STELIXX — WRITING · THE DISPATCH LOG

Signals, engineering records,and market theses.

A living index of our research, edge engineering architectures, market theses, and candid build logs from the frontier of agentic systems.

OBS LOG DISPATCH INDEX · EL −0.3°ENTRIES 030 FILEDCHANNELS 05SCROLL ↓

FEATURED DISPATCH — RESEARCHAugust 2026 · 4 min read

How Answer Engines Evaluate Entity Evidence Beyond Backlinks

As search transitions from blue links to generative synthesis, discoverability is determined by entity consistency and verifiable citations across trusted knowledge graphs.

CORE THESIS

“Generative answer engines do not crawl for keyword density—they extract and cross-verify entity claims against multi-source evidence graphs. Brands without structured citation surfaces are effectively invisible to AI synthesis.”

#GEO
#AI Search
#Entity Models
#Perplexity
#ChatGPT

Read research

01Research

The Six Surfaces AI Engines Read

An AI-visibility audit does not measure a website as one thing. It scores six independent surfaces, each of which gates a different stage on the path from question to citation.

#GEO
#AI Search
#AI Visibility

September 20264 min read

02Engineering

Building a Six-Pillar Audit Pipeline

The architecture behind the GEO Score: fetch a site the way an AI engine retrieves it, run six independent pillar checks, and compile the failures into a prioritized fix list. Independence is the design requirement — a strong average can hide a blocked crawler.

#GEO
#Audit Pipeline
#AI Search

September 20264 min read

03Markets

Search Is Becoming an Answer Market

The engine now performs the synthesis the user once performed, and that restructuring changes what being discoverable means. In an answer market, being described correctly beats being ranked highly.

#Answer Engines
#AI Search
#Distribution

September 20264 min read

04Build Logs

Geolify Build Log: Scoring What We Can’t Fully See

Field notes on building the GEO Score: a 0–100 composite over six pillars that has to stay honest about observing systems from the outside.

#Geolify
#Build Log
#GEO Score

September 20265 min read

05Perspectives

The Evidence Ladder: Narrate Below the Line of Proof

The studio runs a public claims discipline built on six evidence tiers, from Possibility to Enduring Venture. The rule is one sentence: never narrate above the evidence line.

#Evidence
#Claims Discipline
#Operating Code

September 20265 min read

06Research

Measuring Answer Share Between Audits

Answer share — how often tracked prompts return a mention or citation — is measurable between audits only as a trend. A fixed prompt set and repeated runs separate real movement from generative noise.

#GEO
#Prompt Monitoring
#Measurement

September 20264 min read

07Research

Entity Drift: When Models Describe You Wrong

Models reconstruct entities from whatever evidence they retrieve; when that evidence is stale or contradictory, the reconstruction drifts from the canonical truth. Drift is detected in answers and corrected on surfaces.

#GEO
#Entity Models
#Prompt Monitoring

September 20264 min read

08Engineering

Claims That Check Themselves: CI for Public Surfaces

This site’s llms.txt, sitemap, and JSON-LD are not maintained — they are generated from the same source content as the pages, with CI drift checks that fail the build when artifacts go stale. The generator is the enforcement mechanism.

#CI
#Claims Policy
#llms.txt

September 20264 min read

09Markets

Distribution Moves Upstream: The New First Mile

The model describes a company to the customer before the customer ever reaches the company. Consistent entity surfaces are the new first mile of distribution — infrastructure, not marketing.

#Distribution
#AI Search
#Entity Surfaces

September 20264 min read

10Markets

The Economics of the Human Gate

An explicit approval gate reallocates cost rather than removing it: machine throughput concentrates where machines are strong, human attention concentrates where it is scarce, and the gate buys what autonomy alone cannot earn.

#Human-in-the-Loop
#Agentic AI
#Outbound

September 20264 min read

11Build Logs

Outflow Build Log: The Learning Loop in Practice

How the Observe → Research → Act → Learn loop actually behaves once a human gate sits inside it — and why rejection carries more information than approval.

#Outflow
#Build Log
#Human-in-the-Loop

September 20265 min read

12Perspectives

RFC: Human Gates for Consequential Agent Actions

An open proposal: consequential actions by autonomous agents should carry an explicit human approval surface as an industry norm. This RFC specifies what the gate must guarantee and invites comment.

#RFC
#Agentic AI
#Human-in-the-Loop

September 20265 min read

13Research

How Answer Engines Evaluate Entity Evidence Beyond Backlinks

As search transitions from blue links to generative synthesis, discoverability is determined by entity consistency and verifiable citations across trusted knowledge graphs.

#GEO
#AI Search
#Entity Models

August 20264 min read

14Engineering

Zero-Runtime Edge Bundling: Next.js 16 on Cloudflare Workers

A deep dive into our production stack: eliminating Node.js runtime overhead by running pure static and edge SSR builds across Cloudflare’s global network.

#Next.js
#Cloudflare Workers
#OpenNext

August 20264 min read

15Markets

The Shift from Passive Inbound to Autonomous, Reviewable Deal Generation

Why the traditional SDR playbook is collapsing under spam saturation, and how agentic systems with human review boundaries create sustainable revenue velocity.

#Agentic AI
#Outbound
#Revenue Operations

August 20263 min read

16Build Logs

Outflow Build Log: Designing Human Approval Boundaries for Autonomous Agents

First-hand engineering lessons from deploying Outflow as an internal proving ground for Geolify outbound motion.

#Outflow
#Build Log
#Human-in-the-Loop

August 20263 min read

17Perspectives

Automate Execution, Not Responsibility: The Operating Code of a Builder

Our philosophical stance on AI agency: expanding human capability rather than replacing human accountability.

#Philosophy
#Ethical Boundary
#Builder Code

August 20264 min read

18Research

How Answer Engines Choose Their Citations

Citations are assembled per claim at answer time: engines retrieve candidates, extract the spans that carry claims, and credit the sources that state them most verifiably. What a page wins or loses in that contest is decided at the level of the sentence.

#GEO
#AI Search
#Citation Behavior

August 20264 min read

19Research

llms.txt and the Machine-Readable Contract

An llms.txt file is a declaration surface: it states what an entity is and which resources are canonical. It can compress identity and fix vocabulary — it cannot compel a citation or outrun contradiction.

#GEO
#llms.txt
#Machine Readability

August 20264 min read

20Engineering

Deterministic Rendering for AI Crawlers

Render strategy is a visibility surface. A page that only exists after JavaScript runs is a page most AI crawlers never see — which is why the GEO Score treats static content as a pillar, not a preference.

#AI Crawlers
#Static Content
#SSG

August 20264 min read

21Engineering

Agents That Wait: Interface Design for Approval Gates

Autonomous agents earn trust at the approval boundary. The interface rules that keep a human gate load-bearing: briefs that carry the reasoning, options that map to real judgments, and no dispatch path that skips review.

#Outflow
#Human-in-the-Loop
#Interface Design

August 20264 min read

22Markets

Why GEO Emerges Now

Search needed an audit discipline and got SEO. Answer engines need one too — and it is audit-first by necessity, because you cannot fix a surface you cannot see.

#GEO
#AI Search
#Measurement

August 20264 min read

23Markets

Venture Studios in the Agentic Era: The Product Is an Agent

What a venture studio looks like when the product is an agent rather than an app: thesis-driven building, real-world testing, shared primitives that compound — and the places where the model strains.

#Venture Studios
#Agentic AI
#Operating Model

August 20264 min read

24Build Logs

Geolify Build Log: The Least Glamorous Pillar

Crawler governance is the GEO Score pillar nobody demos — robots.txt rules, AI bot access, llms.txt — and the one that decides whether anything else gets measured.

#Geolify
#Build Log
#Crawler Governance

August 20264 min read

25Build Logs

Build Log: One Evidence Standard, Two Surfaces

How stelixx.com and the product surfaces ended up under a single claims discipline — one status vocabulary, generated discovery files, and no shipped claim without a source.

#Build Log
#Claims Policy
#Stelixx

August 20264 min read

26Perspectives

Shared Gravity, Independent Orbits

The studio model explained: a shared core of builders, tools, and operational memory, with each venture on its own orbit — and what every venture owes the one that follows it.

#Venture Model
#Studio Design
#Operating Principles

August 20264 min read

27Perspectives

Why First Light: Seeing Before It Is Obvious

The brand thesis: what "see the first light" claims, the interval between a shift becoming visible and becoming consensus, and why the studio names its frontier instead of chasing every frontier.

#Brand Thesis
#Frontier
#Method

August 20264 min read

28Engineering

Self-Hosted by Default: Fonts to Discovery Files

This site self-hosts every byte it serves — vendored woff2 fonts, discovery files generated from source — and admits third-party calls only under a written exception rule. The discipline buys determinism and privacy at the price of owning every asset and its verification.

#Self-Hosting
#Fonts
#Privacy

July 20264 min read

29Build Logs

Outflow Build Log: Signals Worth Acting On

What separates a market signal worth acting on from noise — and why everything downstream in Outflow inherits the quality of the signal at the top.

#Outflow
#Build Log
#Market Signals

July 20264 min read

30Perspectives

Why We Publish the Record

Building in public as evidence rather than marketing: the record is the claim, and a studio that asks for trust publishes its method, its status labels, and its limits.

#Building in Public
#Transparency
#Method

July 20264 min read

03 — Collaborative research

Have data or observations
from the agentic frontier?

We welcome dialogue with researchers, operators, and engineers working on autonomous systems and search discovery.

Start a conversation Explore our works

Signals, engineering records,and market theses.

How Answer Engines Evaluate Entity Evidence Beyond Backlinks

The Six Surfaces AI Engines Read

Building a Six-Pillar Audit Pipeline

Search Is Becoming an Answer Market

Geolify Build Log: Scoring What We Can’t Fully See

The Evidence Ladder: Narrate Below the Line of Proof

Measuring Answer Share Between Audits

Entity Drift: When Models Describe You Wrong

Claims That Check Themselves: CI for Public Surfaces

Distribution Moves Upstream: The New First Mile

The Economics of the Human Gate

Outflow Build Log: The Learning Loop in Practice

RFC: Human Gates for Consequential Agent Actions

How Answer Engines Evaluate Entity Evidence Beyond Backlinks

Zero-Runtime Edge Bundling: Next.js 16 on Cloudflare Workers

The Shift from Passive Inbound to Autonomous, Reviewable Deal Generation

Outflow Build Log: Designing Human Approval Boundaries for Autonomous Agents

Automate Execution, Not Responsibility: The Operating Code of a Builder

How Answer Engines Choose Their Citations

llms.txt and the Machine-Readable Contract

Deterministic Rendering for AI Crawlers

Agents That Wait: Interface Design for Approval Gates

Why GEO Emerges Now

Venture Studios in the Agentic Era: The Product Is an Agent

Geolify Build Log: The Least Glamorous Pillar

Build Log: One Evidence Standard, Two Surfaces

Shared Gravity, Independent Orbits

Why First Light: Seeing Before It Is Obvious

Self-Hosted by Default: Fonts to Discovery Files

Outflow Build Log: Signals Worth Acting On

Why We Publish the Record

Have data or observationsfrom the agentic frontier?

Have data or observations
from the agentic frontier?