11 · Strategy Narrative: how to explain this to someone
warning · It is a speaking spine: the order in which to tell the story, with each beat pointing to your exact words in the other files. Every quoted phrase is yours. Anything in plain text between quotes is a signpost, not a claim. Where I list a choice you have not settled, it is marked [Your call].
The 90-second version (all in your words)
- The problem, as the organisers framed it. "underneath what sits is a problem of trust and identity — which is currently addressed by a six-digit code." The weak answer "will simply swap the OTP for an agentic handshake and relocate the friction." → 01 Problem Statement
- Who feels it. "buyer is not the receiver": friends, "mom and dad, who are not that well-versed in technology", security guards, maids. "Usually, the buyer is pretty busy during this time." → Map A in 03 Whimsicals Mind Maps; respondent quotes in 06 Your Thinking Log
- Our insight (the one we discovered). "Delivery riders are the forgotten stakeholder. OTP is not just customer friction, but a labor problem." "five rupee penalty for missed orders incentivizes marking customers unavailable and moving on." "location data fails 10 to 15 percent of the time." → 04 Submitted Answers Q2; Map C
- What we built from it. Round 2 design: a household-bound QR badge, geofence, voice priority list, payment release (the "Delivery Concierge Agent"). Round 3 design: a voice agent that calls ahead, a logistic agent that issues a QR per waybill and reroutes, a payment agent for agentic COD. → 04 Submitted Answers, 08 Final Round Written Answers
- Why this company. Round 2 answer: Delhivery "should have built this" (low intent, no churn, public-company incentives, fear of breaking a critical component). Round 3 answer: Flipkart through Ekart, because it "is the only indian company sitting on all three": ekart, super.money, SLAP/ring ai. → 04 Submitted Answers Q9, 08 Final Round Written Answers
- How it makes money and gets its first 1,000 users. Saves RTO (15 to 20% on COD), more stops per rider, "verified delivery" tier for D2C brands, licence the stack ("aws model"). Seed one dense hub, default-on, invite-only agentic COD. → 08 Final Round Written Answers, 09 Final Round Whimsical Transcription 1.5 and 1.7
- What could kill it. Regulation (RBI extra factor for agentic COD), quiet false negatives, internal politics, token cost. → 08 Final Round Written Answers
The strategy as a chain of reasoning (each link is yours)
| Step | The claim | Your words / where it is |
|---|---|---|
| First principles | The OTP is a "metric for finished delivery to the right person" and "prevents against fraud" | Map A |
| Constraint you set | "i dont want the user to do anything. he should not do any work." | 06 Your Thinking Log §14 |
| Why not live tracking | A rider does multiple pickups and deliveries, "nothing is solved and proper eta can not be calculated", plus "also privacy issue" and "tracking live in a huge infrastructure cost" | 09 Final Round Whimsical Transcription 2.2, 2.4 |
| ETA trigger instead | Trigger on the previous stop; or "t-10 order … 10*10 which is 100mins" | 2.4 |
| Why not rider call as the trigger | "usually these delivery people don't even call, and sometimes they fake the call." | 06 Your Thinking Log §9 |
| Verification | One QR per waybill ("namma metro" model), sent as a picture, "pure and black" Level H for cracked screens; GPS only for ETA | 2.2 |
| Fraud / low trust | "India is a low trust society"; leave-at-door gets "a picture of the parcel at the door"; return gets video or "picture of the surrounding" | 06 Your Thinking Log §5, 05 Voice First Solution |
| Voice's job | Call before, read sentiment, decode "low connectivity and noise", three states: available, not available, leave at the doorstep | 05 Voice First Solution, 2.8 |
| Guardrail | Confidence score: below 80% the rider calls manually; 95 to 100% "not available" reroutes the stack | 2.9 |
| Payments' job | "release role, not a collection role" in Round 2; in Round 3, UPI Reserve Pay hold, captured on handoff | 04 Submitted Answers Q4, 2.10 |
| Fourth rail | Round 2 note: consumer AI / personal assistant that knows busy, location, trusted neighbours | 2.5 |
| Annexation | Domestic and gig workers (cooks, maids, house helps); "bound, revocable household badge" | 04 Submitted Answers Q7 |
Second-order thinking you already did (the team's stated edge)
- Always-on GPS rejected → ETA by sequence of stops.
- Rider-trigger rejected → system-trigger.
- User effort rejected → passive evidence (photo/video through the app, sent to buyer on WhatsApp, not saved on the rider's phone).
- A false "not available" "pushes rto up instead of down" → confidence gate and manual call fallback.
- A QR that "completely breaks" when nobody is home → drop-off with photo; alternates registered the same morning.
Questions you left open (do not hide these in the pitch; they are the honest edges)
- "what makes you think we will have the history of who" picks up? (nobody logs it)
- The user is "in a very important meeting" and there are no alternative numbers.
- "wound't that invade the privacy" for video/photo; is "not saved on the delivery agents phone … sent to the buyers' phone over WhatsApp" enough?
- "is that enough?????" for a 15 to 20 minute prior update.
- Whether COD OTP is needed ("yes if the aov is high … or open box delivery"), whether NDR carries reasons ("Ndr needs actual reasons"): these are answered on your board; confirm with an operator.
Back to 00 README Index