If you sit with a credit ops team for a day at any mid-sized Indian NBFC and watch what they actually do, the picture is not flattering. The morning starts with email triage, borrower applications, partner submissions, document follow-ups. By eleven the team is on calls chasing missing files. By two they are downloading bank statements, copy-pasting figures into spreadsheets, opening a separate tab to check GST status, another to pull the bureau, another to check NCLT. By five they have built one due diligence file. The borrower has been waiting four days.
The work is not credit analysis. It is reconstruction.
The cost compounds
Each individual due diligence file is a four- to six-hour exercise. Multiply by ten files a week per analyst, by eight analysts on the team, by fifty weeks a year, and the credit ops team spends roughly twenty thousand person-hours a year reconstructing third parties from scattered sources. At even modest fully-loaded analyst cost, that is a serious line item, and none of it is the work the team was hired to do.
The cost shows up in three less obvious places too.
First, conversion. Borrowers who wait four days for a due diligence verdict ghost. The next-best lender takes them in two. Conversion rate erodes silently as the file-build time grows.
Second, decision quality. A four-hour file rarely covers everything. Litigation gets a quick courts.in search. Related-party exposure gets a glance at the MCA portal. Adverse media gets one Google query. The deeper checks, director-level bureau, three-way reconciliation, secondary sources, are the ones that catch the worst defaults, and they are also the first to get skipped under deadline pressure.
Third, monitoring. Once the loan is on the books, no one re-runs the due diligence. The team has moved to the next batch of files. The original borrower is reviewed at the next quarterly EWS cycle, by which time three months have passed.
Why this hasn't been fixed
The reason credit ops teams still do reconstruction by hand is not laziness. It is that no one has built the infrastructure that does it for them. The data sources exist, bureau APIs, Account Aggregator, GSTN, MCA21, NCLT databases. But knitting them together into a single sourced, audited due diligence report has been beyond what most internal IT teams can ship. So the work falls to humans, every loan, every time.
The lift-and-shift approach, buying a workflow tool that reorganises the manual work, does not move the needle. A reconstruction that takes four hours in Excel takes three hours forty-five minutes in a workflow tool. The bottleneck is the work itself, not the tooling around it.
Due Diligence as infrastructure
The shift that actually moves the needle is treating third-party due diligence as infrastructure, a service that runs continuously, in seconds, against any third party in your pipeline, without human operation. The borrower is onboarded. The due diligence report is generated automatically. The credit officer reviews the output and makes the credit decision. The borrower goes on the books. The same due diligence report continues to refresh in the background. Stress signals fire as they emerge. The audit pack is regenerable on demand.
What changes when due diligence becomes infrastructure: the credit ops team stops reconstructing and starts deciding. The four-hour file becomes a six-second report. Conversion improves because borrowers wait hours, not days. Decision quality improves because every check runs every time. Monitoring becomes continuous rather than quarterly. The audit pack writes itself.
None of this is hypothetical. AICA does it every day for one hundred and twenty-plus banks, NBFCs and enterprises in India. The thing that took your team six hours yesterday takes seven seconds tomorrow.
The bottleneck is not your team's effort. The bottleneck is asking them to do the wrong work. Move that work into infrastructure, and the team you already have can ship two or three times the volume.