Quick answer
The DPDP Act does not mandate a formal record of processing, but almost every obligation depends on one. Rule 3 needs an itemised list of personal data, Rule 7 needs fast breach scoping, Rule 8 needs retention periods and section 11 needs a summary of data and the fiduciaries and processors it is shared with. A data inventory built system by system, before 13 May 2027, is the foundation.
- No DPDP provision names a 'record of processing', yet notices, breach reports and access requests all depend on one.
- Start from systems and data flows, not from policies; interviews alone miss shadow data.
- Record purpose, lawful ground, data categories, location, processors, retention and owner for every processing activity.
- Tag data about children, Indian versus foreign Data Principals, and Third Schedule purposes explicitly.
- Automate discovery where possible and tie the inventory to release processes so it stays current.
Who this is for: Privacy and compliance programme leads, Data engineers and platform teams, Security architects, Product managers, Internal audit
Why does DPDP compliance start with a data map?
The DPDP Act and Rules do not contain a GDPR Article 30-style duty to keep records of processing activities. But try meeting the obligations without one:
| Obligation | What it needs from the inventory |
|---|---|
| Rule 3(b)(i) notice | An itemised description of the personal data you collect |
| Rule 3(b)(ii) notice | The specified purposes and the services they enable |
| Section 6 / 7 | Which lawful ground supports each purpose |
| Rule 6 security | Where personal data lives, who can access it, how it is protected |
| Rule 7 breach | Fast answers on data categories, number of people and location |
| Rule 8(1) / 8(3) | Retention periods and when to erase |
| Section 11 access | A summary of data processed and the fiduciaries and processors it was shared with |
| Section 9 children | Which datasets contain under-18 users |
| Section 17(1)(d) | Which datasets relate to Data Principals outside India |
An inventory is the shared source for legal, security and engineering. Without it, each team builds its own partial picture, and the gaps surface during a breach or a Board inquiry.
Steps 1–2: set scope and assign owners
Step 1: define the unit of the inventory
Inventory at two levels:
- Systems (assets): databases, SaaS tools, data warehouses, object stores, log pipelines, spreadsheets on shared drives, and AI components such as vector stores and prompt logs.
- Processing activities: business-level uses such as "customer onboarding", "payroll" or "fraud scoring", each of which touches several systems.
Linking the two lets you answer both "what does this system hold?" and "where does onboarding data go?".
Step 2: assign owners
Every system and every processing activity needs a named owner who can answer questions and approve changes. For most organisations:
- business function heads own processing activities;
- engineering or IT owns systems;
- the privacy lead owns the inventory format and quality.
Set a scope boundary for the first pass: all production systems holding personal data of customers, employees, candidates and vendors' staff. Defer archives and analytics sandboxes to the second pass, but list them so they are not forgotten.
Step 3: discover where personal data actually lives
Interviews tell you what people think the systems hold. Discovery tells you what they actually hold. Use both.
Discovery sources
- Cloud inventories: list storage buckets, databases, queues and data warehouses across every account and region.
- Schema scanning: scan table and column names and sample values for identifiers such as mobile numbers, email addresses, PAN or Aadhaar-like patterns, addresses and dates of birth.
- SaaS admin consoles and SSO logs: which tools staff actually sign in to.
- Expense and procurement records: subscriptions that never went through IT.
- Log pipelines and observability: application logs and traces often contain request bodies with personal data. See LLM logging and observability.
- Data flows: API gateway routes, ETL jobs, event streams and webhooks to vendors.
Shadow data to look for
Spreadsheet exports in personal drives, data copied into staging environments, CSVs attached to support tickets, WhatsApp groups used for customer coordination, and datasets uploaded to AI tools for analysis.
Step 4: record the right fields
For each processing activity, capture at least the following. A spreadsheet works for small organisations; a data catalogue or privacy tool scales better.
| Field | Example |
|---|---|
| Activity and owner | Customer onboarding, Head of Growth |
| Data Principal categories | Customers; children (yes/no); outside India (yes/no) |
| Personal data items | Name, mobile, email, PAN, selfie, device ID |
| Purpose and service enabled | Open and verify an account |
| Lawful ground | Consent / s.7 use (which one) / legal obligation |
| Source | Direct from user; partner; public |
| Systems | Postgres users, S3 kyc-docs, CRM |
| Processors and recipients | KYC vendor, cloud host, SMS gateway |
| Locations | Mumbai region; vendor in Singapore |
| Retention | Account life; logs one year (Rule 8(3)); other law |
| Security controls | Encryption at rest, role-based access, audit logs |
| Third Schedule class | E-commerce / gaming / social media threshold met? |
Keep the data item list at a level of detail you would be comfortable printing in the Rule 3 notice, because that is where it will end up.
Step 5: draw data flows and flag risks
A table does not show movement. For each major activity, draw a simple flow: collection point → services → storage → processors → outputs → deletion.
Flag these on every flow
- Children's data: any path where a user might be under 18 triggers section 9 and Rule 10.
- Cross-border hops: vendors or regions outside India. The Act takes a negative-list approach under section 16, but sectoral rules such as RBI's payment data storage requirements may still require local storage. See data localisation for AI workloads.
- Purpose drift: data collected for one purpose feeding analytics, personalisation or model training.
- Excess collection: fields that no purpose needs.
- Unbounded retention: data with no deletion job.
- Missing contracts: processors without a Rule 6(1)(f) contract.
Using the flows
The flows become the basis for the Rule 7 breach scoping question "what could be affected?", for security architecture reviews, and, for Significant Data Fiduciaries once notified, for the annual DPIA required by Rule 13(1).
Step 6: keep the inventory current
An inventory that is accurate on the day it is finished and wrong six months later is a liability. Build maintenance into delivery.
- Release gate: any change that adds a personal data field, a new vendor or a new purpose requires an inventory update before it ships. Put a question in the pull request template or design review.
- Procurement gate: no new SaaS tool is approved until its inventory entry and processor contract exist.
- Automated drift detection: rerun schema and storage scans monthly and alert on new tables or buckets that match personal data patterns.
- Quarterly owner attestation: owners confirm their entries are correct.
- Generate outputs from the inventory: the itemised list in the notice, the processor list, retention jobs and the access request export should read from the same source.
Before 13 May 2027
Aim for a complete first pass of production systems by February 2027, so the following months can be spent fixing what the inventory reveals. See the compliance roadmap to May 2027.
Further reading on suhasbhairav.com
- AI Use CasesAI Agent Use Case: Gdpr Consultants Using Website Forms and Data Flows to Identify Privacy Risks
- BlogAI agents for GDPR and CCPA compliance
- AI AutomationAI Automation for Data Governance: How It Works + Example
- BlogEnterprise GenAI data privacy and security guardrails
- AI LabBatuk and Document sensitivity classification
Before 13 May 2027
Compliance checklist
Define inventory units and owners
Inventory both systems and processing activities, each with a named owner.
Section 8(1)
Run automated discovery
Scan cloud storage, databases and logs for personal data patterns across all accounts.
Rule 6(1)(c)
Find shadow data
Check personal drives, staging copies, support attachments and AI tool uploads.
Rule 6
Record lawful ground per purpose
Note consent or the specific section 7 legitimate use for every activity.
Sections 6 and 7
Tag children's and foreign Data Principals' data
Flag datasets with under-18 users and those within section 17(1)(d).
Section 9; s.17(1)(d)
Record processors and locations
List every vendor, its role, processing locations and contract status.
Rule 6(1)(f)
Set retention for each dataset
Assign retention periods reflecting Rule 8(3) one-year logs and any longer legal requirement.
Rule 8(3)
Generate the notice data list from the inventory
Feed the Rule 3 itemised description directly from inventory records.
Rule 3(b)(i)
Add release and procurement gates
Block launches and vendor onboarding without inventory updates.
Rule 6(1)(g)
Answers
Frequently asked questions
Does the DPDP Act require a record of processing activities?
Does the DPDP Act require a record of processing activities?
Not expressly. Unlike GDPR Article 30, neither the DPDP Act nor the Rules name a record of processing. In practice you need one to give an itemised Rule 3 notice, scope breaches under Rule 7, apply retention under Rule 8 and answer access requests under section 11.
What fields should a DPDP data inventory include?
What fields should a DPDP data inventory include?
At minimum: processing activity and owner, Data Principal categories including children and non-Indian individuals, personal data items, purpose and service enabled, lawful ground, source, systems, processors and recipients, locations, retention period and security controls. Add Third Schedule status if you are a large e-commerce, gaming or social media platform.
How long does data mapping take?
How long does data mapping take?
It depends on the number of systems and how well documented they are. A small company with a handful of SaaS tools can produce a first pass in days. Large organisations with many products, regions and legacy systems usually need several months, which is why starting by early 2027 matters.
Can I use a spreadsheet for the DPDP inventory?
Can I use a spreadsheet for the DPDP inventory?
Yes, for smaller organisations. The law cares about accuracy, not tooling. A spreadsheet with the right fields, owners and a review routine is fine. As systems and vendors multiply, a data catalogue or privacy tool that integrates with discovery scans becomes easier to keep current.
How does the inventory help with access requests?
How does the inventory help with access requests?
Section 11 gives Data Principals the right to a summary of personal data processed and the processing activities, and the identities of other fiduciaries and processors it was shared with. An inventory linking each system and vendor to the person's identifiers lets you assemble that summary quickly and consistently.
Should AI systems be in the data inventory?
Should AI systems be in the data inventory?
Yes. Prompt logs, retrieval indexes, vector stores, fine-tuning datasets, evaluation sets and model provider APIs all process personal data. Record them as systems with their own purposes, processors, locations and retention, and flag any use of customer data for training as a distinct processing activity.
Legal references and sources
- Digital Personal Data Protection Rules, 2025 — Rule 3(b)(i) (official text)
- Digital Personal Data Protection Act, 2023 — Section 11 (official text)
- Digital Personal Data Protection Rules, 2025 — Rules 7 and 8 (official text)
- Digital Personal Data Protection Rules, 2025 — Rule 6 (official text)
This guide explains the law for engineering and business teams. It is not legal advice; confirm how it applies to your organisation with qualified counsel, and check the Gazette for later notifications or amendments.
About the author

Suhas Bhairav
Systems architect and applied AI practitioner who designs production AI systems for Indian and global companies, with a focus on privacy-by-design under the DPDP Act and the EU AI Act.
Related guides
- Templates & playbooksDPDP Privacy Notice Template (Rule 3 Compliant)A copy-ready DPDP privacy notice template mapped clause by clause to Rule 3 and section 5, with a short consent notice, a full notice and implementation guidance.
- Templates & playbooksDPDP Audit Checklist: 60 Controls to TestSixty testable DPDP controls grouped by governance, notice and consent, children, security, breach, retention, rights and vendors, each mapped to the Act or the 2025 Rules.
- DPDP Rules 2025DPDP Data Retention and Erasure Under Rule 8 and the Third ScheduleRule 8 sets when personal data must be erased: three-year dormancy erasure for large e-commerce, gaming and social media platforms, a 48-hour warning, and a one-year retention floor for everyone.
- Templates & playbooksDPDP Breach Response Playbook: The First 72 HoursAn hour-by-hour DPDP breach response playbook for the first 72 hours: triage, CERT-In's 6-hour report, Rule 7 intimations to users and the Board, and evidence.
- DPDP Rules 2025Data Principal Rights Under DPDP Rule 14The DPDP Act grants access, correction, erasure, grievance and nomination rights. Rule 14 sets how fiduciaries publish request channels, identifiers and a grievance period of up to 90 days.
- Templates & playbooksDPDP Compliance Roadmap to May 2027A month-by-month DPDP compliance plan from December 2026 to 13 May 2027, when Rules 3, 5–16, 22 and 23 take effect: inventory, notices, security, breach, retention and rights.