Agent-ready web - Jul 28, 2026
PageSpeed Insights Agentic Browsing: What 3/3 Means
A new 3/3 Agentic Browsing result is appearing in PageSpeed Insights. Here is what it means for SEO, AI agents, and financial-services web teams.
13 min read
AI-native products · Lending · Collections · Payments
Make AI deliver measurable financial impact.
IDAI builds AI-native products for non-bank lenders. SelfPay starts with governed collections orchestration: increasing digital self-pay, improving customer treatment, and reducing avoidable field effort. Selected design-partner programs turn live operating problems into measurable pilots and reusable product capabilities.
IDAI operating logic
From AI activity to realized financial value.
A finance-first path: identify the value pool, deploy under governance, and measure what reaches the operating P&L.
- 01
Value pools
Recovery · risk · cost · capital
- 02
Governed deployment
Controls · stage gates · owners
- 03
Realized impact
P&L · cadence · next decision
Credibility
Built from operator work inside regulated retail lending.
Collections digital strategy
Self-pay doubled
End-to-end digital collections redesign on a large retail auto portfolio: propensity selection, payment-friction reduction, and campaign discipline that produced ~₹2.2 Cr annualized savings in the first production cycle.
Deep-bucket recovery
Skip & write-off partnerships
Fintech recovery partnerships and operating design for segments traditional field models under-served — material contribution relative to a full-year PBT base.
Credit decisioning
AI underwriting pilots
Led analytics and deep-learning underwriting pilots across high-volume consumer finance — including design of test / control discipline and executive-ready trade-off frontiers.
Regulated platform build
Online lending & payments
Zero-to-one online seller-finance platform work inside a major private bank, including early cloud-data governance clearance, plus payments P&L ownership on a merchant acquiring surface.
Recognized at NPCI Bharat BillPay's global innovation challenge.
Founder accountability
Senior operator judgment stays in the work.
IDAI products and design-partner programs are founder-led by Muthukkumaran K, whose prior work spans digital collections, AI underwriting pilots, online lending, payments P&L, and zero-to-one AI product building inside and adjacent to regulated financial institutions.
Profile
15+ years
Fintech product, strategy, software engineering, and independent AI product building.
View founder profile →SelfPay
The Cure Ladder for modern collections outcomes.
Identify good-bounce customers. Orchestrate voice, SMS, WhatsApp, and IVR with rising intensity only when justified. Convert a structural share of ~₹200 field economics into digital recovery near payment-rail cost — then hand residual risk to the right specialist team.
India-ready rails. Globally portable logic.
Impact
Three focused ways to validate before scale.
Selected design-partner programs connect product discovery to live operating evidence: a collections diagnostic, a governed pilot, or a wider value-realization mandate for institutions that need evidence before scale.
Diagnostic · Pilot · Governed value realization
Perspectives
Recent Insights
Governed AI - Jul 24, 2026
The AI Security Asymmetry Financial Institutions Should Prepare For
The July 2026 Hugging Face incident is not a simple open-versus-closed model story. It is a warning that an institution's defensive analysis path must remain available when an attacker is unconstrained and hosted models refuse real attack artifacts.
12 min read
Applied AI - Jul 20, 2026
AI Voice Agents for Collections: A Value-Control Blueprint
A regulated lender should treat a voice agent as a controlled collections workflow, not a talking model. The value case depends on selection, conduct, handoff, and measurable recovery economics.
13 min read
Design partnership
Bring a lending problem that could become a measurable product capability.
IDAI works selectively with regulated lenders where a live operating problem can validate reusable product logic. Institutional work email required.