Examine
We look at how decisions are really made today: who decides, on what evidence, and where the picture breaks.
AI and intelligence
Trust and deployment
Requisite Technologies · Intelligence, engineered
Technology is the enabler.
Your institution's intelligence is the advantage.
We build systems that help organisations see the whole picture and decide on evidence, and we help shape the organisation around them. For defence, government, banking and enterprise, inside your walls or fully offline.
What we do
Every system sees a part. We build intelligent systems that see the whole, show their evidence, admit what they do not know, and leave people in command.
The problem
Your information lives in dozens of systems, each with its own version of events, and none of them shows the full picture.
Models answer fluently from stale, partial or conflicting information, and rarely say so.
When something goes wrong, nobody can show what was known, what was assumed and who decided.
The answer isn't more AI. It's better systems around the people who decide.
How we think
The real advantage is what your organisation already knows: the experience of your people, the evidence in your records, and the way decisions get made. We help you put it to work.
We look at how decisions are really made today: who decides, on what evidence, and where the picture breaks.
We build the systems, and shape the roles, rules and accountability around them, so the technology fits how you actually work.
The world keeps changing, so neither the system nor the structure can stand still. We measure, learn and adjust with you.
Grounded in decades of work on how organisations decide well. Read about Requisite Agility →
What we build
Flagship · pilots openKybern · Systemic intelligence
Kybern sits on top of the systems you already run and reads them as one connected world. Every fact keeps its source, every gap is flagged, and people stay in command.
259 of 259 cases passed in our open Reality Challenge, against 74 for a naive baseline. See the leaderboard →
Case studies

Defence · Indian Army · Delivered
HQ 2 Signal Training Centre, Goa, needed its CSBO Grade-II recruitment to stand up to audit and appeal, without depending on an external cloud. RecruitOS ran it end to end: offline application intake, national ID and document verification, officer review, and an immutable record of every decision.

Navy · Fleet readiness
A working Kybern world of ships, crews, spares and maintenance. It answers from the evidence, flags the readiness reports that are stale, and names the check that would settle each doubt. 14 of 14 test cases passed; a naive baseline passed 4.
See how it was tested →
Banking · Financial operations
Feeds, cut-offs, exceptions and the people who own them, in one picture. When a feed goes quiet, Kybern says so instead of assuming it is fine. 14 of 14 test cases passed; a naive baseline passed 4.
See how it was tested →
Healthcare · Hospital operations
Wards, staff rotas, admissions and discharges as one connected system, with missing reports kept visible rather than filled in. 14 of 14 test cases passed; a naive baseline passed 4.
See how it was tested →Why Requisite
How we deliver
Everything we build can run on your premises, or fully offline when it must. No public AI service in the loop, no data leaving the building, no per-token bills.
Entrada AI-in-a-Box
Meru M128 with Entrada Knowledge AI installed. Unpack, connect to your network, start asking.

Sovereign mode: no internet connection, local AI models only, documents from approved sources. Built for defence, security and sensitive government work.
The same software at every point, from fully isolated to fully managed.
Behind it
RCA
Our research programme into how perception, memory, reasoning and checking fit together, so intelligence comes from the whole system rather than one giant model.
Research stageExplore RCA →RA
How organisations stay viable as their world changes, built from what the frontline experiences and what demand requires, not only from the top.
Explore Requisite Agility →Reality Challenge
Our open benchmark: every case scored in public against a simpler baseline. The blind test with expert partners comes next.
See the leaderboard →Consultants, integrators, resellers and research partners: build it with us →
Insights
Confidence is cheap. A system that advises on whether a task force can sail, or whether tonight's settlement run will close, must also say what it does not know, name the evidence that would change its answer, and stop where its authority ends.
Entrada · 5 min readMost document AI answers from whichever version it finds first. When a new circular quietly supersedes an SOP, a confident answer becomes a wrong one. Here is how Change Radar keeps every answer on the latest authoritative version, offline.
Research · Kybern · 7 min readAcross 19 synthetic worlds, Kybern passed 259 of 259 internal cases; a naive baseline passed 74. That is not proof, and we explain why. The real test comes next: expert cases scored blind, and every failure published.
Next step
Bring us one real problem. We demonstrate, then run a bounded pilot inside your environment with the yardstick agreed up front, and you can stop if it does not earn its place.
Or write to reach@requisitetech.com · +91 98150 23001