1,259
HTML pages in the SA4 tree — homepage, trackers, research, compare, guides, blog.
Case study #2 · Receipt AgentGo-20260624 · Three proof angles
From brief to a category surface — with receipts, not slides.
PureFlow (#1) proved local acquisition for trades. AgentGo (#2) proves governed factory throughput: a full GEO intelligence marketing surface — trackers, research, compare — with honest brand separation on disk.
Disk inventory June 2026 · 1,259 HTML pages · 592 research topics · 83 compare · 23 tracker landings · desktop demo :8080 · AgentGo ≠ SourceA product
Who this case study is for
Marketing and SaaS leaders evaluating whether you can ship category-scale surfaces without a six-month content program. Agency CTOs who want deploy proof on a screen-share, not a slide that says “we could.”
Not for: local pool operators (see PureFlow). Not for: buyers who need live GEO billing on day one — we prove factory structure first.
Disk inventory (SA4)
What “factory scale” means — counted, not claimed
592
Research topic folders — long-form authority surface, not five landing pages.
83
Compare routes — competitor-adjacent GEO framing (fair use, no false superiority).
Receipt: data/agentgo-sa4-inventory-receipt-v1.json on disk · refreshed by polish_agentgo_case_study_v3.py
Angle A · Factory scale
GEO intelligence surface — trackers · research · compare
GEO trackers
Vendor-specific visibility landings — ChatGPT · Perplexity · Gemini · Claude · Grok · Qwen · Copilot · Meta AI and more.
Research hub
592 research topics plus 120 guides and glossary — the depth buyers expect from a category player.
Compare matrix
Peec · Profound class comparisons — structure and framing shipped by factory, not one-off copy.
Pure law: prove throughput + structure — not live AgentGo SaaS checkout on day one.
Angle B · Dual desktop deploy
One green-unified SourceA spine deploys to two honest desktop shells — so agencies see deploy, not deck.
- Agent Run ·
:5180— canonical commercial landing + both case studies - SA4 / AgentGo ·
:8080— cinematic stack +/sourcea/portfolio lane - Validate —
validate-sourcea-desktop-landing-v1.shon ship window
production = sourcea.app · demo = localhost — we never blur the two in buyer copy.
Angle C · Wil L3 separation
AgentGo lives in the Wil AI ecosystem intelligence stack. SourceA is the runtime governance lane inside that shell — separate products, separate buyer promises.
- Every buyer page carries: portfolio lane — not AgentGo SaaS checkout
- Wil ship gate (routes · browser · E2E) is L3 law — cited read-only from YA5
localhost:3001sign-in is not sold as production
Build chain (summary)
- Scoped brief — category GEO surface · three proof angles · separation law upfront
- Factory generation — trackers · research · compare grid from governed pipeline
- Case study ship — this page on sourcea.app with commercial + depth gates PASS
- Desktop deploy — Agent Run + SA4 parity via
deploy_sourcea_desktop_landing_v1.py
Technical replay: proof chain demo · self-serve eval: sourcea-boot.
Two case studies · one factory
PureFlow vs AgentGo — pick the story that matches your buyer
PureFlow #1
Buyer: local trades · Proof: live site + visit reports · Timeline: 48-hour acquisition.
AgentGo #2
Buyer: SaaS / marketing · Proof: 1,259-page surface + desktop deploy · Timeline: factory throughput.
Demo: SA4 on :8080 when founder session open
Need factory-scale proof for your category?
Asset B Agent Loop Build $3–10K · fixed scope · verified records — governance factory, not hourly GEO consulting