Studies, articles and open evidence on AI employees, memory and accountable decisions. Explore the work behind JARVIS and System Brain.
Original questions. Transparent methods. Evidence with a stated scope.
What happens when an AI employee resumes work with saved memory and changed permissions? Compare four memory conditions, response-format failures and recovery decisions in a controlled synthetic evaluation.
Compare fixed rules, a single model call, a System Brain pattern and self-review on synthetic decisions. Explore when an AI employee should act, request help or stop.
These computational pilots report controlled synthetic evaluations. Each study includes its methods, limitations and the scope of ZEISMO’s internal verification.
Long-form research on intelligent systems and the conditions in which they operate.
An evidence review of political disruption and overseas growth, with synthetic scenarios and practical planning worksheets for e-commerce resilience.
Read the reportThe foundational technical preprint on an English-speaking JARVIS interface, specialised agent roles, memory and accountable decision-making.
Read the reportA practical dependency test for shared warehouses, payments and operating access across international markets.
Read the articleUpdated 17 Oct 2026 · Version 1.1
An updated memory-lifecycle article connecting expiry and revalidation to the published restart pilot, with a clearer follow-up test.
Read the articleUpdated 17 Oct 2026 · Version 1.1
An updated engineering perspective connecting physical dependencies and bounded AI continuity to the published decision and restart pilots.
Read the articleExplore how role contracts, curated skills, scoped tools and memory fit together in the proposed System Brain architecture.
Read the articleSix documented evidence briefs translate anonymized operational measurements into research questions on AI usage, reliability, corrective work and resilience.
Each brief identifies its method, measured scope and limitations.
Role-specific agents, memory and accountable decisions.
Explore the programme02 / RESEARCH TOPICSBrowse the lab’s connected research themes and evidence.
Browse the topics03 / AI IMPACTExplore published pilots and proposed studies in the AI Impact Observatory.
Visit the observatoryPractical research templates, evidence checklists and integrity tools. A useful starting point for your next experiment.
Free resources. No model subscription required.
A space for thoughtful research discussion, reproducibility questions and potential collaboration.
Tell us about your research interests, a method worth testing or a correction we should examine. Enquiries are recorded for the editorial team.
Start a research enquiryZEISMO builds software for the work businesses do every day: WordPress tools, CRM, WhatsApp AI assistants and custom integrations. Visit our company website to explore the products and discuss your next project.
Explore ZEISMO products & servicesTools for managing and improving your website.
CRM and WhatsApp AI assistants for daily workflows.
Software and integrations shaped around your business.
Follow each marker to the method, record or policy behind it.
ZEISMO practice markers. Verification scope is recorded with each publication.
