14 Products in 20 Days: How SabPaisa Built an AI-First Product Org
SabPaisa (SRS Live Technologies Pvt. Ltd.), an RBI-approved payment aggregator in India, launched 14 enterprise-grade products at once. A 13-person team shipped the entire slate in 20 days. That output would normally take bigger orgs multiple quarters.
Led by Founder & CEO Kumar Manish and Co-founder Abhimanyu Jha, the company leaned into an AI-first operating model and a prototype-first build cycle. The goal wasn't headcount or hype-it was speed, learning, and precision.
What happened
SabPaisa rolled out products across core payments, operations and compliance, and merchant analytics-simultaneously. Each product was developed by individual engineers working with advanced AI systems to architect, build, test, and deploy production-ready software.
The team skipped traditional phase gates. Ideas moved straight to working prototypes, launched internally for immediate feedback and usage data. High-signal prototypes advanced. Low-signal ones were refined or parked.
Why product leaders should care
- Shipping speed compounds learning. Shorten loops; quality rises with each iteration.
- AI paired with single owners reduces coordination overhead and context switching.
- Internal dogfooding creates built-in user research and sharper prioritization.
- Cross-functional AI fluency removes handoff friction between product, engineering, ops, and marketing.
The AI-first operating model
- Every team member engages in AI development-no "AI team," no silos.
- Engineers partner with AI for spec drafting, code generation, test creation, and documentation.
- PMs and ops contribute prompt libraries, test scenarios, and feedback loops.
- Leaders track throughput and quality, not headcount or ceremony.
The prototype-first cycle they used
- Idea → AI-assisted prototype in days, not weeks.
- Immediate internal launch for dogfooding and telemetry.
- Quant + qual review: usage, errors, time-to-first-value, NPS-style feedback.
- Promote high-usage prototypes to production behind flags; refine or pause the rest.
- Repeat weekly with clear owner, SLA, and rollback plan.
Metrics that drove prioritization
- Prototype adoption rate within 48-72 hours
- Time to first successful transaction or core task
- Error rate and recovery time
- Retention across internal cohorts (daily/weekly)
- Support load per feature (tickets per 100 users)
Org design choices that enabled speed
- Single-threaded owners for each product/feature
- Shared AI toolchain and prompt libraries
- Automated test scaffolding and compliance checks in CI/CD
- Feature flags and instant rollback paths
- Company-wide policy: contribute feedback as part of daily work
Fintech guardrails (non-negotiable)
As an RBI-authorized payment aggregator, compliance and risk cannot lag behind shipping speed. Build the rails before traffic hits scale.
- CI hooks for logging, audit trails, and data retention
- Automated checks for KYC/AML workflows and permissions
- PII isolation, access controls, and key rotation by default
- Secure prompt handling and red-teaming for AI features
- Runbooks for incident response and regulatory reporting
Context: RBI guidance for Payment Aggregators is a useful reference point for controls and approvals. See the regulator's site: Reserve Bank of India.
What leadership said
"We're not competing on marketing spend or team size anymore-we're competing on intelligence and speed," said Co-founder Abhimanyu Jha. "This isn't about replacing humans with AI-it's about fundamentally reimagining what becomes possible when you stop treating AI as a tool and start treating it as a collaborative force."
Jha added an analogy: "When cars were first invented, Karl Benz didn't create steel horses with engines inside them-he completely reimagined transportation. We're doing the same at SabPaisa."
What to watch next
- Which of the 14 products hit sustained adoption and revenue contribution
- How the prototype-first model scales beyond the initial team size
- The impact on compliance throughput and operational cost per transaction
Details on the full launch are here: sabpaisa.in/ai-native-fintech.
Steal the playbook (adapt to your org)
- Mandate AI usage across roles; publish prompt and pattern libraries
- Ship prototypes weekly; measure with the same dashboard across teams
- Prioritize features by usage and time-to-value, not opinions
- Bake compliance and security into CI/CD, not post-facto reviews
- Reward shipping and learning velocity over perfect specs
Level up your team's AI fluency
If you're building a similar operating model, align skills to roles and ship cycles. Curated options for product and engineering teams: AI courses by job.
Your membership also unlocks:
