Upcoming · AI Suite

AI Consent Drafter

A Rule 3 consent notice written from the processing you already recorded, in the languages your customers actually read.

Join the waitlist See what it has to get right The notice builder it sits inside is live today.
The actual work

Writing a notice is not typing a notice

Three problems make this take weeks, and not one of them is solved by having a template to start from.

01

Getting the granularity right

Rule 3 wants each purpose itemised with the personal data it needs. Describe them too broadly and consent means nothing; too finely and you hand someone forty toggles.

02

Plain language is a legal test

Not a style note. A notice a person cannot follow has not given the information the Act requires, however accurate the sentences in it are.

03

Translation that keeps its meaning

A notice may be given in English or any language in the Eighth Schedule. Translating a legal instrument is where obligations quietly change shape.

A template library gets you past the blank page and no further. It cannot judge whether “service improvement” is one purpose or four in your business, and it cannot tell you that the Marathi version now says something the English one does not.

Being built to

Five things a model is genuinely good at

Each is a pattern-matching problem across your own records — which is exactly the kind of work worth automating, and exactly what a person does badly at volume.

1

Read your processing, not a sector guess

The purposes, categories and recipients in your registry become the itemised list. The notice describes what your business actually does, rather than what a template assumed a business like yours might do.

2

Normalise the granularity

Where you have described one purpose that is really four, it separates them. Where you have forty that a person could never meaningfully choose between, it groups them into decisions somebody can actually make.

3

Write to a reading level, and check it

“Processing of biometric identifiers for authentication” becomes a sentence a patient understands. Plain language is measured here rather than asserted — a draft that scores badly is rewritten before you ever see it.

4

Translate, then verify by translating back

Each regional version is translated back into English and compared against the original. A clause that has drifted in meaning is flagged rather than published, which is the failure machine translation makes silently.

5

Keep an estate consistent

Twenty notices written over two years describe the same billing process five different ways. It aligns the vocabulary across all of them, so a reader and an auditor see one organisation rather than five.

Watch it work

Three judgements a rule cannot make

This is where the model earns its place. None of these is a lookup, a template or a find-and-replace.

Your registry says

“We process patient data to improve our services.”

It returns four purposes
  • Clinical care and treatment records
  • Billing and insurance claims
  • Appointment reminders by SMS
  • Service-quality feedback calls
Because “improve our services” is a bundle, and a bundle cannot be consented to under §6. Only the last two are optional.
Your legal team wrote

“Processing of biometric identifiers for the purposes of authentication and verification of identity.”

It returns

“We use your fingerprint to check it is you when you arrive.”

Same meaning, scored for readability, and rewritten again if it does not pass. Plain language is a Rule 3 requirement, not a preference.
The Marathi version, translated back

“You may cancel your permission by contacting the hospital.”

It flags the drift

The English says withdrawal must be as easy as giving consent. The Marathi now requires contacting the hospital, which is harder.

§6(4) breached in one language and satisfied in another. Held back for review rather than published.
The Eighth Schedule

Twenty-two languages, and a real problem

The Act lets you serve a notice in English or any Eighth Schedule language. For most Indian businesses that is a translation problem nobody has budgeted for.

01

Your customers are not uniform

A hospital in Indore serves Hindi and Marathi speakers in one waiting room, so a regional language is not optional there. English alone is a notice technically given and never read.

02

Every version is a legal document

Each translation carries the same weight as the English. A mistranslated withdrawal clause is a defective notice in that language, not a typo.

03

And they all have to stay in step

Change one purpose in the English and every other version is now stale. Doing that by hand across eight languages is why notices are updated late or not at all.

The consent module publishes in eight languages today, with a translation per notice per language maintained by hand. Closing the gap between eight and twenty-two — and keeping them all current — is the single strongest reason this module is worth building.

After it is published

Knowing when a notice has gone stale

A notice is accurate on the day it is published and starts drifting the moment your business changes. Nobody notices until somebody asks.

01

It watches the registry

A new system, a new category, a new recipient. Each is a change to what you do, and therefore a change to what your notice should say.

02

It grades the change

A reworded sentence is a minor version and consent stays valid. A new purpose is a material change, and material changes need consent taken again.

03

It drafts the re-consent

Where consent has to be retaken, the campaign and the revised notice arrive together rather than as a project somebody has to schedule.

Version control on notices already ships — a minor bump keeps existing consent valid, a major one triggers re-consent. What is coming is the part that notices the change in the first place.

Your control

Drafted by the model, published by you

A notice is a legal instrument. It carries somebody’s name, and the platform makes sure that name belongs to a person.

1

The compliance gate still applies

The draft goes into the same checker every notice does — standalone, itemised, third parties named, withdrawal as easy as giving, the Board complaint route present. A generated notice earns no exemption from any of it.

2

Your DPO edits and publishes

Every word is editable before anything goes live, and the record shows who approved it and when. That is the signature an auditor is looking for.

3

Your lawful bases stay yours

Whether a purpose runs on §6 consent or a §7 legitimate use is a judgement with real consequences. It drafts from the basis you have set and shows you what follows from it.

The notice builder underneath is live today.

Vetted templates per sector, eight languages, and a compliance gate that will not let a non-compliant notice publish. The drafter replaces the blank page, not the machinery.

Questions

About the AI Consent Drafter

Can we write DPDP consent notices today without it?

Yes, and most of our customers do. The consent module ships vetted templates per sector with purposes, categories and legal bases already set, publishes in eight languages, and runs a compliance gate that will not let a notice go live while a Rule 3 or §6 requirement is unmet. The drafter removes the blank page and the translation grind. It does not remove a step you would otherwise be stuck on.

What makes a generated notice better than a template?

A template cannot see your business. It cannot tell you that what you have written as one purpose is really four, or that forty toggles is not a choice anybody can make. It cannot measure whether a sentence is readable, and it cannot notice that your Marathi version now says something your English one does not. Those are the things that actually take weeks, and they are pattern problems rather than drafting problems.

How do you know the translation is right?

By translating it back. Each regional version is rendered into English again and compared against the original, and a clause whose meaning has moved is flagged instead of published. That matters because machine translation fails silently — the output reads fluently and the obligation has changed. Where the stakes are high, our empanelled privacy lawyers review the notice in the language it will be served in.

Which languages will it cover?

The consent module publishes in eight today: English, Hindi, Marathi, Tamil, Telugu, Kannada, Bengali and Gujarati. The Act permits a notice in English or any of the twenty-two languages in the Eighth Schedule to the Constitution, so there is a real gap between what we do and what the Act allows — and closing it is the main reason this module is worth building rather than a nice addition to it.

Is the output legally reviewed?

Every template in the library goes to Indian privacy counsel before it ships, and we will not describe a machine-written notice as counsel-reviewed simply because it was built from one. A generated draft is a draft: it passes the same compliance checks as anything a person writes and needs the same named approval. For a notice carrying real exposure, our empanelled lawyers review it as a service — that is available today and does not wait on this module.

When does it ship?

No date, on purpose — a date we later move is worse than no date. Join the waitlist and we will tell you when it produces notices we would put our own name to. The consent module it sits inside is live and does not depend on it.

In their words

What compliance teams tell us

“We always thought DPDP compliance was the client’s responsibility since we were only executing services. The evaluation made it clear that how we handle client data creates risk on our side too. It changed how we work internally.”
DSFounderDigital services firm
“We had a basic understanding of DPDP requirements, but the scorecard highlighted gaps we hadn’t identified internally — especially around consent handling and data visibility. It gave us a much clearer starting point.”
BSFounderB2B SaaS company
“The DPDP score was surprisingly insightful. Within minutes we could see where we stood and what needed immediate attention. It simplified something that initially felt quite complex.”
FPProduct HeadFintech platform
“After reviewing our score we opted for a consultation. The discussion was very practical — we got clear direction on what to fix first and how to approach DPDP compliance in a structured way.”
LGFounderLogistics company

Real client quotes, attributed by role and sector — we never name a client.

Insights

DPDP, explained properly

All articles

Working across

Healthcare & HospitalsDiagnostics & Labs Education & EdtechBFSI & Fintech InsuranceLogistics & Mobility Retail & E-commerceIT & SaaS ManufacturingReal Estate
Hospitality & TravelMedia & Publishing Professional ServicesStaffing & HR TelecomOnline Gaming NGO & Non-profitGovernment & PSU Pharma & Life SciencesAutomotive

Publish a notice this week.

Without waiting for any of this. Pick a vetted template, fill in your details, and the gate tells you before you publish rather than after.