interview hypothesis
The interview story connects category insight, customer evidence, Shiprocket economics, and a concrete operating cadence while keeping public-source unknowns explicit.
Rocketizer should become the permissioned AI operations team for connected-storefront Shiprocket Power Merchants: turn commerce exceptions into verified outcomes—not the longest menu of AI employees.
Verify the supported connected storefront, production-safe read path, candidate action rail, context authorization, and metrics under Gate G0; run the roadmap's COD-heavy merchant discovery; then use Gate G1 to continue the hero, pivot to ranked remittance recovery, or stop. The roadmap gate definitions are canonical.
Build the versioned order-risk-contact-payment-shipment-action contract, retrospective baseline, eligibility ledger, and read-only COD-to-delivered recommendation; freeze the roadmap-required evaluation set; let Gate G2 decide whether shadow quality earns live-action work.
Pass Gate G3 on policy, permissions, approval, trace, safe stop, and handoff; implement one lowest-risk COD/recovery action with idempotency and terminal reconciliation; pass Gate G4. This proves controlled execution readiness, not repeatable ROI or first-NDR writeback.
For connected-storefront Shiprocket Power Merchants on a storefront path verified at G0, place only high-risk COD orders into a confirmation queue, convert willing buyers to prepaid, and resolve the first NDR with a logged reattempt/cancel action and delivered-order outcome.
Commit risk queue, approval-gated confirmation/prepaid intent, delivered-order attribution and one verified action on the G0-supported storefront path. Add first-NDR writeback only after the first action, safety and merchant-outcome gates pass.
| Time | Move | Claim type | What to say |
|---|---|---|---|
| 0:00–1:15 | Start with product truth | public fact | Rocketizer has a deployed app, signup, demo, and responding API health surface. The app exposes approvals, artifacts, memory proposals, roles, team invites, and connectors. The demo explicitly uses fixtures and disconnected/manual paths, so these surfaces do not prove production writes, adoption, or outcomes. |
| 1:15–2:15 | Explain why Shiprocket can win | analyst inference | The possible advantage is distribution plus potentially differentiated, merchant-authorized commerce context near action—not the model. Public sources establish Shiprocket scale and action surfaces, not Rocketizer's field-level access, merchant consent, permitted use, retention, cross-merchant rights, or incremental context lift. |
| 2:15–3:15 | Place the category | analyst inference | Horizontal agents, conversational commerce, checkout, logistics, returns, commerce operations, and finance specialists make tools and workflows abundant. The sharper position is a policy-bounded merchant P&L operator that closes a merchant-authorized economic loop and proves whether the context improves outcomes. |
| 3:15–4:10 | Name the initial customer | analyst inference | Start with a connected-storefront Indian Power Merchant that has material COD, recurring high-risk and first-NDR events, a multi-person operations workflow, joinable order/payment/shipment data, and willingness to support a holdout. Sprint 1 selects and verifies the supported storefront, data path, and action path; no storefront is assumed production-ready. |
| 4:10–5:10 | Use cases as a measurement lesson | analyst inference | Narrow workflows create clearer economic narratives, but even strong-looking official cases often omit baselines, windows, formulas, or counterfactuals. Customer-owned corroboration raises confidence without eliminating attribution questions. Rocketizer should make the proof ledger part of the product. |
| 5:10–6:25 | Propose the COD-to-delivered-revenue loop | proposed action | Lead with the ranked COD-to-delivered-revenue control loop. Commit a high-risk queue, approval-gated confirmation or prepaid intent, delivered-order attribution, and one verified action. Add first-NDR writeback only after action, safety, and merchant-outcome gates pass. |
| 6:25–7:25 | Make the product choices | proposed action | Choose outcome over employee persona, one loop over breadth, governance in the core, a shared portfolio contract, bounded pilot pricing before outcome pricing, and stateful memory with explicit policy and retention boundaries. Keep the current promise at approval-ready work until execution proof exists. |
| 7:25–8:30 | Walk through 30/60/90 | proposed action | Day 30 applies Gates G0 and G1 to verify the storefront/read/action path and choose or kill the COD wedge. Day 60 applies Gate G2 to the shadow evaluation. Day 90 applies Gates G3 and G4 to controls and one reconciled action rail. It does not yet prove repeatable ROI or first-NDR writeback. |
| 8:30–9:15 | Show the six-month gates | proposed action | Progress through Gates G0–G8: truth and storefront verification, discovery, data, shadow mode, controls, one action, alpha, controlled outcome, first-NDR depth, pricing, assisted attach, and an invest/change/stop decision. |
| 9:15–10:00 | Close like a founding PM | proposed action | Build a fast, honest learning system around one order-to-delivery P&L loop. Scale only after causal economics, trusted execution, repeatable onboarding, commercial willingness, and Shiprocket-assisted distribution show positive evidence. |
| Unsupported claim | Why it is unsafe |
|---|---|
| Rocketizer serves Shiprocket's 4 lakh+ sellers. | Shiprocket ecosystem scale is not Rocketizer adoption. |
| Rocketizer autonomously writes to merchant systems. | Public surfaces do not establish production write success. |
| Rocketizer runs on Shiprocket MCP or uses TwentyTwo for voice. | Public proximity is not an integration contract. |
| Rocketizer is entitled to use Shiprocket merchant data as a built-in advantage. | Field-level access, merchant authorization, permitted use, retention, cross-merchant rights, and incremental context lift are all unproven. |
| Rocketizer reduced failed deliveries from 14 to 2 and saved ₹2.4 lakh. | The public demo uses illustrative fixture data and is not a named customer case. |
| Power Merchants contribute exactly 85.72% of revenue. | This is an approximate analyst derivation from a rounded published ARPU. |
| The COD-to-delivered loop will create a specified delivery lift or ROI. | No preregistered Rocketizer merchant pilot has publicly established the counterfactual. |
| Outcome pricing is aligned by definition. | Untrusted attribution or uncapped liability can create disputes and adverse selection. |
| The Shiprocket AI portfolio is already unified. | Shared ownership, identity, memory, permissions, pricing, audit, and conflict resolution remain internal questions. |
| Specific Rocketizer security, compliance, residency, retention, or evaluation controls exist. | Current public evidence does not substantiate those claims. |
Evidence boundary: This narrative is a public-source interview hypothesis. Current production actions, supported storefront, field-level context access, merchant consent and permitted use, retention, cross-merchant learning rights, context lift, adoption, customer economics, architecture, governance, and strategic ownership require internal validation.