Appearance

car_sales — vertical discovery brief ​

THIS IS NOT A COMPLETED DISCOVERY. IT IS THE AGENDA FOR ONE.

Nobody has been spoken to. This brief was assembled from desk research on 2026-08-22 and the template's own rule is explicit: "⚠ Inferred from market research is not discovery."

It is filed at status 🔴 not started on purpose, so the G-D gate (QRS-475) refuses any release scope item declaring "vertical": "car_sales" until a real session happens. The market work that is done lives on the five companion pages linked from the index; this page records only what a conversation must still produce, and what desk research already narrowed for it.

Same posture as real_estate, and for the same reason.

0 · Front matter ​

FieldValue
Industry keycar_sales — 🧮 seeded in industries (20260808120000_v2_taxonomy.sql line 281)
Archetypeexpertise — 🧮 as seeded. The card generates the enquiry, never the sale
Primitive composition🧮 ['catalogue','party','schedule','location','resource','asset','campaign'] — 7 of 11, and 5 of the 7 have zero tables
User categories served2 enterprise (primary: the dealer group and its employees) · 1 solo owner (the seeded label's "Sales Agent" case) · 3 consumer (the buyer receiving reminders, anonymous-first)
Release❓ None. Not in any release scope, and must not enter one until this brief is complete
Status🔴 not started
EvidenceNobody. Zero interviews, zero franchise agreements read, zero realised contract values. Desk research only, 2026-08-22

The seeded label and the strategy target are not the same business, and this needs an owner decision

🧮 The industries row reads 'car_sales', 'Car / Bike Sales Agent' — an individual agent, sized like real_estate — while its own description calls it "The flagship Enterprise case (ADR-0022/0023/0024)" and the commercial model targets multi-outlet dealer groups.

🔎 Those are two different merchants with different primitive needs, different price ceilings and different buying units. One row cannot serve both without the archetype layer branching, which 📘 ADR-0021 D4 forbids.

Decision needed before design: either the row is retitled to the dealership case and the solo agent becomes a separate industry, or the reverse. Recorded as QRS-838.

1 · The business, in their words ​

Every question below is unanswered. What desk research established is listed as context for the question, never as the answer.

  1. Who is the merchant? ❓ Who signs, and at what threshold? 🔎 Desk research cannot distinguish 200 signatures from ~100 (outlets) from ~70 (groups), and 🌐 FADA's own ratio is 2:1 outlets to dealerships. Three completely different businesses.
  2. How do they get customers today? 🌐 Context: FADA/NRI puts 49% of new-car leads online, and 🌐 Meta/Kantar reports 48% of new buyers contacting dealerships via WhatsApp. ❓ But the split between OEM-supplied leads, portal leads (CarDekho, CarWale) and walk-ins is unmeasured, and it decides whether this vertical's card has any acquisition role at all.
  3. Vendor pain points, ranked, in their vocabulary. 🌐 Desk research produced a ranked list from audited filings and an association white paper — see operational-gaps. ❓ It is not their vocabulary and it is not their ranking.
  4. Customer pain points. 🌐 Context: 47% of buyers want service reminders by messaging; Indian consumer-rights documentation records the oral-approval pattern (a ₹9,800 estimate billed at ₹27,640).
  5. Existing workflows and tools. 🌐 Context: the OEM's mandated system, plus a separately purchased back-office DMS, plus — in an Indian DMS vendor's own words about its own customers — "sticky notes and notebooks". ❓ Which specific products, at which specific dealer, at what price is the question E2 and E1 answer together.
  6. What information do they already share with customers, and how? ❓ Unasked. 🔎 The template calls this the highest-signal question in the brief and it is completely open here.
  7. Industry-specific interactions. 🔎 Candidate verbs from desk research: test drive · exchange valuation · finance approval · insurance attach · PDI · delivery ceremony · service due · renewal. ❓ Not confirmed with anyone, and the delivery ceremony in particular may be a card moment nobody has considered.
  8. Edge cases they raised themselves. ❓ Unasked. 🔎 Known-likely: month-end registration push, Navratri-Diwali convergence (🌐 FADA records October 2026 as a 40-lakh record month on exactly that), and multi-brand groups running one brand's season against another's.

2 · Content types ​

🔎 Inferred, and every row needs confirming in the session.

Content typeNeeded?Shape / notes
Photographs🔎 likelyStock imagery is OEM-supplied and brand-controlled, so the dealer's own photos matter mainly for used stock
Video🔎 likely, high valueWalkaround for used stock; delivery video as the handover artefact
Before / after🔎 noNot a transformation business
Documents🔎 yesQuotation, finance sanction, insurance policy, RC, PDI checklist, job card
A published number that changes🔎 yes, and it is an ADR-0027 caseEx-showroom price, on-road price, exchange offer, waiting period. ❓ Which of D2/D4/D5 applies is undecided
A schedule🔎 yesTest-drive slots, service bays, advisor availability — the schedule and resource primitives, neither built
Structured specs🔎 yesVariant, fuel, transmission, colour, km, ownership. ❓ Which are filtered or priced upon (typed columns) versus merely displayed (attributes) is undecided and matters
Long-form text🔎 minor

⚠ The ADR-0019 D6 question, answered honestly: 🔎 nothing here obviously needs a new block type. A vehicle listing is a catalogue item with typed attributes; a test-drive request is an enquiry. The asset primitive is the genuinely new thing, and it is already approved — 📘 ADR-0024 D3 passes the ≥3-industry gate with the vehicle, the installed AC unit, the device and the flat.

3 · Acquisition and conversion ​

  1. ⚠ COLD DISCOVERY, asked first. ❓ Unanswered, and it is the question that decides whether this vertical belongs on the platform at all. 🔎 A car buyer almost certainly starts at a portal (CarDekho, CarWale, Cars24) or the OEM's own site, not at a vendor's card. If that is right, the Setu Card has no cold-discovery role in this vertical and the entire value sits in operations — which is a finding, not a failure, and it would make car_sales the first vertical whose card is an internal tool rather than a growth surface. 📘 This is exactly the omission QRS-490 records against the festival-stall brief; do not repeat it.
  2. Why would this merchant share their card? 🔎 Weak. A dealership's brand is the OEM's, not its own. The honest candidate is the individual sales consultant's card, which 📘 ADR-0022's per-agent workspace model already supports and which is also the thing 🌐 29.53% attrition takes away.
  3. What makes their customer forward it? 🔎 Candidate: a delivery video or a service record. ❓ Untested.
  4. Where does the card replace something manual? 🔎 Candidate: the quotation and the service-due call.
  5. Natural referral path? 🔎 Yes, and it is unusually strong: 🌐 twelve groups hold sixty Maruti outlets in Pune, so this is a small, well-connected community where one visible success travels.
  6. Acquisition CTA placement. 📘 Default stands: post-action confirmation and share sheet, never hero.

4 · Monetization ​

  1. What would they pay? ❓ The central unknown. 🌐 Every Indian price found is a list price, an IndiaMART ask, or a directory artefact. Zero realised ACVs anywhere. See commercial-model.
  2. Which capability is the paywall? 🔎 Proposed: feature_code for cross-brand consolidation (the group rollup), because it is the one capability no OEM system can deliver. ❓ Not confirmed with a buyer.
  3. Transaction revenue? 🔎 No, and do not design for it. A vehicle sale settles through the OEM's own financing rails; 🌐 Maruti Smart Finance alone covers over 40% of that OEM's customers.
  4. Is there a second buyer? 🌐 Yes, and it is the largest single fact in the commercial analysis: an OEM. Every deployment reaching 200+ Indian dealerships was top-down. ⚠ But see go-to-market — selling to the OEM forecloses the dealer-trust position, and the two are mutually exclusive.
  5. ⚠ Store-compliance check. 📘 Any upsell converts on web or email, never in-app (ADR-0002).

5 · Governance and verification ​

  1. Legally gated capabilities? 🌐 Yes, three, and all three are hard walls. Motor insurance solicitation is restricted to insurer-appointed MISPs, who may not solicit from customers who did not buy the vehicle from them. Digital lending falls under RBI's Digital Lending Directions 2025. And ❓ whether pooled cross-dealer benchmarking makes QRSETU a Data Fiduciary rather than a processor needs counsel, not research.
  2. Evidence accepted per industries.verification_requirements? ❓ Undecided. 🔎 A dealership is a GST-registered entity with an OEM letter of intent, so verification is cheap here compared with a stall vendor.
  3. A claim we would be the publisher of? 🔎 Yes: price. An on-road price or an exchange offer rendered on a public card is a representation about a regulated, OEM-influenced number. 🌐 Note the CCI record on OEM discount control before designing any published-discount surface.
  4. Needs a legal read before shipping? 🔎 Yes. The DPDP processor contract term is mandatory under the Rules, and the Data-Fiduciary question in item 1 is unresolved.

6 · Divergence seams ​

Required at G0. 🔎 Candidates in this vertical, each needing an implementation or approved fallback per surface (Android native · iOS native · Web PWA) before work starts:

SeamWhy this vertical touches it
CameraUsed-stock photography, delivery video, document capture, odometer photo
PushService-due and test-drive reminders to staff
Storage / offlineA sales consultant on a showroom floor with poor signal
Deep linksA reminder opening the right vehicle record
ShareThe consultant's card, the quotation, the delivery video

🔎 None is novel for this platform, and 📘 check:parity rule R9 already single-sources expo-camera.

7 · Proactive-value answer ​

  • What action does it prompt? 🔎 "These 14 vehicles are service-due this week and nobody has called them." That is the strongest available answer, and it sits directly on the 🌐 41.2%-gross-margin after-sales line.
  • What makes it timely? 🔎 A recurrence per asset, on time or odometer — 📘 the existing engine (QRS-380) with an odometer condition, per ADR-0024 D3.
  • Where does the intelligence come from? 📘 The recurrence derivation in @qrsetu/domain plus the analytics read model. Not a hardcoded nudge — a per-vertical nudge is a defect against ADR-0009.

8 · Part 2 — capability mapping ​

Cannot be completed before the session, and filling it from desk research would defeat its purpose

The template's rule is that anything generalising to one industry only is either genuinely bespoke or mis-modelled. 🔎 Desk research suggests the mapping will be unusually clean here, because 📘 ADR-0024 already did most of it from dealership workflows.

Workflow (their words)Existing capabilityNew capability neededGeneralises to ≥2 others?
❓ awaiting session

🔎 Known in advance from ADR-0024's own P1-P8 analysis: five of the eight dealership pains it lists are already expressible in the model, and the three that were not produced asset, oversight and targets. So the expected outcome of this mapping is few or no new primitives — which is the primitive gate working rather than being bypassed.

9 · Sign-off ​

Session held❌ Not held. E1 in go-to-market is the plan: twelve dealer principals, two weeks
Capability mapping complete❌
New primitives proposed🔎 None expected. asset is already approved by ADR-0024 D3
New block types proposed🔎 None expected
Open decisions for the owner(1) Is car_sales the agent or the dealership? QRS-838 · (2) Dealer-owned positioning or OEM channel? They are mutually exclusive · (3) Does this vertical proceed at all before the pilot returns?
Tracker rows raisedQRS-835 · QRS-836 · QRS-837 · QRS-838 · QRS-839