The thinking behind the platform.
Product updates and essays on lending infrastructure, AI governance, and agentic loan servicing. Guides, the glossary and the RFP toolkit live in the resources library.
More from Lokta
33 articles
Payment appropriation is a configuration science
A single loan product can order its payment appropriation more than 10 million ways. The order decides what a payment retires, and it lands on the product's RoA.
11 min readRead
RBI's proposed revolving-credit restriction: what it means for NBFC lenders
The draft goes beyond reusable credit limits. It could change how NBFCs design, book and service consumer and business credit across the live book.
14 min readRead
Loan Against Property (LAP) lenders in India: the five challenges that start after disbursal
Five things drain a loan against property book after the money goes out: frontline attrition, field collections, balance transfer, the top-ups nobody writes, and the statutory clocks that cost referrals when they slip. With a published signal behind each one.
14 min readRead
We stopped copying dummy rows and built a living, 24/7 digital twin of a company as our QA, just to break our software. Here's why.
Most lending software is tested on stale, made-up data. That fake data, not the feature, is often the real bug. So we built a living NBFC we can test on.
17 min readRead
Inside Moneyview's technology and AI strategy
Moneyview's DRHP reads like a technology company's filing: AI across underwriting, collections and servicing, a hybrid LSP-plus-NBFC book, and a widening product line.
11 min readRead
Loans against mutual funds: fast to pledge, painful to exit
We read every borrower complaint we could find on loans against mutual funds and shares, then sorted each by where it lands in the life of the loan. The pain clusters after approval.
9 min readRead
A loan's profit is decided after approval, not at underwriting
The credit decision is the least important call a lender makes. A loan's profit is won or lost after approval: in servicing, monitoring, and collections.
8 min readRead
RBI Just Redefined What Counts as a 'Model'
I spent the last few days reading RBI's draft model risk guidance. The headline is not AI. It is that RBI just redefined what counts as a 'model,' and a pricing spreadsheet now qualifies.
14 min readRead
Our team is 42. Six of us are human.
Lenders keep asking how a six-person team ships this fast. The honest answer: everyone codes, and a role-specific agent fleet, pointed at four product principles and two decades of lending scars, runs the book of work alongside us.
11 min readRead
Credit is the bloodline of the economy
I'm co-founding Lokta, the agentic loan servicing platform lenders, fintechs and LSPs use to run the live book: servicing, monitoring, collections, recovery.
7 min readRead
What money actually is
Money is a claim on future productivity, and lending is the oldest way of allocating it. This is how capital really flows through India today, and why the NBFC sits at the most exposed node of the whole cascade.
11 min readRead
The price of money is uniform. The outcomes are not.
Same input price, opposite results. The cross-section and the cycle both say the separation between lenders is not the cost of capital. It is the intelligence applied to risk. And digitisation, for all it did, never touched the judgment.
11 min readRead
The loan book that learns
Why a loan book can finally learn, and why it is different from anything that learned before it: its own decisions shape the data it learns from. The flywheel, the borrower it includes, the brakes it needs, and the infrastructure Lokta is building to run it.
23 min readRead
Build, buy, or compose: the credit-decisioning call a lender gets one shot at
Custom credit-decisioning builds fail in three predictable ways. For fast-growing lenders in India, APAC, MEA and Africa, this is how to fund one that lasts.
8 min readRead
Why your LMS ledger should never be model-guessed
LLMs predict; they do not calculate. Loan ledger math must be deterministic and reproducible. Here is what that means for your LMS and your next RBI audit.
9 min readRead
Why Every Lender's Portfolio Analytics Lives in Excel
The LMS data trap: why portfolio managers, CROs and CFOs end up rebuilding loan data in Excel, and the five RFP questions that surface it before signing.
9 min readRead
The Customization Trap: When LMS Vendors Bill Every Change
Why every 'small change' in an LMS or LOS funnels through a billable vendor project, and the RFP questions that surface the trap before you ever sign.
10 min readRead
Ten Things Your LMS Needs to Do in 2026
The 2026 LMS requirements list: AI maker-checker trails, 5-10× tool-call EOD throughput, event-sourced accounting, policy-bounded agents, single-binary deployment, and operability for a 20-person team. Lokta Phase 2 ships against all ten.
11 min readRead
AI in Lending: Where the Real Value Will Land in the Next 3-5 Years
A practitioner's view on AI's real role across underwriting, operations, and collections, and where the lending industry is quietly over-investing today.
7 min readRead
The 30-Call Limit: Why Your LOS API Quietly Caps Your Lending Velocity
LOS and LMS rate limits aren't infrastructure constraints. They are commercial controls, and the 30-call ceiling shapes what a lender's roadmap can ship.
7 min readRead
100 LMS Reviews: The Same Six Complaints Recur
We read 100+ public LMS and LOS reviews from loan officers and servicers. The same six complaints recur across every vendor: here is the theme-by-theme map.
10 min readRead
Agentic AI Is the Next Operating System for Lending
Why the next 10× productivity unlock in credit comes from a workforce of AI agents, not a better underwriting model. A lending-lens view of the shift.
8 min readRead
How to Use AI Agents in Loan Servicing
Loan servicing in plain terms, the ten categories of servicing tickets, which ones an AI agent can run end to end today, and the deployment playbook.
13 min readRead
Every Lender Replaces Their LMS Eventually
Five symptoms tell you whether you replace your LMS on your terms, or the board's. Three or more on the table is a board-level risk, not an IT one to defer.
5 min readRead
Agentic Lending and the 5× Problem
The napkin math no one has done: agentic collections needs 60-120 governed tool calls per account, the 5-10× multiplier no legacy LMS is built for.
7 min readRead
The Lending Technology Pain Map: 3 Audiences
Loan officers, in-house tech teams, and executives describe the same LMS in three different vocabularies. The pain map every buying committee needs to read.
8 min readRead
Why Lokta Is Polylithic: One Binary, Many Modules
Microservices were wrong for lending; monoliths can't scale teams. Polylithic Gradle modules in one Spring Boot binary: 50 engineers do what 500 used to.
6 min readRead
What We Learned Building Apache Fineract
Three architectural lessons from a decade building Apache Fineract, and why incremental upgrades couldn't meet what modern lending now demands.
5 min readRead
The Bullet Train Problem: Legacy LMS vs. AI Lending
Agentic AI is the bullet train. Legacy LMS is narrow gauge. Here's why AI-readiness is a systems property, not a feature you can bolt on.
6 min readRead
Cumulative Impact of Apache Fineract on Global Lending
~$500B, half a trillion dollars, in cumulative loan principal disbursed via Mifos → Apache Fineract (2006-2026). Our central estimate, with full methodology.
7 min readRead
AI in Lending: Why a Deterministic Core
How Lokta deploys AI in three layers (deterministic core, workflow AI, and agentic AI) without breaking trust, compliance, or balance-sheet correctness.
3 min readRead
Lokta's Philosophy for Auditable AI in Lending
Eight truths that define how Lokta uses AI in lending: deterministic core, evidence-grounded outputs, human-controlled decisions, built-in auditability.
3 min readRead
Connected Lending: Fixing Loan Lifecycle Chaos
Why fragmented lending stacks compound costs, NPAs, and compliance risk, and what a truly connected, single-source-of-truth lending platform looks like.
3 min readRead