SupplyShock Sentinel

SupplyShock Sentinel turns uploaded supply-chain data into risk scores, hidden supplier dependency maps, what-if simulations, and AI recovery actions before delays become failures.

Comment

Inspiration

Supply-chain disruption is usually discovered too late, especially by small and mid-sized businesses that still depend on Excel, email, WhatsApp, and manual follow-ups. A supplier delay, inventory shortage, quality hold, or hidden Tier-2 dependency can quietly become a customer delivery failure before the business has time to react.

The idea behind SupplyShock Sentinel came from one practical question:

What will fail first, why will it fail, and what should the business do immediately?

We wanted to build more than a static dashboard. We wanted a visitor-usable supply-chain risk platform where users can upload their own data, validate it, calculate risk scores, simulate disruptions, test what-if scenarios, and receive recovery recommendations before delays become failures.

The MeDo hackathon inspired this build because MeDo made it possible to move from an idea to a working full-stack web application quickly. Instead of spending most of the time on setup, routing, layout, and repetitive frontend/backend wiring, we could focus on the product logic, user journey, and business value.

What it does

SupplyShock Sentinel is an AI-powered supply-chain risk analysis and decision-support platform.

Users can either:

  • try the app with built-in demo data
  • upload their own supply-chain CSV files

The app supports uploads for:

  • Purchase Orders
  • Suppliers
  • Inventory
  • Supplier Relationships
  • Alerts

After upload, the app:

  • validates required CSV headers
  • explains required formats and allowed values
  • previews uploaded data
  • detects missing or invalid values
  • imports valid rows
  • calculates inventory days left
  • calculates purchase order risk scores
  • classifies risk as Low, Medium, High, or Critical
  • updates dashboards reactively
  • maps Tier-1, Tier-2, and Tier-3 supplier dependencies
  • generates alerts and recovery actions
  • allows users to edit data inside the website
  • supports what-if scenario testing
  • simulates random disruptions
  • provides AI-style recovery recommendations
  • exports risk-scored CSV files and reports

The core scoring model is explainable:

$$ Risk\ Score = Inventory\ Risk + Supplier\ Risk + Delay\ Risk + Priority\ Risk + Status\ Risk + DeepTier\ Risk $$

Risk levels are classified as:

Score Risk Level
0-30 Low
31-60 Medium
61-80 High
81-100 Critical

The strongest feature is that the app does not only show current risk. It lets users test future disruptions such as supplier reliability drops, inventory shortages, demand spikes, quality holds, and deep-tier supplier failures.

How we built it

We built SupplyShock Sentinel using MeDo, an AI-powered app-building platform that helped us create and iterate on a full-stack responsive web application through natural language prompts.

MeDo made the build easier in several ways:

  • It helped convert a problem idea into structured pages, workflows, and data models.
  • It accelerated dashboard, table, upload flow, and responsive UI creation.
  • It allowed rapid iteration when testing revealed usability gaps.
  • It helped us add complex features such as CSV validation, reactive dashboards, edit dialogs, AI-style recommendations, and what-if scenarios.
  • It reduced time spent on repetitive setup so we could focus on user experience and risk-analysis logic.

The build process was iterative:

  1. We first built the core supply-chain risk dashboard.
  2. Then we added deep-tier supplier dependency mapping.
  3. Then we added upload/download support.
  4. Then we improved the visitor journey so users could upload their own data.
  5. Then we added CSV format guides, validation, and template downloads.
  6. Then we added reactive editing.
  7. Then we added what-if scenario testing.
  8. Finally, we completed QA fixes for inventory editing, supplier editing, supplier relationship CRUD, and alert editing.

The final app includes:

  • Landing Page
  • Upload Your Data workflow
  • Executive Dashboard
  • Operational Dashboard
  • Risk Analytics
  • Purchase Orders
  • Inventory Intelligence
  • Supplier Risk
  • Deep-Tier Sentinel Map
  • Alert Center
  • AI Recovery Copilot
  • What-If Simulator
  • Data Management
  • Reports
  • Help Center
  • About Page

A key advantage of MeDo was speed of iteration. When testing showed that visitors could see a demo but could not clearly upload and analyze their own data, we quickly restructured the app around a better workflow:

Upload data → validate headers → preview rows → confirm import → run analysis → view dashboards → ask AI Copilot → download reports

That changed the app from a demo dashboard into a usable product experience.

Challenges we ran into

The biggest challenge was turning the app from a demo dashboard into a real visitor-facing product.

A dashboard alone was not enough. A first-time visitor needed to clearly understand:

  • where to upload data
  • what files are required
  • what CSV headers are accepted
  • what each column means
  • what values are allowed
  • how validation errors are shown
  • how imported data drives the dashboards
  • how to download results

To solve this, we added a complete Upload Your Data workflow with:

  • sample templates
  • header explanations
  • preview tables
  • validation results
  • import controls
  • analysis output
  • download options

Another challenge was making the app reactive. If a user changes supplier reliability, delay probability, inventory, demand, or supplier relationships, the app must immediately update risk scores, dashboard KPIs, charts, AI recommendations, and reports.

We also faced a data-format challenge. Some early upload validation expected internal field names such as name and productOrMaterial, while users expected business-friendly headers such as Supplier Name and Product or Material Supplied.

That taught us a direct product lesson:

Users should not be forced to understand internal database naming.

MeDo helped us handle these challenges quickly because we could describe the issue in plain language, request targeted changes, test the app again, and keep improving.

Accomplishments that we're proud of

We are proud that SupplyShock Sentinel became a complete, usable application rather than just a visual prototype.

Key accomplishments include:

  • [x] Complete CSV upload and validation workflow
  • [x] Downloadable sample templates
  • [x] Required header guides and examples
  • [x] Reactive dashboards
  • [x] Explainable risk scoring
  • [x] Editable purchase order, supplier, inventory, relationship, and alert data
  • [x] Deep-Tier Sentinel Map for hidden dependency visibility
  • [x] What-If Simulator with multiple disruption scenarios
  • [x] AI Recovery Copilot for action recommendations
  • [x] Report and CSV export capability
  • [x] Dark/light mode
  • [x] Responsive design for desktop, tablet, and mobile
  • [x] Help Center and About Page for first-time visitors

We are also proud of how quickly the app evolved using MeDo. The platform allowed us to go from concept to a multi-page, full-stack, interactive application much faster than a traditional manual build.

The feature we are most proud of is the combination of Deep-Tier Mapping and What-If Simulation.

A user can see not only that a purchase order is critical, but also why it is critical and which hidden upstream supplier may be causing the risk. Then they can test future scenarios before those risks become real failures.

What we learned

We learned that a useful AI product needs more than AI output. It needs a clear workflow, explainable logic, and user trust.

The biggest lesson was:

A user should be able to understand and use the app without external explanation.

That meant adding:

  • format guides
  • validation messages
  • templates
  • examples
  • help content
  • transparent risk scoring

We also learned that reactive behavior is essential for decision-support tools. When users edit data and immediately see dashboards change, the product feels real. It becomes more than a report; it becomes an interactive planning system.

MeDo also taught us how powerful natural-language-driven app development can be when paired with structured testing. The ease of use helped us build quickly, but quality came from repeatedly testing the app, identifying gaps, and using MeDo to improve specific workflows.

Finally, we learned that what-if testing is powerful for supply-chain risk management. Businesses do not only need to know what is wrong today. They need to ask:

  • What if this supplier delays?
  • What if demand doubles?
  • What if inventory drops?
  • What if a Tier-2 supplier fails?
  • What if we add an alternate supplier?

SupplyShock Sentinel was designed around those questions.

What's next for SupplyShock Sentinel

The next step is to make SupplyShock Sentinel more connected, intelligent, and production-ready.

Future improvements include:

  • ERP integration
  • live shipment tracking
  • logistics carrier APIs
  • weather and port disruption feeds
  • supplier news monitoring
  • automated supplier and customer email workflows
  • role-based access control
  • audit trails
  • multi-user collaboration
  • advanced machine learning models
  • scenario planning with saved playbooks
  • procurement workflow integration
  • accessibility and screen reader testing
  • security testing
  • browser compatibility testing
  • performance testing with larger datasets

We also want to continue using MeDo to iterate on the product quickly. Future versions could add onboarding wizards, guided data cleanup, scenario templates, and supplier communication automation.

The long-term vision is to turn SupplyShock Sentinel into an affordable supply-chain control tower for small and mid-sized businesses.

Instead of reacting after a delay happens, businesses should be able to see risk early, understand the hidden cause, test possible responses, and act before customers are affected.

Built With

  • medo
Share this project:

Updates