Ask a new agent what “data security” means and you will often hear something about hackers and passwords. That is a reasonable first guess and a misleading one. Much of the risk in a BPO has nothing to do with sophisticated attackers. It comes from ordinary people doing ordinary things slightly wrongly. Four beliefs are worth taking apart.
Belief one: leaks come from outside
The reality is far more mundane. A customer’s address photographed on a phone for “later reference”. A screenshot sent to a friend on a messaging app to ask whether a transaction looks odd. A notepad with account numbers left on the desk at shift end. A supervisor’s login shared because the system was slow. Each of these is a leak. None involves a hacker.
This is why production floors commonly have clean-desk rules, restrictions on phones and paper, and monitored or locked-down workstations. They can look heavy-handed, but they target the realistic route: insiders, careless or curious or occasionally dishonest. A related point is access. Agents should be able to see what their task needs, and no more. Looking up a celebrity customer or an acquaintance “just to see” is a violation even if nothing leaves the screen, and well-run organisations log access and review it.
Belief two: personal data means financial data
Card numbers get the attention because the harm is obvious. But a name plus a phone number plus a delivery address is already enough to enable a convincing fraud call, and a health-insurance claim reason can embarrass someone for years. Under data protection laws, “personal data” is defined broadly: information relating to an identifiable person. The EU’s GDPR uses that wording, and India’s Digital Personal Data Protection Act, 2023 (DPDP Act) defines it as data about an individual who is identifiable by or in relation to that data. Some laws also carve out stricter categories (the GDPR has “special category” data such as health data), while the DPDP Act does not use that same split, so check which law applies to the process.
The practical rule for agents: treat anything that identifies or describes a customer as confidential by default, and use it only for the purpose of the call.
Belief three: if the caller knows some details, they must be the customer
Callers who know a name, an address and a date of birth are not proven genuine. Those details are often available from earlier leaks, or from a family member or an ex-partner. Verification exists to prove the caller can reach information that only the customer should have, such as a one-time password sent to the registered number, not simply to recite facts that might be public. A polite refusal to proceed without verification protects the customer, and in a good process it is also protected by the agent’s own quality score.
Sympathy is the lever fraudsters pull. “My mother is in hospital and I need her account balance” should meet the same procedure, delivered kindly, with an alternative path such as asking the account holder to authorise the caller formally.
Belief four: compliance is the client’s concern
Frameworks differ, but a pattern is common. The organisation that decides why and how data is used (the bank, the airline) has the primary legal responsibility. In GDPR language that is the controller; in the DPDP Act it is the data fiduciary. The BPO that handles data on its behalf is the processor, and works under a contract that sets out what it may do. A processor is not a bystander. Under GDPR a processor has direct obligations, including security measures and acting on the controller’s documented instructions. Under the DPDP Act the fiduciary remains accountable for processing done on its behalf, and passes obligations to processors by contract; the commercial reality is that a BPO that mishandles data stands to lose the contract, face indemnity claims and damage its reputation.
The DPDP Act’s rules have been coming into force in phases, so any team handling Indian customers’ data should check the current status and what its client contracts now require, rather than relying on a general summary like this one.
For European customers’ data handled from India, GDPR can apply because of whose data it is and who the client is, even though the people processing it sit elsewhere. Typically that means contractual clauses, restrictions on where data may be stored or accessed, and training records. The floor-level translation is simple: do exactly what the process says with the data, and nothing else.
What replaces the beliefs
A better mental model is three questions asked before touching any customer information. Do I need this for the task in front of me? Have I confirmed I am dealing with the right person? Is the place I am about to put this information (a chat window, a note, my notebook, a personal phone) an approved one? Most incidents fail one of these three tests.
Reporting matters too. If something goes wrong, such as an email sent to the wrong person, tell the supervisor straight away. Regulations and contracts often impose tight deadlines for notifying breaches, and those clocks start when the organisation becomes aware. An agent who hides a mistake for two days has turned a small incident into a serious one.
Leave a Reply