In short
Indian banks do not run one system for post-complaint operations. They run a stack: case management, fraud investigation, complaints handling, workflow, evidence collection and regulatory reporting, each covering part of the lifecycle.
The cost is not in any one tool. It is in the seams between them. In one Indian banking operation, an analyst touched five systems per case, and complaint response ran at about three days of which almost none was decision-making.
What post-complaint operations covers
Post-complaint operations is everything that happens after a customer, a law enforcement agency or a regulator raises something, and before the case is genuinely finished.
For an Indian bank that spans NCRP complaints, I4C portal cases, LEA information requests and court orders, customer and merchant correspondence, money restoration, grievance redressal and post-complaint risk profiling.
It is not fraud detection. Detection decides whether something is suspicious. Post-complaint operations decides what the bank does about it, proves what it did, and answers for it afterwards.
That distinction matters when you are buying. Detection tools are bought by risk. This work is owned by operations, and the two rarely share a budget line.
The government now describes the same scope. The Ministry of Home Affairs SOP of 2 January 2026 covers complaint processing, bank coordination, grievance redressal, removal of lien markings and restoration of defrauded funds. Between 2023-24 and 2025-26, more than ₹56,087 crore was reported through NCRP cyber fraud complaints.
The nine stages a case passes through
- Intake. The case arrives through a portal, an API or an email, each in a different format.
- Triage. Someone decides what kind of case it is and whether it belongs to the bank at all.
- Investigation. Account status, transaction history and KYC records get pulled and read.
- Evidence collection. Statements, documents and logs are assembled into something reviewable.
- Decision. The bank decides what it is going to do, under its own rules.
- Remediation. A hold, a freeze, a reversal or a payment happens, or does not.
- Regulatory reporting. The response goes to the portal or the agency in the required format.
- Closure. The case is marked finished, or marked as something more specific.
- Root cause. The pattern gets fed back so the same case does not arrive again next month.
Most banks have a tool for some of these stages and a person for the rest.
The six tool categories, and what each one misses
| Category | What it does well | Where it stops |
|---|---|---|
| Case management | Holds the record and the status | Does not gather the evidence |
| Fraud investigation | Scores and flags transactions | Ends at the alert, not the response |
| Complaints handling | Tracks the customer relationship | No view of the regulatory queue |
| Workflow platforms | Routes work between people | Someone still has to do the work |
| Evidence and document tools | Store and retrieve files | No judgement about what is needed |
| Regulatory reporting | Formats and submits | Assumes the case is already resolved |
Each category is competent inside its boundary. None of them owns the case from intake to a defensible outcome, which is why the analyst does.
Patched turns messy, manual, spreadsheet-driven operational processes into reliable AI-assisted workflows with human review where needed.
The seams are where the hours go
Stitching is the real cost. An analyst handling one case moves between the core banking portal, the bank's API, the data lake, the internal service desk and the email inbox.
That is five systems for one case, and the cases do not arrive in one queue to begin with.
In one Indian banking deployment, complaint response ran at about three days. Almost none of that was decision-making. It was retrieval, reformatting and waiting.
When that retrieval stopped being manual, response time moved to under an hour and analyst hours per case fell roughly 75%.
Why one managed workflow replaces the stitching
The alternative to buying a seventh tool is to put one managed operating layer underneath the stack you already have.
It reads from your existing systems, assembles what a decision needs, prepares the response in the required format, and keeps a record of what was checked and what was skipped. The bank's existing tools stay where they are.
This is deliberate. Banks do not rip out core banking, ticketing or a case manager because a new tool arrived, and any vendor implying otherwise has not sold to one.
Patched is not trying to rip out a bank's existing tools.
What the stack looks like when it is working
Production outcomes at a leading Indian payments bank, from a 15,000-case starting backlog. Source: caselea.com, 2026.
The fourth number is the one that explains the rest. Coverage across 76 codified resolution scenarios is what lets the operating layer carry real volume rather than a demo subset.
How to evaluate your own stack
Walk one real case end to end and time each stage. Not a representative case, and not your worst one - pick a middling case from last week.
Write down every system the analyst opens and every point where work waits on another team. Then count how much of the elapsed time was decision and how much was retrieval.
If retrieval dominates, buying another tool in one of the six categories will not move your backlog. The problem is the seams, and the seams are not anybody's product.
Who owns this in the org chart
Post-complaint operations rarely has a single owner, and that is the root of most of its problems.
Fraud risk owns detection. Customer service owns the complaint. Compliance owns the regulatory response. Operations owns the queue. Technology owns the integrations. Each of them owns part of a case and none of them owns the outcome.
The practical symptom is that nobody can say what a case costs. Ask five teams and you get five partial answers, none of which includes the other four.
This matters for buying, because a tool bought by any one of those functions will solve that function's slice and leave the seams intact.
The one number that settles the business case
Cost per case, measured end to end across every team that touches it.
Most banks have never calculated it, because the components sit in different cost centres. Building it once is unglamorous work and it is the number every subsequent decision depends on.
Against it, the comparisons become simple. In one deployment, case response moved from about three days to under an hour and analyst hours per case fell roughly 75%, against a backlog that started at 15,000 cases and came down 95%. Whether that is worth doing is arithmetic once the denominator exists.
Sources
- caselea.com production outcomes, one leading Indian payments bank, 2026.
- Patched data points, 2026: five systems per case, 3 days to under 1 hour, roughly 75% analyst hours.
- Patched sales brief by Rohan Sood, CEO, 2026.
- Ministry of Home Affairs, reply in the Rajya Sabha published by PIB, 22 July 2026: more than ₹56,087 crore reported, 2023-24 to 2025-26.
- Ministry of Home Affairs, Standard Operating Procedure for NCRP-CFCFRMS, Custody, Restoration of Money and Grievance Redressal, 2 January 2026: scope of complaint processing, grievances, lien removal and restoration.
