The Agent Operating System
Operate your revenue engine like a system.
At AI speed.
Most B2B companies bolt AI on top of broken processes. Neon Triforce builds the Company Operating System first (strategy, rhythm, data spine), then installs the agents that execute it. This page shows that system, running live, on us.
Architecture
Strategy on top. Agents at the bottom. Signals flow both ways.
Two diagrams explain the whole intellectual property. The Company OS sets direction through three rings. The agent stack executes through four layers. The Obeya is where they meet.
Layer 5 — Strategy (Company OS)
Direction flows down. Results flow up.Ring 1
Sets directionGovernance
Strategy, priorities, KPIs, rhythm. The control system that steers.
Ring 2
Enables executionEnablement
People, process, platforms, data spine. The infrastructure value runs on.
Ring 3
Generates resultsICP Value Loops
Product engine + revenue engine. Where value gets created and learned.
Layer 4 — Interface
How humans talk to the system.
Layer 3 — Intelligence
Skills, scheduled agents, sub-agents. The thinking layer.
Layer 2 — Memory
Long-term memory. INDEX-linked, semantically retrievable.
Layer 1 — Data Sources
Where signals come from. The data spine.
Obeya
Eight tiles. One board. Live applied to Neon Triforce.
The Revenue Insights Room canon: ≤8 tiles, every tile has an owner, target, band, breach rule, latest note. Applied to a one-person business with VAs, the owner column lists which operating role wears the hat for that tile.
Status: static snapshot. Tiles will go live once the JSON manifest generator is built and a nightly scheduled task feeds the dashboard.
Tile 01
liveAgent Health
Owner: Operator-Rutger
Metric: % scheduled tasks ran in window
Breach: Any task >1.5× expected interval
Tile 02
designSkill Library
Owner: Architect-Rutger
Metric: Skills used in 30d, target ≥60% of installed
Breach: Skill unused 90d retire or refactor
Tile 03
designConnector Status
Owner: Operator-Rutger
Metric: MCPs alive, last successful call (target 100%)
Breach: Connector down >24h escalate
Tile 04
designContent Pipeline
Owner: Content-Rutger
Metric: Posts produced + performance vs band
Breach: Output drops 2 weeks. Engagement <50% of 90-day median
Tile 05
designOutput Pipeline
Owner: Operator-Rutger
Metric: Vault freshness. Briefings, drafts, INDEX coverage
Breach: INDEX coverage <95%. Briefing missing 2 days
Tile 06
designDecision Log
Owner: Strategist-Rutger
Metric: Decisions logged per week (target ≥3)
Breach: Change without log entry. Halt and backfill
Tile 07
liveUse Case Maturity
Owner: Strategist-Rutger
Metric: 25 use cases on Ad-hoc to AI-Orchestrated grid
Breach: Stage 4 claim without proof. Downgrade
Tile 08
designExperiment Tracker
Owner: Strategist-Rutger
Metric: Active tests + kill rate (target 20-40% per quarter)
Breach: Kill rate <10% for 8 weeks. Portfolio review
The Bow-tie
25 places AI earns its keep. Drawn the way it actually works.
Acquisition (left) is frequency-based and polynomial. Small conversion gains compound across many steps. Retention (right) is time-based and exponential. Small GRR/NRR gains compound across years. Mutual Commit is the transformation point.
Awareness
- ICP Scoring
- Intent Aggregation
- Website Intelligence
- Data Enrichment
Education
- Website Chat
- Email Personalisation
- Lead-to-Account Routing
- Reply Classification
Selection
- Conversation Intelligence
- Pipeline Intelligence
- Qualification Copilot
- Objection Handling
Mutual Commit
- Forecasting
- Deal Desk / CPQ
- Security & Legal Acceleration
Onboarding
- Success Plan Copilot
- Early Health Signals
- Support Triage
Retention
- Renewal Risk
- Health Scoring
- VoC Sentiment
- Stakeholder Coverage
Expansion
- Upsell / Cross-sell Scoring
- Product-Usage Signals
- Reference Mining
Agents & Schedules
Two schedulers. 33 tasks. One reconciliation.
Cowork has 12 tasks via the scheduled-tasks MCP. Claude Code has 21 in its native Routines panel. Six tasks appear in both, sometimes at different times. The list below is the reconciled view.
CONFLICTS TO RESOLVE
weekly-competitive-scan — Paused in Code, live Thu in Cowork. Reconcile.
content-pipeline — Two different schedules. Pick one or define them as different jobs.
linkedin-engager-scan — Both run daily, different times. Likely duplicate work.
morning-briefing newsletter-digest weekly-competitive-scan paused weekly-visual-inspiration content-pipeline linkedin-engager-scan call-processor weekly-vault-reindex weekly-vault-staleness-audit ai-group-chat-scan authoredup-weekly-performance weekly-call-mining work-order-executor paused weekly-vault-hygiene daily-vault-sync weekly-pipeline-review weekly-newsletter-synthesis daily-job-scanner outbound-dm-coach slack-listener pavilion-cohort-processor-wed pavilion-cohort-processor-fri content-pipeline-build reddit-humour-scan linkedin-connections-scraper allbound-state-scanner allbound-engine Skills Library
65+ skills across three layers. Loaded by directory, not by guess.
A skill is a domain-specific instruction set with embedded best practices. Claude directory-walks the layers and loads only what is in scope for the current context. Same codebase, different mental models loaded depending on the work.
Layer 1 — Global
~23
Loaded every session. Voice, document toolkit, vault infrastructure, sales workflow.
Layer 2 — Neon Business
6
Loaded inside Neon-Business. Content production with brand canon embedded.
Layer 3 — Client Delivery
42
Loaded inside Client-Work. RevOps frameworks, GTM playbooks, Neon canon delivery.
Six on-demand clusters
Every chat starts with a bootstrap question: which cluster to load? Keeps token usage tight, multi-select for cross-cluster work.
Content
neon-linkedin-posting, neon-blog-writer, neon-content-pipeline, neon-gemini-infographic, neon-thought-leadership-architect, marketing-psychology Prospecting
neon-linkedin-lead-triage, neon-positioning-messaging-designer (proposal-generator always-on) Delivery
neon-ai-readiness-day, neon-gtm-ai-use-case-mapper, neon-client-diagnostic, neon-client-skill-library-builder, neon-obeya-builder, neon-operating-cadence-designer, neon-customer-interview-orchestrator, neon-benchmark-reference, neon-company-os-architect, neon-revenue-leak-diagnostic RevOps
all revops-* skills, pipeline-visibility, lead-routing, data-enrichment, revenue-operating-cadence GTM
gtm-planning, gtm-compensation, sales-methodology, neon-deal-velocity-engineer, marketing-operations, neon-icp CS / Expansion
cs-operations, neon-expansion-revenue-architect, neon-partner-ecosystem-architect, partner-channel-operations Principles
The rules that make this an operating system, not a dashboard.
These are not opinions. Skip any of them and you end up with the same telemetry wall everyone else has. Pretty, expensive, ignored.
- 01
≤8 tiles on the executive board. No ninth tile, no shadow dashboards.
- 02
Every tile has target, band, actual, owner (a name not a role), trend, breach rule, latest note.
- 03
48-hour containment SLA on any breach. Owner publishes a containment plan. Logged in Decisions.
- 04
≤3 decisions per weekly session. Anything beyond is tabled for a deep-dive.
- 05
Definition change control. No mid-quarter definition changes without approval.
- 06
WIP caps per function. Explicit limits on active experiments per lane.
- 07
Public kill-rate. Every quarter publish how many experiments were killed and why. Target 20 to 40%.
- 08
Forecast accuracy cadence. Track % of quarters where actuals fell within forecast bands.
Bands are not targets
A band is the normal range. Hitting the band means you are in control. Below band means special cause. Investigate. Do not fire someone.
Annotations are not optional
Every breach, every decision, every experiment needs a note explaining context. This is how you build organisational memory, not just a metric history.
Containment is not root cause
A 48-hour containment plan is a stopgap. Root cause analysis happens in A3s or deep-dives. Log both. Do not conflate them.
Dashboard sprawl is the enemy
One room. Eight tiles. Anything else lives in linked deep-dives. When someone asks for a ninth tile, ask which of the 8 is less critical. Almost always, they redesign one of the 8.