A Mapping Company in Software: Multi-Agent Orchestration of Geospatial Production
AG2I runs geospatial production the way a mapping company is organised — a coordinated team under one conductor, taking a project from raw data to a quality-checked, published deliverable. An orchestrator plans the work and adapts to failures; deterministic specialist agents do the compute; a verifier gates quality at every stage; a shared blackboard is the single source of truth. This reports a working, tested implementation of that backbone and an end-to-end run of a representative production line: a six-way parallel tie-point fan-out, an adaptive bundle-adjustment recovery (σ₀ 1.42 → 0.64 px), and a check-point QA gate at 0.25 m — ingestion to report, with no language model in the compute inner loop.
1. The problem
Continental-scale geospatial production — thousands of images, billions of points, multi-sensor, continuous QA — cannot run as a monolithic script. It must fan out across a coordinated team of specialists, with someone planning the work, someone checking it, and a clear chain of custody. That is precisely how a mapping company is organised; the contribution here is to build it in software, so the team is fast, parallel, and self-verifying.
2. The team
In the established terminology this is an orchestrator–worker multi-agent system — a hierarchical topology in which one coordinator dispatches specialist workers and a verifier gates their output — operated as an agentic organization, or synthetic workforce: each role a mapping company staffs with people is staffed instead with an agent. The mapping is direct.
| Mapping-company role | Agent role | Reasoning or compute |
|---|---|---|
| Production director / COO | Orchestrator | reasoning — plans the DAG, sets params, diagnoses failures, escalates |
| Department managers | Sub-orchestrators | reasoning — own a phase, delegate to specialists |
| Specialists (AT, LiDAR, ortho, cartography) | Worker agents | deterministic compute — the math |
| QA / QC department | Verifier agent | metric (check-point RMS, σ₀) + a reasoning pass that gates |
| Operators / manual review | Human-in-the-loop | exceptions and low-confidence cases |
| Project tracker / shared drive | Blackboard + data plane | single source of truth; data moves as handles, not payloads |
3. The one principle
LLM at the edges; deterministic compute in the core. The specialists running aerotriangulation, LiDAR, and orthorectification are deterministic math — never a language model inside a bundle-adjustment iteration. The reasoning layer (orchestrator and verifier) is where judgment lives: planning, choosing parameters per dataset, diagnosing a bad solution, and gating pass / re-run / escalate. This is the opposite of frameworks that wrap every step in an LLM, and it is what keeps the team fast, inexpensive, and reliable while remaining genuinely agentic.
4. A production line, run end to end
The orchestrator ingests the project, plans the task graph, and dispatches it. Six tie-point specialists run in parallel (≈6× over serial), fanning in to a merge. The bundle adjustment first returns σ₀ 1.42 px and fails the verifier; the orchestrator diagnoses it, switches the estimator to a robust loss, and re-runs — σ₀ 0.64 px, pass. Surface, orthophoto, and HD-map features follow; the QA agent computes an independent check-point RMS of 0.25 m, gates pass, and the report is published. Had the adaptation failed, the task would have escalated to a human operator rather than passing silently.
| Stage | Outcomeverifier-gated | Behaviour |
|---|---|---|
| Tie-point extraction (×6) | 11,325 pts | parallel fan-out |
| Bundle adjustment | σ₀ 0.64px | 1.42 → adapt → 0.64 |
| QA check-points | 0.25m | gate: pass < 0.30 m |
| Job | complete | 15 tasks, self-verified |
5. Verification is first-class
Quality is not bolted on at the end. Every task emits a metric, and the verifier gates the graph: an out-of-bounds σ₀ or check-point RMS triggers the orchestrator to adapt and re-run, or to escalate to human QC. The team checks its own work and produces a provable accuracy report — the distinction from generic multi-agent frameworks, and the same check-point discipline behind AG2I's aerotriangulation and orthorectification case studies.
6. Status
The orchestration core — the Task Contract every module implements, the capability registry, the blackboard, and the DAG scheduler with fan-out, verifier gating, adaptive retry, and escalation — is implemented and covered by a protocol test suite. The worker compute in this demonstration is representative (the figures echo AG2I's measured Graz results — independent σ₀ 0.64 px, check-point RMS 0.25 m); the production modules — OpenStereo, OpenLiDAR, OpenFeatureX and the rest — plug into the same Task Contract with no change to the orchestrator. The next step wires a live module through the team on a real block.
OpenMultiAgent — the orchestration backbone of the AG2I suite. Contact: info@ag2i.ai.