Banking & Capital Markets Systems · Chapter 13 of 13
What you can and cannot find out, and what to ask instead
Roughly half the vendor-to-institution map is unsourceable. That is a fact about the field, and it is usable.
This book set out to answer “which institution runs which system”. On the retail and payments side that mostly worked. On the capital-markets side it mostly did not, and the pattern in the failures is more useful than another vendor list.
What is genuinely knowable
- Core banking at large Indian banks — sourceable, because vendors publicise wins and banks disclose major migrations. SBI, HDFC, ICICI, Kotak, IndusInd, PNB, IDFC First all check out.
- Anything a listed company files — which is why the CAMS and KFin section is the best-evidenced in the book, and why depository market shares are solid.
- Anything a regulator mandates — settlement cycles, the algo regime, digital lending directions, the reporting mandate, the 24×7 rails. These are published, dated and durable.
- Market structure — CCIL's role, the NPCI/RBI split, the aggregator boundary, the bureau licensing. Structure changes slowly and is documented.
What is not, and why
Treasury, trading, wealth, fund accounting and AML platform choices are largely dark. No Indian institution could be confirmed on Murex, Calypso, Kondor, Charles River, Aladdin, SimCorp, Avaloq, Temenos WealthSuite, Broadridge, Actimize, FCCM or OneSumX with a usable citation. Neither could the core platform at Axis, Yes Bank, Bank of Baroda, most public sector banks, either large small-finance bank, any payments bank, or the India operations of Standard Chartered, HSBC or Deutsche.
Canara Bank is the instructive case: the available sources contradict each other outright, one summary using “Finacle” and “i-flex” almost interchangeably within a single paragraph. That is what it looks like when a fact is being copied rather than checked.
The mechanism is worth understanding, because it generalises.
A platform win is announced once, in a press release, at signing. It is then never mentioned publicly again — not at go-live, not at upgrade, not at replacement. So the public record is a set of signing announcements, some of them fifteen years old, with no indication of what survived.
The aggregator sites that publish confident “which bank uses which core banking software” lists are working from exactly that record, and from each other. Two of the most widely repeated claims in this area — HDFC on Finacle, IDFC First on Finacle — are both wrong.
What to ask instead
If you cannot know the platform before the interview, ask about it in the interview — and ask the questions whose answers tell you what the job is:
- “What version are you on, and when did you last upgrade?” A full core migration is a multi-year, board-level event. The answer tells you whether you are joining a migration or a maintenance posture, which are different careers.
- “What is wrapped and what is replaced?” The HDFC pattern — a cloud-native shell in front of a monolith nobody intends to touch — is far more common than replacement, and it is where the interesting engineering is.
- “Where does the batch window still bind?” Since RTGS (Real Time Gross Settlement) and NEFT (National Electronic Funds Transfer) went 24×7, any remaining end-of-day dependency is a live architectural constraint. Whoever answers this well knows their own system.
- “Is this the bank's platform, or the parent's?” In a GCC these are completely different jobs, and the job description often does not say.
And carry the epistemics into the interview itself. Saying “I could not verify which platform you run — the public record on this is mostly signing announcements — so here is what I would want to know” is a stronger opening than reciting a vendor list you got from a blog that got it wrong.
Check yourself