Research

We publish engineering write-ups from our work advancing autonomous search optimization: the systems we build, the failures we document, and the field data behind both. Entries are dated to the work they describe and compiled from the engineering record.

University Partnerships

We partner with programs that value forward-thinking, hands-on training in applied search systems. If that describes yours, reach out; we would welcome the chance to advance this research together.

Client Partnerships

The field data behind this work spans more than twelve months, and our SEO Command Center has run in production on client sites since July 2026. We do not train our models or run automated systems on client sites without consent. Clients who opt in, allowing their site to contribute to our experimental training data, qualify for partner pricing.

  1. October 24, 2026The Closed Loop: Architecture and Production Record of an Autonomous SEO SystemThe SEO Command Center after more than three months in production: the loop and its chains, agents inside client repositories, a measured-outcome decision system, realtime keyword evidence and eleven Full Autopilot properties.Publication
  2. October 18, 2026Where AI Search Visibility Comes From: Five Months of Generative AI Data Across 40 PropertiesGoogle’s first AI Overview and AI Mode report, read for 40 properties beside Bing and GA4: a 2.6× expansion baseline, a share of search inside the answer that more than doubled, and the answer-shaped pages that carry it.Publication
  3. September 30, 2026The Living Strategy: Ninety Days of Autonomous SEO Across Eight PropertiesA longitudinal case study of eight properties whose SEO strategy is recompiled nightly from search, rank, competitor and site evidence: the algorithm advances, verified corrections and windowed results.Publication
  4. August 17, 2026The Strategy Compiler: A Closed-Loop Context Architecture for Autonomous SEO AgentsHow search, analytics, technical, local, site, and research evidence compile into a living strategy that constrains specialized agents, verifies local-first execution, and uses measured outcomes to govern its next state.Publication
  5. August 14, 2026Founding Data: What 12x Impressions Growth Actually Looks LikeSix months of Google Search Console data from two properties running our autonomous SEO pipeline: a 17x impressions curve, a flat one, and the reasons we publish both.Conclusion
  6. August 14, 2026Local-First: Running Autonomous SEO Work on Machines We Already OwnA durable job queue that routes AI work to desktop hardware on subscription pricing first and scale-to-zero cloud second, together with the measured economics that forced the design.Milestone
  7. August 13, 2026Anatomy of a Silent Failure: Five Days of a Dead IsolateA scheduled step hung past its runtime kill ceiling and wrote no ledger rows, not even deferrals, for five days. We present the forensic analysis and the observability doctrine that emerged from it.Publication
  8. August 13, 2026Reserve, Settle, Floor: Metering AI Spend as a First-Class ContractEvery expensive bug in an autonomous platform is eventually a money bug. We document the reservation ledger, the under-charge floor, and the timeout arithmetic that keeps agent spend honest.Publication
  9. August 10, 2026Guardrails That Failed by Being Too StrongOur file-allowlist guardrail held every blocked post on a file the agent legitimately needed. The resulting guarded-open design combines advisory scope, hard fences, and always-run verification.Publication
  10. August 2, 2026Honesty Engineering: Teaching an Autonomous Platform to Tell the TruthSampled metrics presented as sitewide, two meters on one budget, fabricated URLs, and phantom drift banners: a defect taxonomy for systems that report on their own work.Publication
  11. July 29, 2026Context Is an Engineering Surface: How AI Agents Read a WebsiteCompiled context packs, read-order indexes, and the incident that proved "the file exists in the workspace" is not the same claim as "the agent will read it."Publication
  12. July 18, 2026Porting a 46,000-Line SEO Engine Into a Multi-Tenant Platform in Five DaysCensus-first planning, a fourteen-PR chain, and adversarial review panels that caught two SSRF vulnerabilities and a cross-tenant traversal before any of it reached production.Milestone