:

  1. 1

    Rainfall Monitoring (desktop web) — turning raw data into a planning surface

  2. 2

    Paddock planning (mobile web) — map-first interaction for in-field decisions

Feature 1 - Rainfall Monitoring

Turning rainfall data into a planning surface, not a logbook

The problem

Farmers were tracking rainfall in spreadsheets, diaries and notebooks. Fine for records, useless for making actionable decisions. The platform needed to convert that data into something they could plan grazing rotations and feed demand against.

Constraint: No usability research budget, and a competing product racing to market. We had to commit to an interaction model from day-one and refine through the Product Owners SME knowledge, not user testing.

The solution

Used AI UI generation tools to rapidly visualise solutions and played them back to the PO for same-day approval. Led with a calendar-first view (familiar mental model from existing tools) and layered visualisation underneath. Calendar handles daily entry, chart handles trend reading.

1Concept design
readily.ai
readily.ai concept
v0.app
v0.app concept
replit.com
replit.com concept

What we rejected

Initial calendar view didn't scale when multiple rainfall sites were linked. Too much visual scanning for what should be at-a-glance. Pivoted to a linear table view that lets users compare site status from a single bird's-eye row.

Calendar design
2Calendar design
Table design (Final)
3Table design (Final)

Outcome

  • Stakeholder confidence

    Increased delivery velocity built confidence in both product direction and continued investment.

  • Design ahead of engineering

    Design consistently stayed ahead of engineering, reducing ambiguity and downstream rework.

  • Patterns / +1 designer effect

    Established patterns compounded design impact, enabling one designer to support more surface area.

Rainfall dashboard shared live in a stakeholder video call
Isometric montage of Agri-Tech screens across desktop, web, and native