How it works
Six steps from human intent to accepted work — this is how Talomnia performs every task, including building itself.
Process diagram
How it works
-
01
The client describes the task
What must be delivered, what materials exist, what the constraints are, by when, and what counts as successful completion.
-
02
Talomnia determines what and who is required
Roles, Skills, Blueprints. One task does not equal one specialist. The client does not manage this internal complexity: they buy the completion of a task, not a set of agents.
-
03
The economics are determined before execution
What will be done, the expected result, timeline, budget estimate, the maximum allowed budget (Budget Limit), acceptance criteria. New clients: deposit or prepayment; trusted B2B: contractual postpay and credit limits.
-
04
The work is performed
AI agents, software tools, humans, or their combination — whatever the task requires.
-
05
The client receives more than an answer
The result plus Evidence: what was done, on what basis, which checks passed.
-
06
The work is accepted or returned for rework
Acceptance or Rework under the agreed terms, then Settlement — the post-acceptance financial closure.
One example, start to finish
You ordered a market research study. The system selected the roles and knowledge, fixed the budget limit and the acceptance criteria, performed the work and returned the report. The request walked all six steps above — from describing the task to acceptance.
The result is the report. The Evidence is everything that substantiates it: the sources, the methodology, the executor roles, the checks passed, the limitations of the study, the rework history and the cost. All of it ships with the result and stays verifiable.
What this rests on: the Knowledge Contract
Every executor works not from a "prompt" but under a Knowledge Contract — a bounded execution context answering four questions: what the executor must know, what it may do, what it must not do, and how success is defined.