DorsiumDorsium
Distributed Task Estimation

Define task size with machine consensus

Independent Nodes analyze the task and reach a matching decision on its size.

Prevents incorrect size assumptions before work begins. Replaces repetitive manual pre-evaluation.

No guessing. Only matching decisions.

Get early access
  • More accurate planning
  • Fewer wrong estimates
  • A clearer picture of the task
How it works

How a task's size gets decided

Three clear steps from task input to an accepted decision.

  1. 1

    Task enters analysis

    A backlog task or ticket is submitted to the analysis queue before any implementation starts.

  2. 2

    Multiple Nodes evaluate the task size separately

    Several independent Nodes read the same task and each decides on its expected size.

  3. 3

    Only matching decisions are accepted

    When the Nodes agree, the decision is accepted and written to the audit trail.

Live flow

A verification round

Task in, 5 Nodes analyze, one final decision out.

The task

#2104Planning
Integrate OAuth 2 login flow
Submitted by
pm.alex
Expected size
Medium
State
Ready for analysis
Dispatch to Nodes

Nodes

Node 01
Parallel analysis
Waiting
Ready for the next analysis round
Node 02
Parallel analysis
Waiting
Ready for the next analysis round
Node 03
Parallel analysis
Waiting
Ready for the next analysis round

5 Nodes run in parallel in the background - 3 shown here for visual clarity.

Consensus Result

Matching decisions0
Diverging decisions0
Awaiting enough matching decisions

Rule example: at least 3 of 5 Nodes must agree on the same size before the decision is accepted.

What Nodes can read

  • All Nodes read the same task input
  • All Nodes apply the same rules
  • Nodes run independently, without talking to each other
  • No single Node can decide the size alone
  • Decisions are compared, not averaged
  • A diverging decision is excluded from the consensus
  • Audit trail entry can be created safely

Final action

  • Accepted scope: Medium
  • Task status: Ready for planning
  • Decision recorded in the audit trail
Core principles

Why this model holds up

Three principles keep every analysis round trustworthy.

Deterministic

Same task, same rules, same decision - every run.

Auditable

Every decision writes an immutable audit trail entry.

Consensus-backed

No single Node decides alone - multiple Nodes must agree before a decision is accepted.

Where it's used

Projects where this model is already in use

Internal systems, delivery environments, and backlog workflows where this model is live.

ProjectDorsium Mobile Platform
ProjectDorsium Corporate Website
ProjectDorsium Community Platform
Technical layerDistributed Scope Agent
Technical layerAudit Trail Layer
Technical layerPublic API Layer
Early access

See the real size of your tasks

We're onboarding pilot teams before the public rollout.

Book an intro call
Explore other products
Distributed Work Review

Reviews completed work with independent machine consensus.

Distributed Ticket Validation

Validates completed ticket closures with independent machine consensus.