Every row below is a capability in the virtual organisation. Every column is a Knowledge Strata layer. The matrix shows the diagnostic we apply to any organisation, starting with ourselves.
A living snapshot · 1 founder · 26 capabilities · 56 skills · 45 commands
Headline counts current to June 2026. The detailed matrix below is the most recent full capture (April 2026); a refresh is in progress.
L3b — the human source
Steve sits outside the matrix below. L3b is human territory by definition: tacit judgement that has not yet been articulated. The matrix maps what has been externalised; the source remains a person.
Capabilities · 20 of 26 captured here · operate at L1/L2/L3a · L3b is the human ceiling
Chief of Staff
Enabling
Hot
5
1,475
352
ceiling
Security Engineer
Platform
Hot
4
1,505
225
ceiling
Legal & Risk Counsel
Stream-aligned
Hot
3
1,177
150
ceiling
Videographer
Output
Hot
6
1,112
269
ceiling
Platform Engineer
Platform
Hot
4
1,016
223
ceiling
Quality Reviewer
Platform
Cool
3
754
240
ceiling
Commercial Adviser
Stream-aligned
Hot
2
503
48
ceiling
Content Strategist
Output
Warm
4
445
90
ceiling
Technology Strategist
Enabling
Cool
2
439
53
ceiling
UX Designer
Output
Warm
3
433
148
ceiling
3D Artist
Output
Warm
3
423
112
ceiling
Data Engineer
Platform
Cold
2
386
149
ceiling
Product Manager
Enabling
Cold
2
344
58
ceiling
Framework Specialist
Stream-aligned
Cool
3
322
44
ceiling
Financial Adviser
Stream-aligned
Warm
2
312
54
ceiling
Enterprise Architect
Platform
Hot
2
292
46
ceiling
Marketer
Output
Warm
3
184
58
ceiling
Agent-Readiness
Stream-aligned
Cool
2
170
60
ceiling
Research Analyst
Platform
Cool
5
119
57
ceiling
Information Designer
Output
Warm
2
78
42
ceiling
Why the L3b column is blank for every capability: by definition, a capability is articulated knowledge: a role profile, documented patterns, externalised thinking. Tacit judgement cannot be articulated yet; that’s what makes it tacit. The codification ceiling for any agent or AI workforce is L3a. L3b is and remains human territory. This is the architectural boundary most AI workforce products miss.
What L1 looks like — the skills pool
L1 is a shared pool the whole organisation draws from. Any capability can invoke any skill. The four categories below are the April capture.
Production (14)video-produce · remotion-render · quick-image-v2 · refresh-diagrams · defined-terms · nano-banana-pro · blender-mcp · & 7 more
Research & intake (10)deep-research · deep-dive · ingest · ingest-video · daily-scan · scan-state · knowledge-health-check · & 3 more
Governance & meta (16)dashboard · learning-loop · update-config · capability-create · brain-entry · memory-route · & 10 more
Compounding signals
The organisation grows through daily scans, brain journal entries, commits, and extracted skills. Each compounds: what enters as L3a today may become L1 six months from now.
47
Daily scans
processed
9
Brain journals
evolving
62
Commits
past 7 days
56
Skills in the
orchestrator
This dashboard is a product preview
What you’re looking at is Knowledge Strata applied to a real organisation: mapping where each capability’s knowledge sits, where the codification gaps are, and where the compounding is happening. The same diagnostic applied to your organisation is the core of Wellversed, our forthcoming product for distilling and operating organisational knowledge with AI.
Wellversed is in development. Get in touch to see it applied to your organisation.
One person’s L3b, externalised into 7,873 lines of L3a thinking, stabilised into 11,611 lines of L2 patterns, extracted as 52 L1 skills, and running against itself every day.