Many data leaks in Indian companies are not dramatic hacks. They are a sales executive forwarding the customer list to his personal Gmail before joining a competitor. An accountant copying salary data to a pen drive to work from home. A customer service agent pasting a complaint, complete with the customer's name, phone number and account details, into a public AI chatbot to draft a polite reply. Each act is small. Together they are how customer data walks out of the door.
Data loss prevention (DLP) is the set of policies and tools that stop this. Done badly, it blocks everything and people find workarounds within a week. Done well, it targets the few channels where sensitive data actually leaks.
Why does data loss prevention matter more in India now?
Two reasons. First, the Digital Personal Data Protection Rules, 2025 were notified on 14 November 2025, operationalising the DPDP Act, 2023, with an eighteen-month phased compliance period. Under the Act, the highest penalty, up to ₹250 crore, applies to a Data Fiduciary's failure to maintain reasonable security safeguards. Failing to notify the Data Protection Board or affected individuals of a personal data breach can attract up to ₹200 crore. When a breach happens, the Rules require affected individuals to be informed without delay, in plain language, explaining what happened and what is being done.
Second, CERT-In's April 2022 directions include data breach and data leak among the cyber incidents that body corporates must report to CERT-In within six hours of noticing them.
A leak through an employee's personal email is still a leak. DLP is one of the clearest ways to show that "reasonable security safeguards" were in place.
Where does data actually leak in Indian offices?
| Channel | Typical scenario | Risk level |
|---|---|---|
| Personal email | Forwarding files to Gmail "to work from home" or before resigning | High |
| USB pen drives and external disks | Copying folders for convenience or deliberately | High |
| WhatsApp (web and mobile) | Sending customer documents, price lists, screenshots of dashboards | High |
| Public AI chat tools | Pasting customer data, contracts or source code to get help | High and growing |
| Personal cloud storage | Uploading to a personal Drive or Dropbox | Medium to high |
| Over-shared links | "Anyone with the link" sharing on company drives | Medium |
| Printouts and phone photos | Photographing screens | Medium, hard to stop technically |
| Departing employees' devices | Laptops not returned or not wiped | Medium |
The AI chat problem deserves its own mention
Public generative AI tools are now one of the easiest ways for well-meaning staff to leak confidential material. In 2023 Samsung imposed a temporary ban on staff use of generative AI tools after an apparently accidental leak of internal source code linked to staff use of ChatGPT, according to reporting on an internal memo. The memo cited growing concerns about security risks presented by generative AI. Banning tools outright tends to push usage to personal phones. The better answer is an approved, company-controlled AI tool plus DLP rules that stop sensitive data reaching unapproved ones.
A phased DLP rollout that does not kill productivity
Phase 1: Know what you are protecting (weeks 1 to 3)
- List your sensitive data types: customer personal data (names, phone numbers, Aadhaar, PAN, bank details), employee data, pricing, designs, source code, contracts.
- Find where each lives: CRM, ERP, shared drives, email, laptops.
- Classify simply: Public, Internal, Confidential, Restricted. Four labels people can remember beat ten they cannot.
Phase 2: Fix the basics before buying a DLP product (weeks 3 to 6)
- Access control. Most leaks are possible because too many people can see too much. Remove access that is not needed for the job.
- Company accounts only. Business email, storage and messaging on company-managed accounts with two-factor authentication.
- Sharing settings. Turn off "anyone with the link" sharing by default on company drives; allow external sharing only for named domains.
- Offboarding checklist. Same-day account disablement, device return, and a review of recent large downloads or forwards.
- Provide an approved AI tool with a clear policy on what may be pasted into it.
Phase 3: Monitor first, then block (weeks 6 to 12)
- Switch on DLP rules in audit mode: email DLP in your mail platform, endpoint agents for USB and uploads, browser controls for unapproved AI and file-sharing sites.
- Watch for two to four weeks. You will learn which alerts are real and which processes legitimately move data.
- Tune rules to cut false positives. A DLP system that alerts on every invoice will be ignored.
Phase 4: Enforce on the highest-risk channels
- Block USB storage by default, with approved encrypted drives for people who need them.
- Block or warn on emails with Restricted data to personal domains.
- Block uploads of Restricted data to unapproved AI and cloud sites.
- Use "warn and justify" pop-ups for medium-risk actions. They stop most accidents without blocking work.
Phase 5: Review and improve continuously
- Monthly review of incidents and exceptions.
- Quarterly access review for sensitive systems.
- Short, real-example training. Show staff the exact pop-up they will see and why it exists.
What DLP cannot do
Be realistic. DLP will not stop a determined insider photographing a screen with a personal phone, and it will not fix bad access control. It works best as one layer alongside least-privilege access, logging, and a culture where people understand why customer data is protected. Logs matter too: CERT-In's directions require logs of ICT systems to be kept for a rolling 180 days within India, and those logs are what let you investigate a suspected leak.
How do you handle the employee experience?
- Tell staff what is monitored and why. Secret monitoring damages trust and may create legal issues of its own.
- Give a fast exception path. If a legitimate task is blocked, the manager should be able to approve it the same day.
- Measure friction: the number of exception requests and how quickly they are resolved.
A ten-minute DLP readiness check
- Can you list the five most sensitive data sets in the company and who can access each?
- Do all staff use company-managed email and storage accounts with two-factor authentication?
- Is USB storage controlled on machines that handle customer data?
- Is there an approved AI tool, and a written rule on what may be pasted into public ones?
- Are "anyone with the link" shares switched off by default?
- Is there a same-day offboarding checklist that someone actually follows?
- Would you know within a day if a departing employee forwarded the customer list to a personal address?
Every "no" is a gap to close before buying any DLP software.
How Cehpoint can help
We help Indian companies classify data, tighten access, roll out DLP in monitor-then-enforce phases and prepare for DPDP obligations. Our cyber security and investigation service covers the technical controls, and our DPDP compliance service covers the legal and process side. If staff need AI tools that keep data inside your control, see private LLM deployment.