HOME/TECHNIQUE/Agentic Orchestration/Orchestrator–workers

TECHNIQUE

Orchestrator–workers

Agentic Orchestration

3APPLICATIONS
3OBSERVED OPERATORS
01

State of Practice

CROSS-VALIDATED — 7 OPERATORS

Across deployed cases, orchestrator–workers is implemented as task decomposition plus routed specialist agents/tools, with topology and control depth varying by workload.

Observed Practices

Decompose a complex task into staged subtasks or workers, then coordinate those steps through an orchestrator, planner, executor, graph, or pipeline.

7 of 7 operators with deployed/pilot evidence.
AtlassianBuild.incDropboxGrabMetaSlackUber

Give workers or subagents narrow responsibilities, prompts, tools, or data scopes rather than exposing all work to one general agent.

7 of 7 operators with deployed/pilot evidence.
AtlassianBuild.incDropboxGrabMetaSlackUber

Attach workers to operational tools, data sources, or enterprise systems so orchestration can retrieve context and act in the target workflow.

7 of 7 operators with deployed/pilot evidence.
AtlassianBuild.incDropboxGrabMetaSlackUber

Add separate quality-control stages such as a critic, grader, LLM judge, confidence scoring, filtering, validation, or human oversight around worker outputs.

5 of 7 operators with deployed/pilot evidence.
AtlassianDropboxMetaSlackUber

Use parallel or asynchronous execution when subtasks are independent or when the workflow includes long-running jobs.

3 of 7 operators with deployed/pilot evidence.
AtlassianBuild.incMeta

Keep workflow state, evidence, logs, or metadata outside a single prompt through journals, event streams, graph state, persistent memory, or context passing.

5 of 7 operators with deployed/pilot evidence.
Build.incGrabMetaSlackUber

Where Operators Converge

Every deployed operator decomposes work into multiple coordinated steps, roles, tools, or subagents instead of relying only on one monolithic prompt.

Every deployed operator connects the orchestration layer to real workflow context: enterprise knowledge, code-review systems, search indexes, internal tools, security data, schedulers, proprietary datasets, or external APIs.

Where Operators Diverge

Operators use different orchestration topologies.

APPROACH 01

Hierarchical routing or delegation to specialized subagents/personas.

AtlassianBuild.incSlack

APPROACH 02

Planner/executor or planner-plus-specialist split.

DropboxMeta

APPROACH 03

Graph/ReAct workflow executor with nodes, edges, states, and tool-enabled apps.

Grab

APPROACH 04

Prompt-chained, multi-stage pipeline with pluggable assistants rather than a named hierarchy of agents.

Uber

Operators differ in how they control output quality and risk.

APPROACH 01

Use a critic, grader, LLM judge, confidence score, or filtering stage to review generated work.

AtlassianDropboxSlackUber

APPROACH 02

Use validation, persistent experiment logging, human oversight, compute-budget checks, and predefined guardrails around autonomous execution.

Meta

APPROACH 03

Configure guardrails for the worker, with no cited critic or judge stage in the teardown evidence.

Build.inc

Operators differ in how they expose tools to the orchestrator.

APPROACH 01

Route to domain subagents or experts that encapsulate specialized tools and instructions.

AtlassianBuild.incSlack

APPROACH 02

Collapse retrieval complexity behind one purpose-built tool or a dedicated search agent.

Dropbox

APPROACH 03

Offer plugin/native/community tool frameworks so app builders can add capabilities to agents.

Grab

Watch Items

Context overload is a practical limiter in long or tool-heavy orchestrations: Slack reports accumulated message history filling the context window and investigations spanning hundreds of inference requests and megabytes of output; Dropbox says precision in context is critical and calls out trimming tool/retrieval context; Atlassian says 1000+ Jira issues cannot be stuffed into the LLM context.

False positives, hallucinations, and evidence-interpretation variability require downstream checks: Uber reports standalone prompts creating many false-positive code-review comments, while Slack uses a weakly adversarial Critic to mitigate hallucinations and variability in interpreting evidence.

Tool and data access are deliberately bounded for safety: Slack safely exposes only a subset of data sources through the tool-call interface; Dropbox cites granular access controls and security reviews for safe code execution; Meta’s announced data-access work adds rule-based risk controls and output guardrails.

Some autonomous workflows remain bounded by oversight or guardrails: Meta keeps human oversight at key strategic decision points and checks compute budgets up front; Build.inc describes agent “training” as giving the agent correct guardrails in a JSON file.

02

Implementation Menu

CURATED DEFAULTS
NameKindMaturity
LangGraph supervisor topologylibraryestablished
Temporal workflowsserviceestablished
Claude Agent SDKlibraryemerging
03

Observed in Production

3 APPS