Campaign control surfaces
Cockpit, campaign manager, strategy room, pollster, media monitor, social-signal views and field intelligence pages.
Behaviour turns campaign advice into operator surfaces, creative workbenches, polling and signal views, experiment ledgers, approval gates and learning loops. The first layer is simple: what can be used now, where it sits, what it connects to, and how a tactic becomes evidence.
This estate is the visual operating background for clients and investors: current interfaces, tactical assets, experiment ordering, process maps, architecture diagrams and the evidence library behind the system.
Cockpit, campaign manager, strategy room, pollster, media monitor, social-signal views and field intelligence pages.
Advertising video workbench, performance maps, run packets, approval states and platform export controls.
Decision engine, Albert Orchestration, signal ingestion, data stores, task loops, approval gates and feedback into corpus learning.
Inputs, reasoning, operator decisions, deployment, polling, media signals, social response and measured learning.
Visual QA, manuals, logs, architectural diagrams, deploy reports, case records and structured performance artifacts.
Campaign operators can order a draft experiment from the cockpit and then inspect linked tactics, assets, results and signals by experiment_id.
The pages share the same white Behaviour look and route through Behaviour domains. Older protected pages have been carried over where they still show important operational depth.
| Surface | Question it answers | Where to probe | Key integration |
|---|---|---|---|
| Campaign cockpit | What work is live, what has been ordered, and what has been measured? | campaign.behaviour.ai/cockpit | Experiment API, plans, CRM, products, approvals |
| Campaign manager | What should the director do next, and what tactical records support it? | campaign-manager.html | Pollster API, budgets, approvals, experiment filters |
| Strategy room | What does the corpus recommend, and what polling brief should be formed? | strategy-room.html | Pollster API, corpus search, reasoning workspace |
| Media monitor | What narratives, signals and response drafts require attention? | media-monitor.html | GDELT, social signals, escalation queue |
| Field intelligence | What are field-level observations saying, and what needs escalation? | field-intelligence.html | Booth notes, field reports, chat escalation |
| Video workbench | What creative packet is approved, what assets are attached, and which experiment owns it? | advertising-video-workbench.html | Video workbench API, KV/R2, campaign experiment marker |
| Search inventory | What assets, manuals and visual records exist? | search.html | Protected visual index and asset metadata |
The compact map below is the first layer. The full architecture file includes the comprehensive subsystem map, input/output contracts, controls, feedback paths and glossary.
The comprehensive process and ecosystem diagram from the architecture estate is included here so the high-level story can be probed immediately.
The useful material is listed by function, so a client can understand the system first and an investor can then inspect the proof trail.
| Asset class | What it contains | Current route | Why it matters |
|---|---|---|---|
| Architecture diagrams | System, campaign and strategy flow maps. | architecture-map.html | Shows how information moves through Albert Orchestration. |
| Campaign interface pages | Manager, strategy, media and field surfaces. | campaign-manager.html | Shows the tactical deployment layer. |
| Creative workbench | Video packet, approvals, assets and export shape. | advertising-video-workbench.html | Shows how media production is governed and measured. |
| Performance maps | Advertising performance flow and asset dependencies. | performance-map.svg | Shows paid and creative feedback structure. |
| Search index | Manuals, logs, maps, documents and visual records. | search.html | Makes the asset estate inspectable. |
| Strategy architecture | Reasoning, data, evaluation and deployment layers. | strategy-engine-architecture-2026.html | Shows the decision-to-action path. |
The cockpit can now create a draft experiment. It remains non-executing until a human approves the tactic, assets, measurement plan and risk guardrails.
Operator enters objective, hypothesis, tactic, audience, channel, primary metric and owner in the campaign cockpit.
Creative packets, polling results, media alerts and social signals carry the same experiment_id.
Results are recorded against baseline, measurement window and guardrail metrics before any learning claim is promoted.
Learning events can update advice only when integrity status, attribution scope and human review are recorded.