Vault — Multi-Role Project Delivery Cockpit & AI Executive Assistant

Published on July 21, 2026 • 12 min read • Category: Strategic Product Management & AI Ops • By Nikhil Udgata

Modern software engineering organizations frequently suffer from costly operational friction. Enterprise clients submit ad-hoc feature requests; Technical Program Managers (TPMs) manually calculate capacity in static spreadsheets; Account Engineers (AEs) fill out isolated standup forms; and C-Suite leadership lacks a single source of truth for portfolio health.

To solve this enterprise delivery bottleneck, I architected Vault — a multi-role project management cockpit and AI executive assistant designed to streamline engineering execution, optimize workforce capacity, and eliminate administrative overhead across four critical business functions.

🎯 Executive Summary & Business Impact

Vault replaces fragmented spreadsheets and manual reporting cycles with an automated, real-time operational workspace:

| Business Metric | Before Vault | With Vault Cockpit | Strategic Impact | | :--- | :--- | :--- | :--- | | Backlog Prioritization | Subjective stakeholder debates | Mathematical RICE Scoring | Objective, ROI-driven feature roadmapping | | Capacity Allocation | Manual spreadsheet planning | Visual Drag-and-Drop + Overload Alerts | Zero developer burnout & 100% capacity balance | | Status Reporting | 4–6 hrs/week manual compilations | Automated SSE & Jira Sync | Instant cross-team operational transparency | | AI Insights Cost | Expensive per-query LLM calls | Redis 4-Hour Cached Summaries | >90% reduction in AI API token costs |

🗺️ Enterprise Delivery Journey & 4-Role Value Loop

Vault aligns four key organizational stakeholders into a continuous, real-time delivery lifecycle:

[1. Client Portal: RICE Intake] ──> [2. TPM Workbench: Capacity] ──> [3. AE Workspace: Execution] ──> [4. Executive Cockpit: AI RAG]
         ▲                                                                                                    │
         └────────────────────────── Real-Time Redis SSE & Bi-Directional Jira Webhook Loop ──────────────────┘

Figure 1: High-level architectural flowchart of Vault showing the real-time event pipeline and Jira integration.

Enterprise Technology Stack:

• Frontend: React SPA, Vite, Tailwind CSS (V3), custom SVG Gantt charting engine, real-time Server-Sent Events (SSE) state hooks.
• Backend: Node.js, Express, cursor-paginated Jira Cloud REST API integrations, Redis Pub/Sub broker, PostgreSQL database with dynamic SQL schema translation.
• AI Orchestration: Dify AI orchestration stack (wrapping Ollama / Llama 3), implementing the lean-prompt pattern with a 4-hour Redis TTL response cache to minimize token costs.

💼 Business Use Cases by Stakeholder Role

1. Client Portal — Objective RICE Intake & Milestone Visibility

• Business Challenge: Clients struggle to track feature delivery timelines and often push unquantified requests that derail product roadmaps.
• Vault Solution: External client partners gain a dedicated dashboard with real-time milestone progress. New feature requests are evaluated using the mathematical RICE Prioritization Model:

RICE Score = (Reach × Impact × Confidence) ÷ Effort

• Value Delivered: Data-driven backlog grooming that replaces subjective negotiation with objective financial and operational scoring.

Figure 2: Client Portal dashboard showing interactive RICE scoring sliders and milestone execution progress.

2. TPM Workbench — Resource Utilization & Burnout Prevention

• Business Challenge: Over-allocating software engineers leads to burnout, delayed releases, and expensive context switching.
• Vault Solution: TPMs schedule sprints using an interactive calendar where Jira issues are dragged directly onto engineer timelines. The workspace automatically computes daily work hours and triggers Overload Safety Badges if an engineer's allocation exceeds 8 hours a day.
• Value Delivered: Maximizes engineering throughput while preventing burnout and ensuring predictable delivery dates.

Figure 3: TPM Capacity board illustrating calendar allocations, engineer utilization rates, and overload warnings.

3. AE Workspace — Frictionless Compliance & Automated Jira Sync

• Business Challenge: Engineers dislike spending valuable coding time filling out duplicate status reports across multiple tools.
• Vault Solution: Account Engineers (AEs) land on a clean daily view of assigned work items. Completing EOD standups or status changes in Vault automatically triggers bi-directional Jira Cloud API updates, posting comments directly back to the Jira issue history.
• Value Delivered: Saves 2+ hours of administrative work per engineer per week while maintaining a 100% compliant Jira audit log.

Figure 4: Account Engineer inbox featuring daily task lists, status toggles, and automated Jira comment syncing.

4. Admin Cockpit — Executive Portfolio RAG & Cost-Effective AI Insights

• Business Challenge: Leadership needs portfolio oversight without paying thousands of dollars in LLM API token fees for continuous status summaries.
• Vault Solution: The Executive Cockpit aggregates cross-project metrics into portfolio heatmaps with manual RAG (Red-Amber-Green) status overrides. AI summaries utilize a lean-prompt pattern cached in Redis with a 4-hour TTL.
• Value Delivered: Provides C-Suite executives with real-time risk visibility while cutting AI operational costs by over 90%.

Figure 5: Executive Admin Cockpit rendering cross-project health matrices and AI summary widgets.

⚡ Core Operational Engines & Real-Time Sync

Server-Sent Events (SSE) Real-Time Architecture

Rather than forcing manual browser reloads or maintaining heavy WebSocket connections, Vault implements a Server-Sent Events (SSE) pipeline:

1. When a TPM assigns a ticket, Express writes to PostgreSQL and publishes an event to the pm:events Redis channel.
2. The SSE handler broadcasts the payload to all active client portals and dashboards.
3. The frontend usePmEvents() hook intercepts the payload, debounces by 300ms, and updates the UI instantly without page reloads.

Bi-Directional Jira Cloud Webhook Loop

1. When an external update occurs in Jira, a webhook is delivered to POST /api/jira/webhook.
2. Vault updates internal project states and alerts the assigned engineer's inbox.
3. Engineer explanations submitted in Vault are automatically pushed back to Jira via jira.addComment().

🌐 Enterprise Deployment & Cost-Optimization Strategy

Vault was engineered to demonstrate enterprise-grade capabilities at zero cloud infrastructure overhead:

• Multi-Stage Containerization: Built with Docker to package the React frontend and Node.js backend into a lightweight Alpine container.
• ₹0 Production Cloud Stack: Hosted using Vercel (Frontend SPA), Render (Express Backend API), Render Postgres (PostgreSQL Database), and Upstash (Redis Broker).

Vault demonstrates how unifying structured intake, intelligent resource capacity planning, and automated event synchronization creates an indispensable delivery cockpit for modern tech enterprises.