FAQ
Practical questions, answered plainly. Buyer problems live on the Questions page; open research questions live at /research; this page is operational.
What is Upstream Zero?
A research company studying commercial evaluation — how organizations are assessed, recommended, validated, and eliminated before selection. The full, canonical answer lives on the What is Upstream Zero? page.
Who is it for?
Organizations that suspect AI-mediated evaluation is affecting whether they are recommended, shortlisted, or eliminated — and researchers, buyers, and machine systems who want inspectable evidence about how commercial evaluation behaves. The problem map lives on the Questions page.
What do you actually do?
Two connected things. Research: observe and measure evaluation behavior, publishing every claim at its evidence tier. Commercial: apply those methods to a client's situation — observing how evaluators assess them, which requirements they appear to satisfy, where evidence gaps eliminate them, and how stable the outcomes are.
How is this different from SEO, AEO, GEO, or AI-visibility tools?
Those disciplines optimize appearance: rankings, citations, mentions, visibility. Upstream Zero measures evaluation: what happens between appearing and being selected. It does not optimize websites for rankings and does not sell visibility improvement. Normal technical hygiene matters, but it is infrastructure, not the work.
Do you optimize websites?
No. An engagement may reveal that evaluators cannot verify a requirement you satisfy, and you may choose to publish better evidence — but the deliverable is the measurement and diagnosis, not site changes, and no ranking or recommendation outcome is promised.
Do you guarantee recommendation changes?
No — structurally, not just rhetorically. Engagements promise deliverables (artifacts), never evaluator behavior; the site's object model has no field an outcome-promise could live in, and the build fails if research objects cite commercial ones.
What does an engagement include?
Provisional engagement shapes are published with their deliverable specifications on the Services page: audit, requirement and evidence gap analysis, stability measurement, ongoing monitoring, machine representation advisory, and research partnership. Names, scopes, and pricing are provisional pending founder confirmation (FD-6), and nothing is bookable until that decision is made.
What evidence do you need from a client?
The claims you make commercially, the requirements you believe you satisfy, and the artifacts that could verify them — plus permission to observe how evaluators currently describe and assess you. The contact page lists what to bring to a first conversation.
What is Client Zero?
Upstream Zero itself. The observatory runs its methods on itself first: this site is the first measured subject, its publication is treated as an intervention in the environment it studies, and the first experiment (EXP-0001) observes whether its own published objects propagate into evaluator behavior. The concept: Client Zero.
How is the research connected to the commercial work?
The research produces methods; the commercial work applies them as measurement and diagnosis. A firewall keeps the direction one-way: research objects never cite commercial objects (build-enforced), capabilities stay marked experimental until they derive from published method, and measured engagement outcomes will publish as tiered evidence with client consent.
How do you handle corrections?
Corrections are published with the same dignity as findings: a recorded revision object, never a silent edit or deletion. Superseded material remains public with a forward pointer, and the public git history makes every change diffable. The full policy is on the Philosophy page.
How can someone contact or work with Upstream Zero?
Through the contact page, which describes what to bring and what a first conversation covers. There is no sales funnel; the register is a calm, specific conversation about what can currently be measured.