Organization Context Research Process

capture → source gate → enterprise → deep research (5 phases × human gate) | smb → prioritization → assessment → verdict → write once, read everywhere → the pipeline

01Context input & capture

Context types (13 enterprise / 6 SMB) enter through one of two paths — written directly or uploaded as a file — then the active entries are selected for use.

source — writtenDirect entry

Written by the org, versioned and toggleable per downstream stage.

source — uploadedFile upload

Auto-extracted from PDF, DOCX, or CSV.

Active entries selected

The chosen context records feed the next decision and every downstream stage.

02Source of truth — deep research or prioritization

On an enterprise license, the selected entries trigger the adversarial deep-research pipeline; otherwise the Strategic Prioritization Agent scores the candidates directly. Both converge on scored project candidates that seed the portfolio and roadmap.

enterprise onlyDeep research — 5 phases

Scope → Search (Tavily + SEC EDGAR) → Fetch → Verify (two-vote adversarial) → Synthesize. Each finding is merged with provenance and passes a human review gate before any injection.

human gateApprove or revise

Approve sends findings to prioritization; revise loops back to the scope phase.

smb / directStrategic Prioritization Agent

Scores the project candidates and seeds the portfolio and roadmap.

03Assessment — readiness verdict

In parallel, the org context builds a dysfunctionality profile, which feeds a five-dimension delegation-readiness assessment. The verdict decides whether work proceeds.

Dysfunctionality profile

Built from the active context entries.

Delegation readiness

A five-dimension assessment of the organization’s readiness to delegate to agents.

ReadyProceed to the pipeline.

ConditionalProceed only with a remediation plan required.

Not readyGate closed — revisit after the underlying changes land.

04Write once, read everywhere

The scored context and approved findings become the grounding layer that every stage reads continuously:

Discovery

Org context injected into discovery interviews and prioritization.

Design

Injected as constraints and criteria for design decisions.

Build

Injected for tone, voice, and boundary consistency.

Evaluate

Org-specific standards inform rubric generation and scoring.

Operationalize

Adoption and rollout calibrate to the organization’s real structure.

05What the context types carry

Context type	Injected into	Purpose
Company profile	Discovery	Organizational background used in every discovery interview.
Strategic goals	Discovery + scoring	Alignment for discovery and portfolio scoring.
Brand guidelines	Build	Tone and voice consistency in build synthesis.
Org chart	Discovery + design	Who decides what — maps authority for readiness scoring.
Governance policy	Governance	Risk tolerance, compliance, and decision-rights framework.
Evaluation criteria	Evaluation	Org-specific standards for rubric generation and scoring.

Write once, read everywhere. Organization Context is not a stage you finish and leave. It is the foundation the pipeline stands on — captured once, versioned, and read continuously by discovery, design, build, evaluation, and operationalization. Human-in-the-loop early and often: every context entry and deep-research finding clears a human review gate before it touches a downstream agent.