Problem Statement
After the Mindanao M7.8, DTI imposed a 60-day price freeze, yet monitoring teams can’t be in every sari-sari and market. Shoppers don’t know SRP by heart; receipts fade; violators claim “new stock.” Reports to Negosyo Centers die from incomplete details.
Proposed Solution
Snap tag + receipt: OCR reads item, price, store; app compares to SRP/freeze table and verdicts “OVER — ₱12 above SRP” with a ready DTI complaint draft (store, date, evidence photos). Works via SMS for non-smartphones. Market masters see a heatmap of repeat violators.
Target Market
- Primary users: Shoppers, nanays, students in quake/flood zones.
- B2G: DTI monitors, LGU price councils.
How AI Is Used
- Receipt OCR + normalization: Handles faded thermal prints and handwriting.
- Complaint drafter: LLM writes complete Tagalog complaints so reports don’t get rejected.
PLDT / Smart / ePLDT / Rokid Angle
- Smart: SMS reporting for zero-load shoppers.
- ePLDT: Hosts SRP tables + evidence vault.
- Rokid: Inspectors see SRP overlay in markets (stretch).
Demo Scenario (3 minutes for judges)
- Snap mock canned-goods tag ₱35 vs SRP ₱27; verdict OVER appears.
- Complaint draft generates with store, date, photos.
- Show compliant item for contrast: “TAMA — within SRP.”
- SMS query returns same verdict on a basic phone.
- Close: “Picturan, isumbong, may resibo.”
30-Hour Build Plan
| Hours | Task | Role |
|---|---|---|
| 0–3 | Mock SRP tables + receipts | All |
| 3–10 | OCR + verdict logic + drafter | AI lead |
| 3–10 | SMS + heatmap | Backend |
| 10–20 | Shopper UI | Frontend |
| 20–26 | Faded-receipt testing | All |
| 26–30 | Pitch, disclosure | PM / Pitch |
Suggested stack: Flutter + on-device OCR, DTI SRP table, LLM complaint drafter, SMS report fallback. Public data / APIs: DTI SRP bulletins, Price Act RA 7581 guides.
Risks & Privacy Notes
- False accusation: Verdicts advisory with SRP citation; complaints go through DTI process. Demo stores fictional; no real store data into AI tools.
- Privacy: Store names only with reporter consent; buyer info minimized.
AI Use Disclosure (draft)
Uses OCR plus LLM complaint drafting. Coding assistants helped with camera code. All demo verdicts human-verified vs mock SRP.
Rubric Scoring
| Criterion | Weight | Score (1–5) | Why |
|---|---|---|---|
| Innovation & creativity | 25% | 4 | Receipt-to-complaint in seconds is practical |
| Problem–solution fit | 25% | 4 | Post-disaster overpricing is rampant |
| AI application & value | 20% | 3 | OCR + drafting, modest AI |
| Feasibility | 15% | 5 | Very buildable |
| Impact & scalability | 15% | 4 | Nationwide DTI use |
| Weighted total | 4.20 |