Omnichannel Banking Needs a Single View of the Complaint, Not Just the Customer

Omnichannel banking has made it easier for customers to move between mobile apps, email, WhatsApp, branches, and other service channels. The systems supporting those channels, however, still need a reliable way to determine when multiple interactions belong to the same unresolved issue. 

That distinction matters in grievance redressal. 

Suppose a customer reports a failed UPI transaction through the banking application and follows up later through email and WhatsApp. If each channel independently creates a case, one underlying grievance can generate multiple tickets. 

The immediate consequence is duplicated investigation. Different agents may retrieve the same transaction details, review the same account history, and work towards separate resolutions. 

There is also a regulatory dimension. Under the RBI Integrated Ombudsman Scheme, complaint handling is governed by defined timelines. Duplicate cases can therefore create multiple operational records and SLA clocks around an issue that, from the customer’s perspective, has never changed.

Solving this requires more than maintaining a unified customer profile. 

A bank may already know that three messages came from the same customer and still need to determine whether those messages concern one complaint, related complaints, or entirely different issues. 

That requires identity resolution and semantic deduplication to work together. 

An explicit complaint reference provides the strongest connection. Messages belonging to an existing channel thread provide another. Where neither exists, semantic similarity can help determine whether the intent of a new interaction matches an open complaint. Ambiguous matches can be linked for review, while sufficiently different issues continue as separate cases. 

Auditability matters here too. If a system decides that two grievances are the same, the bank needs a record of how that conclusion was reached. Automated consolidation therefore needs explicit rules, confidence thresholds, and supervisory intervention for uncertain cases. 

The architectural implication is important: the complaint needs its own identity independent of the channel carrying it. 

We applied this principle while developing Samadhan, our grievance-redressal accelerator. Its ingestion layer normalises interactions from multiple channels before identity resolution and deduplication determine how they relate to existing cases. In the project baseline, this allowed a customer issue represented by three tickets and three SLA clocks to be consolidated into one canonical ticket and one clock. 

As banks add more service channels, that connective layer will become increasingly important. Omnichannel service works best when the bank can preserve continuity of the issue as effectively as it preserves continuity of the customer.