Niyatna Docs

The Formation Roadmap

A practical formation path for companies moving from scattered AI use into governed agentic operations.


Niyatna formation starts with company intent, not tooling. The goal is to understand how work should move through people, agents, memory, permissions, and proof before automation is introduced.


Phase 1: map company intent

Identify the work that should become askable, assignable, remembered, automated, and proven.

  • Intent mapping: Clarify the company outcomes that matter.
  • Drag audit: Find recurring work, repeated questions, handoff delays, and review bottlenecks.
  • Boundary design: Decide which actions can move automatically, which require review, and which should stay blocked.

Phase 2: define the operating map

Niyatna needs a company map before agents can act safely.

Define:

  • roles and departments;
  • authorized requesters;
  • review and approval owners;
  • external partners or clients with scoped access;
  • SOPs, standards, and recurring workflows;
  • evidence required before work is considered complete.

Phase 3: create the first OpenIntent flows

Start with a narrow set of repeatable requests.

Each flow should include:

  • the request shape;
  • context and source materials;
  • permission boundaries;
  • assigned human or agent path;
  • acceptance criteria;
  • proof and report-back format.

Phase 4: connect memory, route, and proof

Once the first flows are clear, Niyatna connects the system surfaces:

  • Memory provides relevant company context.
  • Route chooses the right human, agent, model, or tool path.
  • HQ tracks owned work and review gates.
  • Proof of Intent records evidence and final status.

Phase 5: expand deliberately

After the first workflows prove useful, expand by department or operating area. Niyatna should grow as a governed company system, not as a pile of disconnected automations.