Self-assessment
Score yourself
Answer each question honestly. Yes = 2, Partly = 1, No = 0. Keep a running total.
Your score
0/ 24
0 of 12 answered
Not scored yet
The baseline
0 / 601 Is all personal data encrypted at rest and in transit, including at vendors?
02 Are encryption keys managed and rotated?
03 Do backups work, and have you tested a restore recently enough to trust it?
The access
0 / 604 Is access role-based, least privilege, with segregation of duties, so no single account can both move and approve?
05 Is every privileged account on multi-factor authentication?
06 Are there zero shared admin accounts anywhere?
07 Does joiner-mover-leaver handling actually remove access when people change roles or leave?
The logs
0 / 608 Are audit logs tamper-evident, showing who accessed or changed personal data?
09 Are logs kept long enough to support an investigation, and does a named person actually review them?
The proof
0 / 610 Has every system holding personal data had vulnerability assessment and penetration testing before go-live and at intervals since?
11 Does every finding have an owner, a due date and a closure clock, with none sitting open through the next audit?
12 Can you produce, for one customer's record, the log of who accessed it last month?
| Score | Where you stand | What to do next |
|---|---|---|
| 0 to 8 | Exposed | Protection is asserted, not real. Encryption and privileged access first; they carry the most risk per rupee. |
| 9 to 16 | Halfway | Controls exist but evidence does not. The log review habit and the finding-closure clock are the usual gaps. |
| 17 to 21 | Strong base | Run the three-artefact test below on your most sensitive system. |
| 22 to 24 | Leading | Keep the baseline mapped as sectoral frameworks change, and watch cloud posture continuously. |
Any single "No" is a named gap, whatever the total says. Nothing leaves your browser.
What the duty expects
The Act and Rules name the expectation: measures like encryption, obfuscation or masking, access control, and logging and backups so breaches can be detected and investigated. "Reasonable" scales with the sensitivity of the data, which is where your classification work feeds in.
Where sectoral frameworks apply, they set the floor: the banking regulator's cyber security framework for the entities it regulates, and the national incident-reporting directions for everyone. The DPDP duty stacks on top; it does not replace them.
The framework: four moves
Move 1: Build the control baseline. Map each control to the safeguards duty and to any sectoral framework you answer to. One baseline, many obligations. The control set: encryption at rest and in transit, role-based access with least privilege and segregation of duties, multi-factor authentication for privileged access, tamper-evident audit logs, tested backups, and vulnerability assessment with penetration testing.
Move 2: Design the access model. Roles, privileges, joiner-mover-leaver handling, and the approval path for elevated access. Shared admin accounts are an audit finding and a breach waiting.
Move 3: Set the logging standard and test the backups. What is logged, where logs live, how long they are kept, who reviews them. Backups exist so breaches and losses can be investigated and recovered from; test restores, not just backup jobs.
Move 4: Run VAPT with a closure clock. Scope and frequency per system criticality. Every finding gets an owner and a due date; the tracker is the evidence. Where a security operations centre exists, plan how privacy-relevant events feed it.
Template: the control checklist
One row per control per system. Copy this and fill it in.
| Control area | The standard | System | Status | Evidence | Last evidence | Owner |
|---|---|---|---|---|---|---|
| Encryption | At rest and in transit, keys managed and rotated | Example: lending app | Met | Config extract, key rotation log | Security | |
| Access | Role-based, least privilege, MFA for privileged access | Example: lending app | Partially met | Access review; two service accounts pending MFA | Security | |
| VAPT | Before go-live and at intervals, findings closed on a clock | Example: lending app | Met | VAPT report, closure tracker | Security | |
The downloadable file adds the full control areas table.
What good looks like at the end
You should have: a control baseline mapped to the safeguards duty, an access model, a logging standard, a running VAPT programme, and a remediation tracker that shows findings closing.
The live test: pick one system holding personal data. Produce its last access review, its last VAPT report with closure status, and the log of who accessed a sample record. If any of the three cannot be produced, the safeguard is asserted, not real.
Where enterprises go wrong
- A certificate on the wall treated as a control.
- Shared admin accounts.
- Logs that exist but nobody reviews.
- VAPT findings that stay open through the next audit.
- Confusing the privacy control floor with the whole security programme. This page is the floor the Act demands; the wider programme goes further. The cyber safety page covers incident prevention more broadly.
The decisions to make
- VAPT scope and frequency, per system criticality.
- Whether privacy tooling feeds the security operations centre.
- How cloud posture is monitored, and by whom.
- Which sectoral frameworks stack onto your baseline.
Next step
Making privacy work inside your systems, where these controls meet your architecture. If you need help with any of this, write to us and we will help.