TaskMentor.ai — AI-Powered Enterprise Task Decomposition & Onboarding Platform
Target User: Enterprise HR managers, project leads, legal/tech department heads, and newly onboarded corporate employees. The Problem: Corporate onboarding and enterprise workflow execution suffer from vague task descriptions, long adaptation periods, and task hoarding. Traditional PM software (Jira, Trello) tracks progress but does not assist employees in "how" to execute complex, multi-step B2B assignments (e.g., legal compliance, technical architecture reviews). The Solution & Outcome: TaskMentor.ai bridges this gap using Google Gemini API: 1. AI Task Decomposition: Automatically breaks vague enterprise titles into 3 actionable micro-tasks with estimated duration, difficulty, and XP gamification rewards. 2. AI Vision Proof Audit: Validates uploaded proof-of-work documents and screenshots using Gemini Vision OCR before marking tasks complete. 3. Anti-Hoarding Policy: Dynamically locks tasks to prevent resource bottlenecking. Outcome: Onboarding adaptation time is reduced by 70%, management overhead is minimized, and compliance is verified in real-time.
Preliminary score
- Value for the user15.0 / 25
A specific problem and a clear outcome from solving it.
- Prototype and use of AI14.5 / 30
A working core scenario, and what AI actually contributes to the result.
- Quality testing9.5 / 20
Tests, examples of failures, and comparison with the current approach.
- Feasibility5.0 / 15
Data requirements, running costs and a clear next step.
- Originality5.0 / 10
A meaningful difference from common solutions.
What the judges wrote
- GPT-6 AstraPreliminary52/100
- Strengths
- Your project addresses a real gap between assigning an enterprise task and helping a new employee execute it. Your proposed combination of task decomposition, proof-of-work checks, and anti-hoarding rules goes beyond a generic onboarding chatbot. Your testing description identifies specific failures—rate limits, vague inputs, and unreadable uploads—and explains the mitigations you attempted.
- To improve
- Show one complete assignment-to-proof-review workflow with actual model outputs and task-state changes; the supplied materials do not demonstrate that core loop. Validate your 70% onboarding reduction and sub-two-second breakdown claims with measured baselines, and report test counts and false approvals for proof uploads. Add human review for uncertain proof checks rather than treating document recognition as compliance verification. Move the API key out of the browser bundle, explain confidential-document handling, and provide per-task costs and a concrete pilot plan.
- Claude Opus 5.5Preliminary46/100
- Strengths
- Your problem framing is sharp: Jira and Trello track progress but do not help a new employee work out how to execute a vague, multi-step assignment. Your core loop of AI decomposition into micro-tasks, followed by Gemini Vision checking proof-of-work before completion, gives AI a real job in the workflow. Your testing write-up names concrete failure modes: 429 rate limits, overly broad output from single-word inputs like "Contract", and blurry screenshots that failed OCR. It also covered Azerbaijani as well as English inputs.
- To improve
- Your 70% onboarding-time reduction and 20–30 minutes saved per task are stated without any measurement. Even a small timed comparison against a manager doing the same breakdowns, or a rubric rating subtask quality, would make these claims credible. Report the vision audit as numbers, for example how many valid, partial and blurred proofs were correctly accepted or rejected, including false accepts. Your README is still the default AI Studio template and the Gemini key ships in the client bundle. Add a backend proxy, a cost-per-task estimate and a concrete pilot plan with one company department to strengthen feasibility.
The team
Mirsadig Javad
developer
Elvin MasimovLead
developer