Responsible AI
What our products do not do.
In governance and policing, the limits matter as much as the features. These commitments apply to every Nayatra product, stated plainly.
We do not decide anything. We assemble what is on record so that you can.
Our commitments
Six commitments.
We assist. We do not decide.
No Nayatra product makes a determination. None recommends a charge, an arrest, a grant or a refusal. Our products assemble what is on record and on the ground so a person can decide, and that person remains accountable.
We do not score or predict people.
No risk ratings for individuals. No likelihood-of-offending. No flagging of people for attention. Our products read records that already exist; they do not generate assessments about individuals. If that is what you need, we are not the right company.
What we do analyse are places, patterns and workloads: where incidents cluster, which assets fail, which complaints are urgent. Arrivo's criticality rating applies to a complaint, not to the person who filed it.
We do not create facts.
We retrieve, organise and cite. Where the record is silent, our products say so. They do not infer what probably happened.
We retrieve what helps and what hurts.
Material that undermines a line of enquiry is surfaced on the same terms as material that supports it.
You own the data and the deployment.
Our products run under your control, and can be audited, restricted or switched off by you without reference to anyone.
A person reviews before anything leaves.
Nothing a Nayatra product produces goes into a file, a submission or a decision without a person reading the source. In Arrivo, no complaint closes without a supervisor's verification.
On AI regulation
Built around oversight, not in spite of it.
Several jurisdictions now treat AI used in law enforcement and public administration as high-risk, with duties around transparency, human oversight and record-keeping.
Our products are built around those same ideas: cited answers, human decisions and a complete log. We do not claim compliance on your behalf. We will work through your obligations with your legal advisers.
Questions we expect
Asked plainly, answered plainly.
Can a Nayatra product identify a suspect?
No. Sage can show what is on record and where records connect. Identifying anyone is an officer's judgement, made on the sources.
Does Kshetra predict crime?
No. It shows where incidents have occurred and lets you define hotspots by your own rules. Deployment decisions remain with your officers.
Does Arrivo rate the people who complain?
No. It rates how urgent a complaint is, so the right officer sees it first.
Can we switch it off?
Yes, at any time, without reference to us.
Who can see the log?
Your administrators and auditors, under the access you define.
Bring your legal adviser.
We will walk your legal, oversight and IT teams through exactly what each product does and does not do.