Faster Property Due Diligence Cuts Credit Risk, Not Just Waiting Time

Many credit risk officers file due diligence speed under borrower experience: a TAT figure on a dashboard that operations worries about and risk puts up with. That view is mistaken, and it quietly costs portfolios money. Speed in property due diligence reduces risk. The way it does so is more interesting than the obvious point that borrowers drop out.

Here we argue that DD turnaround time deserves a place beside LTV, FOIR and bureau score as a first-class risk metric. The readers we have in mind are CROs, credit committee members, heads of mortgage products and legal-vendor managers at banks, NBFCs and HFCs.

What it means to treat DD speed as a risk lever

By due diligence speed we mean the clock time between receiving the property documents and issuing the final legal opinion, or the Title Search Report (TSR) signed by an advocate. If it drops from the usual 5-7 working days to below one, several risk vectors shrink at once: borrowers have less time to switch lender, file-level review improves, deadline shortcuts disappear, court data is fresher and a re-check before disbursal becomes affordable. Speed is a covariate. Improve it and several risk outcomes get better together, beyond the headline TAT.

A CRO who sees speed only as a borrower satisfaction issue misses the larger point. The benefit to the portfolio appears in NPA correlation, in the risk of vendors missing SLAs and in the audit trail.

Six ways speed lowers risk

Speed is six levers and not one. Each of the mechanisms below touches a different risk vector in a typical mortgage portfolio.

1. The window for borrowers to switch closes

Between sanction in principle and final disbursal there is a gap of 5-7 days, and in that gap competitors quote a better rate and the borrower leaves. NBFCs report that 23% of borrowers drop off during this stretch. A good share of these exits are not about price but about confidence: the borrower reasons that a slow lender must also be a disorganised one, and goes elsewhere.

The risk here is eroded portfolio yield and not credit loss. The bank spends on origination, marketing and underwriter time, and books no asset. Faster DD narrows the window where this happens.

2. Reviewers keep their attention

Compare a legal vendor who handles 30 files a day at 4 hours apiece with one who handles the same 30 files at 30 minutes apiece, AI-assisted retrieval doing the mechanical work. The mental states differ. In the first case attention is rationed and the last files of the day get only 60% of the effort. In the second the reviewer has attention left to read the document chain, where defects hide.

When DD gets faster through parallel work and not through cutting corners, more files are reviewed properly, not fewer files reviewed badly. That is the counter-intuitive part. People usually set speed against thoroughness, but the two clash only when speed comes from skipping steps. Speed that comes from parallelism adds to thoroughness.

3. Rushed files stop happening

Every credit committee has heard of a file that closed within 48 hours because a stamp duty deadline loomed, its legal opinion written overnight. Such opinions are almost always lighter: fewer portals, an eCourts search stopping at the district court, no IBBI search for a corporate seller. The defects come to light later, often after disbursal.

When the standard DD timeline is under one day, the "rush file" no longer exists as a category. No file needs rushing, since every file is already quick. The checks that deadlines usually cut, notably the valuable ones such as the SARFAESI search, lis pendens and MCA charge search, are done as a matter of course.

4. Court data is fresher

Think of a case filed against a property or its seller only last week. For a loan underwritten this week it is a live risk. Manual DD that spans seven days and queries eCourts on day one reads data that is a week old when the opinion is signed. AI-assisted DD that runs the search as the opinion is issued, minutes before sign-off, reads data that is live.

The gap matters. Lis pendens, partition suits, injunctions and SARFAESI proceedings can all be filed between pulling the documents and issuing the final opinion, so a shorter window leaves less of a blind spot. In high-ticket files the effect is out of proportion. The dispute that brings down a title is seldom the one from 1997. More often it was filed last month by an aggrieved co-heir who has just heard the property is being sold.

5. Timestamps in the audit trail help compliance

A DD pipeline that takes 30 minutes produces an audit log by machine: each portal queried, each response, each red flag and each advocate sign-off, all timestamped. This is what a regulator or internal auditor wants for a stressed account. A 7-day manual process leaves a final report and little more, with no record of what was actually checked.

For HFCs supervised by the NHB, and for banks that face RBI's periodic audits of legal vendors, the timestamped trail is a risk-reduction asset in itself. It turns the question "did the vendor really run all the checks?" into "show me the log".

6. Re-verification before disbursal gets cheap

Say the first DD costs Rs.2,000 per file and needs 5 days. Repeating it just before disbursal is painful, in operations and in money. Most lenders skip it, even though sanction and disbursal are often 30-60 days apart, long enough for encumbrances, court cases and the seller's status to change in a material way.

Make the same DD cost a fraction of that and run in 30 minutes, and a check before disbursal is no burden. Some institutions already run a "freshness check" within an hour of releasing the cheque, re-querying the EC, eCourts, MCA and SARFAESI portals lightly. It costs little. The NPAs it heads off, even at an incident rate of 2-3%, are many.

How each risk responds to speed

• Borrower leaving for a competitor · Slow DD (5-7 days): wide window, drop-off of 20%+ · Fast DD (under 1 day): the window closes and fewer borrowers leave
• Reviewer fatigue, steps skipped · Slow DD (5-7 days): the last 30% of files are reviewed in a hurry · Fast DD (under 1 day): each file receives its full share of attention
• Portal checks dropped to meet deadlines · Slow DD (5-7 days): SARFAESI and IBBI checks are often left out · Fast DD (under 1 day): all 20-plus portals swept for every file
• Out-of-date court data · Slow DD (5-7 days): a blind spot lasting 5-7 days · Fast DD (under 1 day): data only minutes old at sign-off
• How complete the audit trail is · Slow DD (5-7 days): a final report with no granular log · Fast DD (under 1 day): every check logged with a timestamp
• Re-verification ahead of disbursal · Slow DD (5-7 days): costs too much to run · Fast DD (under 1 day): costs little and runs when the cheque goes out
• Vendor SLA breaches · Slow DD (5-7 days): SLAs are missed often in busy weeks · Fast DD (under 1 day): spare capacity takes up spikes in volume
• Defending a file when it is disputed · Slow DD (5-7 days): hard to rebuild what was checked · Fast DD (under 1 day): everything is logged and defensible
• Risk building up in rushed Friday files · Slow DD (5-7 days): shows up in NPA cohorts · Fast DD (under 1 day): the cohort disappears

The risk of vendors missing SLAs

Slow DD creates a structural vendor-management problem too. Take a bank whose legal vendor needs 5 days on a baseline file. A sudden jump in volume (a month-end disbursal push, a festival quarter, a CRE deal closing) stretches that SLA to 8-10 days. Under pressure on throughput, the product team either pushes files through on a thinner check or sends the overflow to a second, less-vetted vendor.

Both choices are bad. The first brings credit risk into the portfolio. The second brings vendor risk, because the second vendor has different rubrics, different portal access and a different quality of sign-off.

A fast pipeline takes volume spikes in its stride and quality holds. In that sense throughput lowers risk. A bank that can clear 10x its normal volume on a peak day without thinning checks avoids the rushed-file cohort that turns up as NPAs six quarters on.

Why NPA correlation falls

Lenders that have shifted to AI-augmented DD in the last 24 months report two things that go together. More files are flagged for extra checks, typically 5-10% more, because more portals are queried and more red flags emerge. And the NPA rate on disbursed files declines across a 24-month window.

There is no mystery in the mechanism. Slow manual DD systematically misses some risks. These lenders are not becoming luckier on files that turn into NPAs. Those files were always going to go bad, and the earlier process failed to catch them at underwriting. A faster and more thorough DD catches them, the credit committee declines or restructures them, and they never enter the book.

For a CRO this is the most important change of framing: faster DD does more than lower loss-on-default. It lowers probability-of-default, because it changes which files are sanctioned in the first place.

What "faster" must not mean

All of this holds only if the speed comes from parallel work, automation and better access to data, and not from cutting corners. A DD that is faster only because the vendor dropped the eCourts case search is worse than a slow DD that ran it. The same goes for the depth of the encumbrance chain, SARFAESI checks and lis pendens screening.

A credit risk team judging a faster DD vendor can apply three tests:

• Request the list of portals the vendor queries for each file. A good vendor's list lengthens with time. A shrinking list is a warning.
• Ask for a sample audit log. A vendor that cannot produce a timestamped log of each portal queried and each check raised is buying speed by skipping steps.
• Ask what happens to a file with low-confidence OCR. If the answer is "we accept the lower confidence to keep the SLA", that is a red flag. The right answer is "we route it for human review and the SLA absorbs it".

One concrete example of 20-plus checks in parallel

An AI title search engine of today, LegiScore among them, writes a 29-section legal opinion in under 15 minutes. It runs parallel queries on 20-plus portals and on the eCourts network spanning 28 states and 600-plus districts. The figure is quoted not for show but as a benchmark: it indicates what "fast and thorough" can mean in 2026, and a 5-day vendor SLA should be judged against it.

Where a rival vendor takes 5 days to do less, the CRO's question is not "is the SLA acceptable?" It is "what am I getting for the SLA, when a peer benchmark of 30 minutes with wider coverage exists?"

Putting speed to work as a risk metric

A CRO who takes this case to a credit committee will do better with three practical steps than with theory:

• Start with the NPA files of the past 24 months. Reconstruct each file's DD log, where one survives, and count how often a portal skipped at underwriting later turned out to hold the defect. Those are the files that slow DD missed.
• For the same 24 months, measure the borrower drop-off rate by DD turnaround bucket: files closed in under 2 days, in 2-5 days and in over 5 days. The drop-off rate is usually monotonic with TAT.
• Work out the all-in cost of a slow file: the cost of drop-off, the cost of reviewer attention, the NPA contribution of missed checks and the cost of defending the audit trail. Set it against the per-file cost of a faster DD vendor. At almost any volume the economics favour speed.

To read more on the property risks that fast DD catches more dependably:

• Property due diligence for home loans: what banks require
• Property verification for banks and NBFCs: a guide
• Title chain verification: 13 years against 30 years
• The SARFAESI Act and property: what lenders must check
• Common property frauds in India and how to avoid them