Dispute Resolution: Ideal vs Reality

The first time the gap really shows up is rarely in a risk committee.

It’s usually in a complaints review that was supposed to be routine.

Someone has put up a slide:

“Credit bureau–related complaints – Q3.”

The numbers are small compared to the overall complaint volume.

A few hundred, maybe a couple of thousand in a large institution.

The Head of Service says something along familiar lines:

“Most of these are routine disputes – customers claiming wrong DPD, closed loan showing as open, limit not updated.

TATs are within policy. No systemic issue.”

A business head nods and asks what sounds like a reasonable question:

“Any impact on our lending? Are we wrongly declining good customers because their disputes are stuck?”

The analytics team answers carefully:

“We don’t see a material effect at a portfolio level. Individual cases, yes, but not something that moves the needle overall.”

The room is satisfied.

The slide moves on.

Nobody asks the impolite version of the question:

“When a customer contests our data in a bureau, and everyone in our ecosystem is supposed to fix it quickly and fairly, what actually happens?”

You see that version later.

It appears when:

·         A senior customer escalates through the Board or RBI because they’ve been blocked from a loan for six months due to an incorrect write-off remark.

·         An ombudsman order lands on someone’s desk with phrases like “deficiency in service” and “delay in rectification of credit information”.

·         A partner quietly pushes back on a pool, pointing to an unusual cluster of bureau disputes linked to your organisation code.

By then, no one is discussing the theory of dispute resolution.

They’re dealing with how it really behaves in practice.

 

The belief: “There is a clear dispute process, and it works roughly as designed”

If you strip away the careful language, the working belief inside many institutions sounds like this:

“If bureau data is wrong, the customer can raise a dispute.

The bureau routes it to the reporting lender.

We verify and correct it within defined timelines.

The system isn’t perfect, but it’s broadly fair.”

You see this belief baked into:

·         Internal training decks:

– “Customer may approach us or the bureau directly.”

– “All such cases are handled under the dispute resolution process.”

·         Policy documents:

– “We shall respond to credit information disputes within X days as per regulatory expectation.”

·         Board notes:

– “Bureau-related complaints form a small share of overall complaints. All are addressed within defined TATs.”

From a distance, it feels tidy:

·         Data is either right or wrong.

·         There is a channel to contest it.

·         Everyone has an interest in keeping information accurate.

The unspoken assumption underneath is simple:

“The dispute process is a clean exception path around an otherwise orderly system.

It may be slow at times, but it is fundamentally neutral.”

If you sit in the rooms where these disputes actually move, the contact centre floors, operations queues, bureau desks, and occasional escalation calls with bureaus and RBI, the picture is rougher.

Dispute resolution is not neutral.

It is shaped by incentives, bandwidth, and wiring.

And most senior forums only see the version that fits on one line: “Within TAT, no material issues.”

 

What dispute resolution looks like on paper vs in the wild

On paper, the ideal flow is simple:

1.      Customer spots an error (through a lender, bureau, or free report).

2.      Customer raises a dispute with bureau or lender.

3.      Bureau forwards the dispute to the reporting member.

4.      Lender verifies in core systems and relevant records.

5.      If data is wrong, lender sends corrected data to the bureau.

6.      Bureau updates the report; customer is notified.

Straightforward. Mutual interest in accuracy.

Neat arrows on a slide.

In practice, three frictions show up again and again.

1. No one owns the customer’s “full story” end-to-end

When a dispute is raised today, it travels through multiple systems:

·         A CRM or complaints module where the customer first logs the issue.

·         A bureau-dispute queue in operations, often with its own ID and TAT logic.

·         The core or LOS, where actual account behaviour lives.

·         The bureau’s own ticketing system, where the member’s response is recorded.

In one bank, an internal review mapped a single high-profile dispute:

·         Customer wrote to the bank on 18 April.

·         Complaint was logged on CRM the same day.

·         It was re-classified twice (“general complaint” → “credit card” → “credit bureau”).

·         The bureau desk got the ticket on 24 April.

·         The dispute from the bureau’s side for the same issue had a different reference number and arrived on 28 April.

·         Operations verified the internal account behaviour on 2 May.

·         The correction file to the bureau went in the next scheduled batch on 10 May.

·         The customer’s report actually reflected the change in mid-June.

If you asked, “What was our TAT?”, the answer depended on the system:

·         CRM: closed within the internal 30-day limit.

·         Operations: responded to bureau dispute within their own TAT.

·         Bureau: updated as per their service window.

If you asked, “Who owned the customer’s problem from discovery to resolution?”, nobody could answer without telling a long story.

The ideal assumes one path and one owner; actual practice breaks it into segments, queues, and metrics.

2. “Data is correct as per our records” becomes a shield

In theory, the lender is supposed to verify:

·         Did we report the right DPD?

·         Did we update closure correctly?

·         Is the write-off or settlement remark accurate?

·         Did restructuring or moratorium treatment follow agreed norms?

In many disputed cases, the internal account view is itself not straightforward:

·         There may have been re-aging, reversals, charge-offs, or manual adjustments over months.

·         Different systems (core, collections platform, legacy LOS) can show slightly different dates or codes.

·         Manual notes sit in email threads, not in structured fields.

Operational teams, under volume pressure, often fall back on the safest defensible statement:

“Data is correct as per our records.”

Sometimes that’s true.

Sometimes it means:

·         “The way we posted this is consistent with our own rules, even if the customer experiences it as unfair.”

·         “We don’t have the time or mandate to question whether our treatment will survive a more detailed regulatory or ombudsman review.”

On a dispute MIS, it gets coded as:

·         “Dispute not accepted – data found correct.”

·         “No correction required.”

On a credit review call six months later, when a regulator or ombudsman points out that three similar disputes were all rejected on the same grounds, the institution suddenly discovers that:

·         Their internal records do not fully explain the chosen treatment.

·         The customer’s version of the story is not obviously wrong.

·         A blanket “data correct as per records” doesn’t sound so solid when read aloud in an order.

The ideal assumes correctness is objective.

Reality is that “correct as per our records” is often a defensive position, not a neutral fact.

3. TAT metrics hide the cases that matter

Internal MIS on disputes tends to look like:

·         Total disputes received.

·         % resolved within defined TAT (15/30 days).

·         Broad break-up: “DPD related”, “closure/limit issues”, “ownership mismatch”.

On paper, you might see:

·         “Overall TAT compliance: 96% within 30 days.”

·         “No ageing disputes beyond 60 days.”

What this doesn’t show:

·         How many disputes have been closed with “no change” despite weak narratives.

·         How often the customer came back to complain again, now via RBI or social media.

·         Cases where the underlying issue was systemic (e.g., a product’s closure logic) but each dispute was treated as isolated.

In one NBFC, when someone finally pulled repeat complainers on bureau issues, they found:

·         A small segment of customers had raised 3+ contacts across channels for the same underlying dispute.

·         Their individual tickets were all “closed within TAT.”

·         Their bureau data was still not corrected.

The report that went to senior management each month had been telling a comfortable story:

·         “Low dispute volume, high TAT compliance.”

The small—and more important—story never made it into the slide:

·         “For customers who really push, the process feels circular and opaque. They don’t trust our closure.”

The ideal assumes that TAT and closure are good proxies for fairness.

Reality is that they mostly measure how quickly the organisation can move tickets through its internal pipes.

 

Why the ideal story survives in senior forums

If reality is this messy, why do we keep hearing:

“Dispute volume is low, TATs are fine, nothing systemic”?

Partly because of how dispute work is framed and where it lives.

Disputes are treated as service issues, not risk artefacts

In most organisations:

·         Customer Service / Operations owns credit bureau disputes.

·         Their dashboards focus on:

– Volume per channel

– Average TAT

– Ageing buckets

– Complaint codes

Risk and credit teams see bureau disputes as:

·         “Noise around the portfolio”,

·         Or something that will surface only when there is an inspection observation.

So when a CRO asks for dispute data, they often get:

·         A polite, aggregated view designed for service reporting, not for risk insight.

What doesn’t get discussed is:

·         Whether disputes cluster around specific products, branches, partners, or time periods.

·         Whether certain kinds of treatment (e.g., post-COVID restructures) are generating disproportionate dissatisfaction.

The topic stays on the service side of the house, even though it touches the core of the credit-information ecosystem.

The customer’s reality rarely shows up in the documents we read

A customer who has been wrongly marked settled or written off for months does not experience it as a technical issue.

They experience:

·         Being declined at another bank for a home loan or business facility.

·         Being told repeatedly to “wait 30 days” after “necessary corrections have been done”.

·         Getting canned bureau-dispute emails that do not match what they see in their report.

Internal documents reduce all of that to:

·         “Customer disputes DPD; found correct, closure sent.”

·         “Customer alleges closed loan is open; updated in next cycle, informed.”

The emotional reality of the dispute, the loss of face, the frustration at being bounced between lender and bureau, the sense that nobody is accountable, never appears in any pack.

So, in one room:

·         Leaders see a small box on a slide.

In another room:

·         The same leader might get a WhatsApp forward of a thread where their bank is being named and shamed for exactly the same process.

Because those two experiences are rarely connected, the belief that “our process is basically fine” survives.

No single metric captures “how often the system is plainly wrong”

There are measures for:

·         Complaint count.

·         TAT.

·         Ombudsman cases.

·         RBI escalations.

What’s missing is a simple, uncomfortable number:

·         “In how many disputes did we eventually acknowledge that the bureau data was actually wrong?”

And, equally critical:

·         “In how many ‘no change’ closures did ombudsman or RBI later tell us we were, in fact, wrong?”

That number requires:

·         Joining CRM, bureau desk, legal, and regulatory data.

·         Admitting that the institution’s first instinct was not always right.

Because no function is naturally incentivised to surface that, it stays buried.

 

How more experienced teams hold the “ideal vs reality” gap without drama

The institutions that seem calmer when bureau disputes land in inspection rooms don’t have perfect processes.

They just stopped telling themselves that the ideal and reality were the same thing.

A few patterns repeat.

They separate three realities: policy, process, and lived experience

Instead of saying “we have a dispute resolution process”, they do a simple mapping exercise:

·         Policy reality

– What do we claim we will do?

– TATs, checks, communication touchpoints.

·         Process reality

– How does a dispute actually flow through CRM, ops, bureau, and core?

– Where do hand-offs happen?

– How many different IDs does one dispute generate?

·         Lived experience

– What does a real customer see on their report and in their inbox at each step?

– How many times do they need to follow up?

– What are we telling them that later turns out to be untrue?

They don’t turn this into a big transformation project.

They pick a handful of actual disputes and walk them end-to-end.

In those walkthroughs, the myth of a neat, neutral process falls away quickly.

That alone changes how leaders talk about “ideal vs reality”.

They treat bureau disputes as a risk signal, not just a service metric

In one lender, the CRO insisted that bureau dispute MIS show up next to:

·         Decline rates due to bureau reasons.

·         Vintage performance in segments where disputes were concentrated.

That simple juxtaposition led to uncomfortable questions:

·         “Why are these specific branches or partners generating more disputes about ‘loan closed but showing open’?”

·         “Is there a connection between our treatment of charge-offs in product X and later disputes?”

·         “Are bureau-related escalations clustering around customers who are otherwise good risks?”

The message was clear:

·         These are not just service irritants.

·         They are telling us something about how our own credit information behaviour is perceived.

Once dispute data enters risk discussions, the ideal story starts losing its automatic privilege.

They are honest about where disputes are structurally hard to fix

Some issues are genuinely complex:

·         Legacy systems that did not capture certain fields properly.

·         Acquired portfolios where past posting practices are messy.

·         Historical treatments (e.g., COVID-era moratoria or restructures) that do not map neatly into bureau formats.

More mature teams don’t pretend that every dispute can be resolved cleanly within 30 days.

They do something else:

·         Identify categories of disputes where the root cause is structural.

·         Acknowledge, in internal notes, that these may require:

– Longer timelines,

– Manual intervention,

– Or even a conscious decision to give the benefit of the doubt to the borrower where evidence is ambiguous.

You sometimes see small, practical decisions like:

·         “For this narrow vintage of accounts, if bureau disputes arise on DPD in these months, we will bias towards correction unless we have strong proof otherwise.”

Is that perfect? No.

Is it closer to fairness than a blanket “data correct as per our records”? Yes.

 

A quieter way to think about “ideal vs reality” in dispute resolution

It is tempting to keep the neat story:

“There is a clear bureau dispute process.

Customers can raise issues.

We respond within regulatory timelines.

Hardly any cases reach the ombudsman.

Therefore, things broadly work as intended.”

If you stay with that, dispute resolution will remain:

·         A small box in a service slide,

·         An occasional regulatory observation,

·         A topic that feels marginal compared to credit growth, GNPA and recovery.

If you accept a more uncomfortable reality:

·         That between customer, lender, bureau and regulator, the dispute process is fragmented and often opaque,

·         That “correct as per our records” is sometimes a way to avoid revisiting past decisions,

·         And that TAT and closure rates say more about our internal efficiency than about customer fairness,

then the question changes shape.

It stops being:

“Do we have a compliant dispute process and decent TATs?”

and becomes:

“If we quietly sample the last year of bureau disputes,

how often would we still stand by our first response in front of an ombudsman or RBI –

and in how many cases would we, in hindsight, side with the customer instead?”

Most institutions don’t like the answer the first time they ask it.

The point is not to chase perfection.

It’s to make sure that the distance between the ideal and the reality is something you choose to live with, eyes open –

not something you only discover when someone outside your walls writes it into an order.