mondaiy — AI Work Navigator
Evidence-backed personal work assistant for fragmented work sources
Built and led delivery of mondaiy (P-004), an AI work navigator that brings together user-permitted messages, email, calendar events, tasks, and long-term memory. The product creates evidence-backed daily briefings, answers questions with source links, extracts actionable work, and prepares actions for user approval.
Timeline
N/A
Type
Project
Status
ongoing
My work
- •Owned product definition, roadmap, source-of-truth documentation, and delivery coordination across a four-person team
- •Authored the PRD, user stories, acceptance criteria, and product safety boundaries for provenance, source conflict handling, and approval-gated actions
- •Integrated and reviewed vertical slices spanning FastAPI, LangGraph, React, browser extension, provider connectors, and PostgreSQL/pgvector retrieval
Outcome / Impact
- •Delivered an active MVP/pilot product that consolidates permitted Gmail, Google Calendar, Discord, Jira, and Outlook sources
- •Implemented evidence-backed daily briefs, source-linked Q&A, actionable-work extraction, and user-approved task or calendar actions
- •Established product guardrails: source provenance and version history, explicit conflict handling, cross-user isolation, and no external action before approval
Tech / Skills
Case Study
1) Context / Problem
Work information is fragmented across messages, email, calendar, and task systems. People lose time finding the latest authoritative source and need a safe way to turn information into action.
2) Your Role
As Product Manager and Product & Delivery Owner, I owned the product definition, roadmap, source-of-truth documentation, acceptance criteria, and cross-functional delivery coordination.
3) Approach
Designed the product around user-permitted sources, persisted source evidence and version history, and scoped LangGraph answers to connected data. The delivery combines FastAPI services, React web UI, browser extension, provider connectors, PostgreSQL/pgvector retrieval, and user approval for external side effects.
4) Result / Impact
Delivered an active MVP/pilot with source-linked daily briefings, grounded Q&A, actionable-work extraction, and approval-gated task or calendar workflows. The live demo is available at mondaiy.io.vn/briefing.
5) Learnings
Reliable personal AI needs product boundaries as much as model quality: preserve evidence, represent conflicts, isolate user data, and keep the human in control of external actions.
6) Links
See links above.