Free template

DPDP data mapping template.

Most DPDP duties depend on knowing where personal data lives. This template gives you the fields to record for each system, and a second sheet for AI tools. Copy the layout or download both as spreadsheets.

16 map fields11 AI fields
Data map

One row per system

Data and purposeItemised fields and why you hold them.
Vendors and flowsProcessors, copies and exports.
RetentionHow long, and how deletion happens.
AI tools sheetTraining, retention, stored and processed where.
GapsThe start of your gap register.
Two sheets, CSV downloadNothing stored or sent
For the 13 May 2027 start

Get the template.

The core DPDP duties start on 13 May 2027, by our count. Each file has the headings, one illustrative example row and ten blank rows. The example is made up, not a real company. For why the map comes first, read why DPDP readiness starts with a data map. Once the map is done, work through the DPDP technical readiness checklist.

Nothing is stored or sent. The files are made in your browser and open in Excel, Numbers or Google Sheets.

How to read the fields

The Act and Rules do not prescribe a data map. The fields and examples are our suggestions, chosen so each DPDP duty has the answer it needs. The citations show which duty a field serves.

This page is general information, not legal advice. Mono is not a law firm, an auditor or a certification body.

Sheet 1: the data map.

One row per system, or per dataset inside a system. The last field is the point: once the rows are filled in, the gaps tend to show themselves.

1. System and dataset

The tool and the specific data in it. Split a system into datasets when purpose or retention differs.

Example: Support desk: tickets and attachments

Needed for: Everything else hangs off this row.

2. Data categories

An itemised list of the personal data fields. Name them one by one.

Example: Name, email, phone, order ID, free-text messages, uploaded files

Needed for: The notice, which Rule 3 says must itemise the data (section 5).

3. Data Principals

Whose data it is: customers, employees, candidates, vendors’ staff.

Example: Customers

Needed for: Rights requests and breach notices to the right people (sections 11 and 12; Rule 7).

4. Purpose

Why you hold it, in the words you used in the notice.

Example: Resolving customer support requests

Needed for: The notice and consent (sections 5 and 6), and deletion once the purpose is served (section 8(7)).

5. Basis

Consent, or the specific legitimate use in section 7. DPDP has no general legitimate-interests or contract basis.

Example: Consent at sign-up; legal review pending

Needed for: Section 4. Mark rows where you are unsure and ask your lawyer.

6. Source

Where the data comes from.

Example: Web form, email, in-app chat

Needed for: Knowing which notice and consent screen covers it.

7. Owner

One named person who answers for the row.

Example: Head of support

Needed for: Someone to keep the row current and to act when it has a gap.

8. Location and region

Where it is stored and processed, by vendor and region.

Example: Vendor cloud, Singapore region

Needed for: Section 16 and any sector storage rule, such as RBI’s for payment data.

9. Processors and contract

Each vendor that receives the data, and whether you hold a contract with it.

Example: Support desk vendor; its AI reply feature, which is on; contract signed

Needed for: Sections 8(1), 8(2) and 11(1), which cover the processors you shared data with.

10. Flows out

Where the data goes next, including exports, the warehouse and any country outside India.

Example: Nightly export to the data warehouse

Needed for: Erasure that reaches every copy, and breach scoping.

11. Retention and deletion

How long each kind of data is kept, and how deletion happens, including in backups.

Example: No limit set; manual deletion only

Needed for: Section 8(7) and Rule 8. The Rules do not spell out backup erasure, so document your backup expiry.

12. Identifier

How you find one person’s data in this system.

Example: Email address or customer ID

Needed for: Access, correction and erasure requests (sections 11 and 12).

13. Safeguards

Encryption, access control, logging and backups for this store.

Example: SSO and MFA on; audit log kept 1 year

Needed for: Rule 6, including keeping logs for at least one year (Rules 6(1)(e) and 8(3)).

14. Children’s data

Whether anyone under 18 is in the data, and whether the system tracks or monitors behaviour.

Example: No. Age gate at sign-up

Needed for: Section 9: verifiable parental consent before processing, and no tracking, behavioural monitoring or targeted advertising directed at children unless an exemption applies.

15. AI features

Any AI feature inside this system or fed from it. Point to a row on the AI tools sheet.

Example: Reply suggestions: see AI tools row 3

Needed for: Keeps AI tools from sitting outside the map.

16. Gaps and actions

What the row shows is missing, with an owner and a date.

Example: No retention period; warehouse copy not covered by erasure

Needed for: This column becomes your gap register.

How to fill it in.

No single source is complete. Interviews tell you what people believe happens, and settings and scans tell you what does. You need both. We set out each step in the guide to building a data map in practice.

  1. Step 1

    List the systems

    Start from your single sign-on dashboard, software spend, cloud accounts, integrations and the SDKs in your app builds.

  2. Step 2

    Interview the owners

    Ask what goes in, why, who can see it, where it goes next and when it is deleted. Include HR, finance, sales and support.

  3. Step 3

    Check the data itself

    Read schemas and sample content, with approval. Where an interview and a setting disagree, trust the setting and write down the gap.

  4. Step 4

    Draw one person’s journey

    Follow a customer, and an employee, from collection to every copy and every vendor. Mark where data leaves India.

  5. Step 5

    Keep it current

    Give the map an owner. Update it when a vendor, SDK, AI feature or personal data field is added, or a vendor changes its terms.

Sheet 2: AI tools.

Prompts, uploads and transcripts hold personal data, and AI features are often switched on inside tools you already map. You stay responsible for processing a vendor does for you (section 8(1)), so each AI tool needs its own row. If you do not know which tools your staff use, a shadow AI audit finds them, and DPDP and AI tools covers what to change.

1. Tool and feature

The product and the exact feature, because the same vendor can have a safe plan and an unsafe one.

Example: Meeting recorder, summary feature

Needed for: A row per feature, not per vendor.

2. Where it sits

A stand-alone assistant, an AI feature inside a SaaS tool, an API call in your own product, a coding assistant or a meeting recorder.

Example: Feature inside the CRM

Needed for: Features switched on inside existing tools are the ones missed most often.

3. Account and plan

Work account on a business plan, or a personal or free account.

Example: Business plan, SSO

Needed for: A personal account has no contract with you behind it.

4. Data that goes in

Prompts, uploaded files, transcripts and connected sources, with the personal data fields in them and whose they are.

Example: Call transcripts: customer name, phone, order details

Needed for: The same itemised view as the main sheet.

5. Training setting

Whether the vendor trains on your content, the setting you chose, and the date you checked.

Example: Not used for training on this plan; checked 20 Nov 2026

Needed for: Vendor terms differ by plan and change over time.

6. Retention

How long chat history, files and logs are kept, including any abuse-monitoring logs, and whether zero retention is available to you.

Example: History 30 days; vendor logs per its terms

Needed for: Deletion duties reach processors (section 8(7)).

7. Stored where, processed where

Two separate answers: where data sits at rest, and where the model runs on it.

Example: Stored: India. Processed: US or EU

Needed for: Section 16, and the transfer disclosure in your vendor list.

8. Model provider and sub-processors

Who actually runs the model, which may not be the vendor you pay.

Example: Vendor X, which calls provider Y

Needed for: Section 11(1) asks for every processor you shared data with.

9. Contract

The agreement or data processing terms you hold, and the review date.

Example: DPA signed; review due in 12 months

Needed for: Section 8(2): a valid contract with each processor.

10. Output and decisions

Where output is stored, who checks it, and whether it feeds a decision about a person.

Example: Stored in the CRM; agent reviews each reply

Needed for: Output can be personal data too, and decisions about people need a named reviewer.

11. Owner and last checked

The named person, and the date someone last read the vendor’s terms.

Example: Head of IT; 20 Nov 2026

Needed for: A row older than the vendor’s last terms change is a guess.

Stored where, processed where.

AI vendors often give two different answers. Where data is stored at rest is one question. Where the model runs on it is another. DPDP does not require data to stay in India, but hosting in India avoids future section 16 restrictions and sector rules. These are examples, checked on 20 November 2026. Vendors change plans often, so read the current terms before you record a region.

ChatGPT Enterprise and Edu

India is available for storage at rest. Inference residency covers only the EEA and Switzerland, the US and the UAE, not India. So a workspace can store data in India while the model still runs elsewhere.

Claude API, first-party

The inference_geo setting takes “global” or “us”, and workspace geography is US only. There is no India option. By default, API data is deleted within 30 days and is not used for training without express permission.

Azure OpenAI

Microsoft says data is not available to OpenAI and is not used for training. Standard deployments process data in the geography you choose, Global deployments may process it anywhere, and DataZone deployments within the US or EU zone. Stored data stays in the resource’s geography.

Claude on Amazon Bedrock

Since 29 September 2026, Claude Opus 5, Sonnet 5 and Haiku 4.5 are available through an India inference profile that routes only between the ap-south-1 and ap-south-2 regions. Check the model you use is in that profile before you record India as the processing region.

Microsoft 365 Copilot

Microsoft says prompts, responses and Graph data are not used to train foundation models. Outside the EU, queries may be processed in the US, the EU or other regions. Microsoft has announced in-country processing for India, so confirm what applies to your tenant.

Sources: OpenAI help centre, data residency and inference residency for ChatGPT. Claude Platform documentation, data residency. AWS Machine Learning Blog, Amazon Bedrock India cross-region inference. Microsoft Learn, Azure OpenAI data privacy. Microsoft Learn, Microsoft 365 Copilot privacy.

Where Mono fits in.

Data mapping is the first part of our DPDP gap assessment. We build the inventory and flow diagrams with read-only access, including SaaS tools, vendors and AI tools, then compare each system with the Act and Rules and give you a gap register and a plan. When the gaps are clear, we build the fixes through DPDP implementation. Mono does not certify you, audit you or give legal advice. Our partner law firm gives the written legal opinion.

Data mapping FAQs.

Does the DPDP Act say I must keep a data map?

No. The Act and the Rules never use the term. But several duties only work if you know where personal data sits: an itemised notice, access requests that name the processors you shared data with, erasure that reaches those processors, breach reports and the Rule 6 safeguards. A data map is the practical way to meet them.

How detailed should one row be?

One row per system, or per dataset inside a system when its purpose or retention differs. List the actual fields, not “customer data”, because Rule 3 asks for an itemised description of the personal data in the notice. If you cannot say how to find one person in that system, the row is not finished.

Why do AI tools get a separate sheet?

Prompts, uploads and call transcripts are personal data when they contain it. The questions for an AI tool also differ from a database: does the vendor train on your content, how long are logs kept, and is data stored in one place but processed in another. Those answers change by plan and by date, so the sheet records when you checked.

Does DPDP require personal data to stay in India?

No. The Act lets the Government restrict transfers to countries it notifies under Section 16(1), and none were notified as of November 2026. Section 16(2) keeps stricter sector rules, such as RBI’s payment data storage direction. Record where each vendor stores and processes data so you can react if that changes.

Is a filled-in template enough to show compliance?

No. A data map is a working record, not a certificate, and the Act has no DPDP certificate. It shows where the gaps are. Mono does not certify anyone and is not a law firm, so have your own lawyer confirm how the Act applies to you.

Skip the spreadsheet. We map it for you.

Sources: Digital Personal Data Protection Act, 2023, Sections 4, 5, 6, 7, 8, 9, 11, 12 and 16. Digital Personal Data Protection Rules, 2025, Rules 3, 6, 7 and 8 (G.S.R. 846(E), Gazette of India, 13 November 2025). Vendor documentation as named above, checked 20 November 2026. Last reviewed 20 November 2026. See also our DPDP compliance page.