We build civic intelligence infrastructure.
Civiq.plus exists because the hardest part of improving public services is not effort or intent. It is that nobody in the system can see the same picture at the same time, at the level where the problem actually lives.
2024
Founded
Hyderabad
Headquarters
8 levels
Geographic depth supported
India-first
Built for this context
Why we exist
The gap we kept running into
Ask five departments in the same district how many open water complaints exist in a given ward and you will get five numbers. None of them are wrong. They are answers to five slightly different questions, because each department holds a different definition of the ward, a different notion of open, and a different idea of what counts as a complaint.
This gets diagnosed as a data quality problem and treated with a cleaning exercise. It is actually a modelling problem, and cleaning does not fix it, because next month’s data arrives with the same divergence. Meanwhile the citizen who reported the problem is told it is resolved, and reports it again in six weeks.
Civiq.plus is our attempt at the infrastructure layer that makes this tractable: a live geographic model that every module writes into, so that a ward is one ward, an issue is one issue, and a claim that something was fixed carries the evidence that it was.
We are deliberately not a social listening product. The platform is a decision-support and operations system: it has to explain what changed, where, why it may have changed, what the options are, what was actually done, and whether the measurable outcome improved afterwards. Anything that stops at the first step is a scoreboard.
Principles
What we hold to
Geography is the organising primitive
Most civic data problems are not analysis problems. They are the consequence of several departments describing the same place in incompatible ways. Fixing that first makes everything downstream tractable.
A number without confidence is a guess
A detailed dashboard built on thin data is more dangerous than no dashboard. Confidence and provenance travel with every score we expose, and low confidence is displayed rather than smoothed away.
Resolution has to be demonstrated
Closure rates are the easiest civic metric to improve and the least meaningful. Where it matters, closing an issue should require evidence and an independent check.
Aggregate, not individual
We build for aggregate geographic, cohort and community intelligence. We do not build covert individual-level profiling, and that is an architectural constraint rather than a policy setting.
AI drafts; people decide
Agents cluster, summarise and notice at a scale people cannot. They do not get to be the last step before something reaches the public.
The audit is part of the product
A civic decision is judged on whether it was reasonable on the evidence available at the time. Keeping that evidence, and its confidence, is not overhead.
Want to work on this?
We hire people who are more interested in whether the ward improved than in whether the dashboard looked good.