Delivery systemGovTech
Know who you serve, and prove how you served them.
Registries connected, cases routed and closed, service levels visible by district and ward, commitments tracked from announcement to commissioning.
Public Proof
When a citizen asks whether the promise was kept — who answers?
Publytics builds the data and AI infrastructure public institutions run on: one evidence chain from what was committed to what was delivered — verified, dated, in Tamil, and able to survive a challenge.
Eight institutions · Six product families · One evidence chain
The shared core · five layers · illustrative
Method Standardv1.0 · published
Publytics Tracker 131 Mar 2027
DPDP obligations13 May 2027
TN low-value linebelow ₹25 lakh
Start with your institution
One evidence chain. Eight institutions. Each buys it in a different shape.
A secretary, a compliance officer, a programme officer and a data editor each ask a different first question. Each pathway reflects its own procurement route, budget head and definition of value.
/0.1DeliveryAccountability
Government & public sector
“We have the announcement and the spend. We cannot produce the line between them, by ward, in Tamil, on the day someone asks.”
See the pathway/0.2Delivery
Municipal & urban bodies
“Complaints reach us through 1913, the app, councillors and social media, and they get closed. We cannot show, ward by ward, what was actually fixed, when and by whom.”
See the pathway/0.3Obligation
Regulators & supervisory bodies
“Returns arrive on a cycle, self-declared, in formats that resist comparison. We are supervising a population we cannot see.”
See the pathway/0.4Obligation
Enterprise & GCCs
“We have a policy. What we do not have is a record that would survive an inspection in May 2027.”
See the pathway/0.5Accountability
Foundations & philanthropy
“Our partners report outputs; the board asks for outcomes. We cannot show which programme produced which result — or how sure we are of it.”
See the pathway/0.6Accountability
Universities & research
“Every project rebuilds the same dataset from Tamil sources, and it dies with the grant. We need data we can cite, reuse — and disagree with.”
See the pathway/0.7Accountability
Newsrooms & media
“We need the dataset, the method, and the ability to disagree with you in print.”
See the pathway/0.8DeliveryAccountability
Legislatures & public offices
“Assurances given on the floor, petitions handed in at the office — we cannot show what became of them, and the record leaves when the term ends.”
See the pathwayMost institutions can report what they announced and what they spent. Almost none can show the chain between them.
The gap is structural, not personal. A department can produce a budget line and a press release. It usually cannot produce, on demand and in the language the citizen speaks, the record that connects the two — what was committed, what was verified, by whom, when, and with what stated limits.
That is not a reporting problem. It is an infrastructure problem — and it is the one we build for.
Read the full argumentThe operating loop every product implements
Capture the commitment as it is made.
Every commitment, delivery event, and grievance enters as a typed record at the moment it happens — not reconstructed later from a spreadsheet.
- One shared definition of a promise, across every department.
- Multilingual at the schema, not at the presentation layer.
- The record is addressable, so it can be cited.
What you buy
One core. Three systems. Six product families.
Every family addresses the same record model, ledger and language layer — so the second purchase is configuration, not a new implementation.
Obligation systemRegTech
Show the regulator the discharge, not the intention.
Regulatory obligations held as typed, dated records — built around India’s compliance calendar and the DPDP obligations that apply from 13 May 2027.
Accountability systemEvidence
Publish a number that survives a challenge.
Commissioned trackers and research, multilingual narrative monitoring, and the verification layer that connects a CSR mandate to a programme outcome.
The shared core
Five layers beneath every product.
Three are ours outright — the record model, the evidence ledger and the disclosure gate — with the Tamil civic evaluation set inside the language layer. The rest is assembled on open Indic models and India’s public digital infrastructure.
The platformProprietary
/0.1
Civic record model
A typed model — promise, scheme, obligation, department, ward, grievance, delivery event, beneficiary, regulation — that all six product families address. Tamil and English labels at the schema, not the presentation layer.
Proprietary
/0.2
Evidence ledger
Source, collection method, version, effective date, confidence and error rate attach to the fact, not to the report. Corrections are permanent notices at the original address. No number renders without them.
Assembled on open models
/0.3
Language layer
Retrieval, classification and summarisation over Tamil, Tamil–English code-mixing, dialect and transliterated civic text. Assembled on open Indic models; the Tamil civic evaluation set and its published benchmark are ours.
Assembled on open models
/0.4
Publication engine
Ships a tracker on an announced date, versioned, with a diff against the previous edition and a machine-readable dataset that can be cross-listed on open-data portals.
Proprietary
/0.5
Disclosure gate
Where AI is used, what it does not do, and who reviews it — with the Method Standard checklist enforced as rules. A publish is blocked if any rule fails.
The solution map
Which product each institution leads with.
Research & Policy Intelligence is the most widely shared family — the lead product for four of the eight. Compliance & Regtech is the fastest to deploy.
- Primary — lead with it
- Secondary — expand into it
- Optional — on request
- Not offered
Delivery systemObligation systemAccountability system
How every engagement begins
One way in, for every institution. Each step is credited against the next.
Start with a free 90-minute Briefing on one question you cannot evidence today. Prove it on your own data in a day. Pilots are scoped to fit public low-value procurement lines, and the annual price is written into the pilot.
- /01
Briefing
Free
90 minutes
- /02
Bootcamp
₹2–4 L
100% credited within 6 months
- /03
Record Model Workshop or DPDP Readiness Assessment
₹8–12 L · ₹6–12 L
Credited against what follows
- /04
Pilot
Scoped
Within the public low-value line
- /05
Annual platform
By institution
Annual price written into the pilot
- /06
Expand
Configuration
A second family or department
Start where the deadline is
The dates that frame every engagement — statutory, and our own.
Two belong to the law, one to the budget cycle, and two are commitments we made in public — so you can hold us to them.
4 Sep 2026
Method Standard v1.0
The five-item publication checklist, open and versioned.
Publytics commitment
13 Nov 2026
DPDP Rule 4 in force
Consent Manager registration opens, 12 months after notification.
Statute
Jan–Mar 2027
Budget and CSR planning
State budget preparation; corporate CSR annual action plans set for the next financial year.
Budget cycle
31 Mar 2027
Publytics Tracker 1
The first public edition, on an announced date, under the Method Standard.
Publytics commitment
13 May 2027
DPDP obligations apply in full
Including Significant Data Fiduciary DPIA and audit duties. Penalties up to ₹250 crore.
Statute
AI at Publytics
Every product uses AI somewhere. We state exactly where — and where it does not.
Institutional trust requires bounded systems, not broad claims. Each product publishes its own disclosure, and the disclosure gate blocks a workflow that cannot produce one.
01
Where AI is used
Classification, multilingual retrieval, anomaly detection, summarisation, and decision support within defined workflows.
02
What it does not do
AI does not make high-impact public decisions, determine eligibility, or operate as a political persuasion system.
03
Human review
People remain accountable for interpretation, escalation, approval, and any action that affects a person or institution.
Tamil-first, benchmarked
Built for Tamil, code-mixed, dialect and transliterated civic text, with a Tamil civic evaluation set so accuracy can be checked before acceptance.
Data stays bounded
Client data is never used to train models outside the client’s governed environment. Residency is written into the contract.
Non-partisan by design
We do not sell political targeting or persuasion systems to parties or candidates — and we log what we decline.
Why trust a young company
The method is part of the product — and it is already public.
A field note proves practice. A tracker proves method. An impact study proves outcome. Until our first tracker ships on 31 March 2027 , the strongest proofs are the ones live today: the Method Standard, the refusal log, and a corrections policy that publishes our errors first.
The Method Standard, v1.0The Method Standard
Publication integrity checklist
- 01Source and collection method disclosed
- 02Funding or commissioning party named
- 03Update date and version history visible
- 04Limitations stated in plain language
- 05Download format and reuse terms specified
Published 4 Sep 2026 · open and versioned
Tracker 1 · 31 Mar 2027
Trackers
Published on an announced date. Versioned. Downloadable data. Named commissioning party. Stated error rate. Diff against the previous edition.
ScheduledImpact studies
One named institution, one number, on the record — before and after. Never an anonymised “a leading department”.
In preparationBenchmarks
Named baselines declared in the same document that defines the task — including the Tamil civic evaluation set, so a technical evaluator can check accuracy before acceptance.
Policy live · none recordedError rates & corrections
A stated error rate on every published figure. A correction notice stays permanently at the original URL, dated, with what was wrong and how it happened. The count is published.
Our boundary
We sell to institutions. Never to contestants.
Publytics does not provide political targeting or persuasion systems to political parties or candidates.
Where our buyer is an elected representative, the engagement is with the public office — paid from public funds, discharging public duties, and its outputs pass to that office’s successor. Where an output concerns an election, it is published openly or not produced at all.
Start a conversation
Bring us the institutional question — not a software shopping list.
Every first meeting is the free Briefing: 90 minutes on one commitment, obligation or figure you cannot evidence today. No deck.