AION Rakshak
Plan the week before it arrives.
AION Rakshak forecasts the registration load on each police station seven days ahead, with uncertainty bands. Every override, surge alert and dispatch decision is written to a tamper-evident ledger the moment it happens.
Request a demoDownload the brochure PDF, 7 pages, 1.8 MB
Forecasts for stations, not people
Rakshak forecasts workload per station and day. It does not score, rank or predict individuals.
A paper trail by default
A hash-chained ledger with a daily seal records each decision, and tampering with any entry breaks the chain visibly.
Starts from the public record
The first forecasts are built from a state's own published FIR data, before any new dataset is requested.
The signal was already there
Every station keeps the same shape. Almost nobody was reading it.
Public FIR registers carry minute-resolution timestamps. In the registers we have read in full, station after station traces the same curve: quiet before dawn, a slow climb through the afternoon, a hard peak at night, and a quieter Sunday.
AION Rakshak reads a state's own published record first, and turns it into a seven-day forecast before a single new dataset is requested.
The product
A forecast the control room can act on, with a paper trail behind it.
Station-by-day demand forecasts, seven days out, with uncertainty bands rather than a point guess. Every override, surge alert and dispatch decision is written to a tamper-evident hash-chain ledger the moment it happens.
- Watchfloor
- Station map with a satellite base layer, a 24-hour district ribbon, a coverage timeline and per-station drill-down, deployed on AION infrastructure with single sign-on.
- Evidence and audit
- Hash-chained event ledger, periodic seals, tamper-cascade detection and electronic-record certificates under section 63 of the Bharatiya Sakshya Adhiniyam, 2023.
- Pattern surface
- Station rank, share, year-on-year change, seasonality and a 7 × 24 registration heat map, built on the full public ledger.
- Scoped roles
- Five roles, administrator, controller, station house officer, auditor and viewer, each scoped to its own police zone.
- Reconciled counts
- Public FIR counts reconciled against the state crime records bureau and NCRB submissions before any forecast is built.
- Honest bands
- Uncertainty bands are checked against what happened in backtest, so a band means what it says.
Data protection
What DPDP-by-design means in Rakshak, in plain terms.
The Digital Personal Data Protection Act, 2023 applies to personal data. The simplest protection is not to hold any, so Rakshak's forecasting runs on data that identifies no one. Where personal data may one day be needed, the safeguards come first and the data second.
How access is controlled
- Five roles, scoped to a zone
- Administrator, controller, station house officer, auditor and viewer, each limited to the police zones assigned to that account.
- Changes take effect at once
- The role is read from the database on every request, so demoting or disabling an account cuts off an open session immediately.
- Strong sign-in
- Bcrypt password hashing, optional TOTP second factor, and lockout after repeated failed attempts, with every sign-in written to an access log.
- A record that cannot be quietly edited
- Every override, alert and dispatch decision enters a hash chain. Changing any entry breaks every entry after it, and auditors can verify the chain and issue a section 63 certificate for any range.
| Held | Not held |
|---|---|
| Police station, FIR number, district and registration time, from the state's own public FIR list | Names of complainants, victims, accused or witnesses: the forecasting store has no column for them |
| Forecasts per station and day, with uncertainty bands | Any score, rank or prediction about an individual |
| Who did what, when: logins, overrides, alerts and dispatch decisions | Passwords or machine tokens in readable form: only their hashes are stored |
Who is responsible for what
Under the Act the police force is the Data Fiduciary, and AION acts as its Data Processor, processing only on the force's documented instructions under a written agreement.
Retention is set by the force
The Act exempts much State processing for the prevention and investigation of offences, including some erasure duties (section 17). How long data is kept, and for what purpose, is decided by the police force as Data Fiduciary, and AION processes it on those instructions.
Before any personal data
Any person-level feature ships only behind a formal data extract, an auditor's co-sign on every list and an outside ethics review, and it is never used to rank people for patrol.
What we can add
Not a roadmap promise: a list of what is ready, gated or borrowed.
- State-system integrations
- Signed, outbound event push to existing state systems. Code ready; wired when a receiving endpoint is agreed.
- Traffic and camera fusion
- Accident records and camera challans folded in so patrol and traffic forecasts share one demand picture. Needs a data MoU.
- Seasonal surge model
- Tourist arrivals, occupancy and the event-permission calendar as drivers, for states whose demand is seasonal. Needs a data MoU.
- Hotspot beat layer
- Beat polygons and a shared GIS layer, adapted from AION's existing patrol-optimiser module. Code ready.
- Investigator lookup
- A case lookup for investigating officers in the CCTNS style, never a ranked list pushed to patrol and never complainant or victim data. Ships only behind a formal extract, an auditor co-sign on every list, and an outside ethics review.
- Multi-state replication
- The same engine already runs a live deployment outside law enforcement; expansion is a data agreement, not a rebuild.
What it does
- Station-by-day forecasts seven days ahead with uncertainty bands modelled
- Watchfloor map with district ribbon and per-station drill-down
- Role-based access for administrators, controllers, station house officers, auditors and viewers
- Forecasting on public FIR data that names no one; roles, sign-in and ledger built for the DPDP Act, 2023
- Single sign-on, deployed on AION infrastructure or yours
Victim and complainant data is never used for forecasting.
Who it is for
- State police control rooms and command centres
- Police IT and modernisation directorates
- State crime records bureaus
How it starts
Win on evidence, one stage at a time.
Stage 0: prove it on the public record
Pattern surface and backtest built entirely from data the state already publishes. Nominal cost, recovered only on a go decision.
Stage 1: pilot on the formal extract
A read-only CCTNS or 112 feed, a block-level backtest and two control stations.
Stage 2: fixed-fee pilot
The state pays for the pilot, and the reference story builds alongside the rollout.
Stage 3: statewide, outcome-linked
An annual subscription with a component tied to the same test-and-control measures the pilot was judged on.
The record already exists. The only question is who reads it first.
Write to us for a demo of the watchfloor and the ledger.
Email info@aionpl.com