01
Start with the economic constraint.
Define the value pool, the baseline, and the decision that must change before discussing models or platforms.
Founder
Muthukkumaran K
Founder & Product Director, IDAI
An operator-builder across retail lending, collections, payments, and applied AI.
Before founding IDAI, Muthukkumaran spent 15+ years moving between software engineering, banking product management, and fintech strategy. His work has centred on getting difficult programs through data, policy, technology, operations, and executive scrutiny until the economics can be measured.
15+ years
Across engineering, banking products, fintech strategy, and independent product building
2x self-pay
From an operating redesign in a large retail lending portfolio
~INR 2.2 Cr
Annualized savings from the first scaled collections cycle
Operating thesis
“AI creates value only when the model, operating workflow, decision rights, and financial measure are designed together.”
02
Design for operating reality.
A useful pilot must survive existing systems, policy, data quality, frontline incentives, and limited change capacity.
03
Make the result defensible.
Use explicit controls, clear ownership, and a finance-readable result so a promising signal is not mistaken for realized value.
Selected operating stories
The work behind the claims.
Collections economics
2x self-pay · ~INR 2.2 Cr annualized savings
A small operational pilot that earned the right to scale.
An early analytics-led collections experiment produced ambiguous attribution and exposed a deeper problem: the operating model, not only the score, needed to change.
Muthukkumaran reframed the work around payment friction, BBPS behaviour, customer response, and controlled escalation. He kept the first pilot deliberately light, using simple scripts and existing campaign teams so the institution could test the economics before committing to a large technology program.
The approach won executive backing, overtook existing digital payment modes, and doubled self-pay within six months. That operating logic is part of the foundation for SelfPay.
Regulated platform build
70% of the workflow digitized · 3,000+ UAT cases
A new lending platform, including the constraints that limited it.
At a major private bank, his first product assignment was to help build an online seller-finance platform across credit, policy, legal, information security, technology, operations, and external partners.
He drafted the cloud and customer-data proposal that cleared information security review, led a cross-functional test program, and helped take the platform live. Adoption was ultimately constrained by credit filters for early-vintage sellers.
The lesson was durable: digital execution cannot compensate for a product-policy mismatch. The platform's operating learning later informed a more successful overdraft product for existing current-account customers.
AI decisioning discipline
A structured pilot · no overstated production claim
Promising model evidence, followed by an inconclusive live test.
In a consumer-finance underwriting program, historical model work showed an attractive approval-rate and bad-rate frontier. The live pilot used explicit test and control populations while most applications remained in business-as-usual flow.
A control-group integration defect made the live comparison unreliable. Muthukkumaran brought the product, analytics, startup, and lending-system teams into one debugging session, helped isolate the fault, and treated the production evidence as inconclusive.
For IDAI, this matters as much as a successful model: governance includes knowing when the evidence is not strong enough to support a value claim.
Working with IDAI
What a partner can expect.
Founder attention
The senior operator shaping the engagement remains close to the data, working sessions, decisions, and result readout.
Cross-functional translation
Finance, risk, collections, product, analytics, technology, and partners work from one problem definition and one value baseline.
A clear stop-or-scale decision
Every pilot is designed to end with evidence strong enough to scale, revise, or stop without protecting sunk effort.
Operating experience
IDAI / Imayan Digital LLP
2021 - Present
Founder-led AI value realization and the SelfPay collections product direction.
TVS Credit
2018 - 2021
AVP, Fintech & Strategy: digital collections, AI underwriting, recovery partnerships, long-range planning, and initiative P&L.
ICICI Bank
2015 - 2018
Product management across online lending, payments, strategic projects, and the founding team of ICICI Innovation Labs.
Aricent, now part of Capgemini
2008 - 2013
Software engineering for high-availability network protocol products and internal automation tools.
Confidential discussion
Bring the value question, the operating constraint, and the evidence available today.
The first conversation is used to determine whether IDAI can add value, what would need to be true, and whether a focused engagement is warranted.