12.7 Phase 7 — Scale and automation
Move from “lender that built a platform” to “lender that runs at scale on a platform that is also a product”.
Features
Section titled “Features”- Full multi-tenancy — schema isolation, per-tenant config, per-tenant white-label.
- Self-serve partner onboarding — partner can configure their flow without engineering involvement.
- New product launches — LAP / equipment / new segments — without code change for
~80%. - New geographies — additional state coverage, language coverage, regulatory variant.
- Advanced ML — segment-specific models; ensembles; ECL / LGD / EAD models.
- Self-serve analytics — business users query the warehouse.
- Productised SaaS offering — sold to external NBFCs (per 1.10).
- API-first for all customer scenarios.
Integrations
Section titled “Integrations”- Additional vendor categories as needed.
- White-label channel SDKs.
150 – 250+total per 9.2.
Ongoing. Continuous improvement, not a fixed-duration phase.
Build vs buy
Section titled “Build vs buy”- All competitive moats are own.
- Primitives still bought.
- Vendor mix optimised based on volume + economics.
Acceptance criteria
Section titled “Acceptance criteria”- Multi-tenant operations stable.
- Per-tenant data isolation verified by audit.
- New product launched in
< 6 weeksfrom decision. - Self-serve partner-onboarding completed in
< 30 days. - One external SaaS customer paying.
- Stretching the team across many priorities.
- Multi-tenancy bug → cross-tenant data leak.
- Customisation requests fragment codebase.
End state
Section titled “End state”A platform that:
- Runs a profitable own + co-lent book.
- Powers
2 – 3other NBFCs as a SaaS customer. - Onboards a new co-lending partner in
< 60 days. - Has model-based underwriting maintained continuously.
- Maintains audit-grade compliance at all times.
- Operates at
~60%cost-income ratio. - Generates
₹15 – 25+ Crof annual profit.