All work

BugPilot — AI Bug Reporting Assistant

Reporting POS issues was slow and lossy. BugPilot lets anyone — client, QA, ops or developer — capture an issue in one tap; AI converts the recording into a structured, developer-ready bug report with the technical context attached. The concept placed Top 5 of 26 teams at Foodhub's AIXCELERATE Hackathon 3.0.

Category
AIXCELERATE Hackathon 3.0 · Top 5 of 26 teams
Role
Product concept, problem framing & pitch
When
Foodhub Hackathon · 2025
AIDeveloper toolingIssue reportingWorkflow automation
Problem

Reporting POS issues meant calls, messages, screenshots and back-and-forth. Details got lost and developers spent extra time reproducing the issue — delaying fixes.

Approach

One tap captures screen, voice and user actions; AI converts it into a clear problem statement with steps, expected vs actual and auto-attached logs/API data — instantly creating a developer-ready ticket.

Impact

Bug reporting shifts from 'what is the issue?' to 'here is the issue, why it happened, and everything needed to fix it' — Top 5 of 26 teams.

My Role

Product concept, problem framing & pitch

Users

Clients, QA, operations and developers reporting POS issues

Impact

Top 5 of 26 teams — a faster, lossless path from issue to developer-ready ticket

Context

In a POS platform used across thousands of stores, issues surface from people who aren't technical — store staff, ops, support — and land on developers who need precise, reproducible detail. BugPilot was a hackathon concept aimed squarely at that gap.

Problem / Opportunity

Reporting an issue meant a chain of calls, messages and screenshots. Context leaked at every hop, and developers burned time just reconstructing what happened before they could fix anything.

User & Business Need

  • Non-technical reporters needed a way to flag issues without writing a technical report
  • QA and ops needed less manual documentation and fewer follow-up rounds
  • Developers needed complete, reproducible context to cut investigation time

My Role

  • Framed the problem and the target user segments the concept had to serve
  • Shaped the end-to-end flow from one-tap capture to a developer-ready ticket
  • Defined what 'complete context' means — steps, expected vs actual, logs, API data, device details
  • Pitched the concept to judges, where it placed Top 5 of 26 teams

Product & Delivery Approach

  • Capture: one tap records screen activity, a voice explanation and the user's actions
  • AI understanding: generate a structured report — problem, steps to reproduce, expected vs actual, impact
  • Auto technical data: attach application logs, the click journey, API requests/responses and device details
  • Instant ticket: the finished report becomes a development ticket automatically

Constraints & Trade-offs

  • AI-generated reports must be trustworthy — the design leaned on captured evidence, not guesswork
  • Capturing logs and API data raises real privacy and scope questions to handle deliberately
  • The reporting flow had to stay one-tap simple for non-technical users while still satisfying developers

Outcome & Impact

  • A concept that reframes reporting from 'describe the problem' to 'here's everything to fix it'
  • Clear value mapped to every stakeholder — clients, QA/ops and developers
  • Validated by judges with a Top-5 finish out of 26 teams

Supporting documents

BugPilot — Demo Download
BugPilot — Screenshot Save
BugPilot — Concept Write-up
PDF

Key Learnings

  • A reporting tool's real product is the quality of context it hands to the next person in the chain.
  • Mapping value to each stakeholder separately made the pitch land — one feature, three distinct wins.
  • AI is most convincing when it's grounded in captured evidence rather than asked to invent detail.