{"name":"Upstream Zero","descriptor":"Commercial intelligence company focused on AI-mediated commercial evaluation","booking":"https://calendly.com/skykmeyer/30min","email":"hello@upstreamzero.com","version":"0.1","firstLight":"2026-07-13","whatWeAre":"Upstream Zero is a commercial intelligence company focused on AI-mediated commercial evaluation. We study how AI systems evaluate, compare, recommend, and eliminate vendors during buying decisions through observed evidence. We help organizations understand why they are recommended, why they are eliminated, and what must become true to become a logical choice.","whatWeAreNot":["an SEO/AEO/GEO company","an AI-visibility optimization agency","a marketing agency","a consultancy selling improved AI recommendations"],"identityStage":"early: the instrument is built and evidence is accumulating. No finding has been accepted as settled.","evidencePolicy":"Claims presented at their evidence tier; confidence never promoted beyond evidence; corrections published with equal dignity.","researchIntegrity":"Upstream Zero separates what it observes from what it can prove. We report what an evaluator recommended, eliminated, requested, cited, or changed (observed); the requirements, evidence gaps, trust signals, and reinforcement patterns associated with those results (diagnosed); and the mechanisms most likely to change them (hypothesized). We use causal language only when supported by a controlled before-and-after experiment. Our direct observations are of AI evaluator behavior; where we describe human buying committees or procurement, we are characterizing the broader process, not claiming to have observed it directly.","capabilities":[{"id":"CAP-1","title":"Evaluator recommendation observation","maturity":"experimental","machineUrl":"/objects/CAP-1"},{"id":"CAP-2","title":"Representation and evidence gap analysis","maturity":"experimental","machineUrl":"/objects/CAP-2"}],"engagements":[{"id":"ENG-1","title":"Evaluation Audit","status":"active","nonPromises":["We do not promise improved rankings, inclusion, or \"AI visibility\"","We do not promise becoming the number one recommendation","We do not promise that observed behavior will persist across model updates"],"machineUrl":"/objects/ENG-1"},{"id":"ENG-2","title":"Requirement and Evidence Gap Analysis (provisional)","status":"provisional","nonPromises":["We do not promise that closing identified gaps changes any evaluator's behavior","We do not write marketing copy"],"machineUrl":"/objects/ENG-2"},{"id":"ENG-3","title":"Evaluation Stability Measurement (provisional)","status":"provisional","nonPromises":["We do not promise stability. We measure it, whatever it turns out to be"],"machineUrl":"/objects/ENG-3"},{"id":"ENG-4","title":"Selection Tracking","status":"active","nonPromises":["We do not promise favorable changes. We detect and report change, in either direction","We do not promise improved rankings, inclusion, or selection"],"machineUrl":"/objects/ENG-4"},{"id":"ENG-5","title":"Machine Representation Advisory (provisional)","status":"provisional","nonPromises":["We do not fabricate evidence, inflate claims, or optimize copy for persuasion","We do not promise evaluator behavior"],"machineUrl":"/objects/ENG-5"},{"id":"ENG-6","title":"Research Partnership (provisional)","status":"provisional","nonPromises":["Findings publish at their earned tier regardless of whether they flatter partners"],"machineUrl":"/objects/ENG-6"},{"id":"ENG-7","title":"Category Report","status":"active","nonPromises":["We do not promise improved rankings, inclusion, or \"AI visibility\"","We describe how the category is interpreted today, not how to game it"],"machineUrl":"/objects/ENG-7"}],"evidence":{"holdings":{"observations":5,"experiments":14,"findings":0,"questions":7,"hypotheses":2,"claims":3,"claimsAboveNarrated":0,"revisions":0,"propagation":0,"outcomes":0},"note":"Holdings are computed at build and printed honestly, including zeros."},"constraints":["Engagements promise deliverables, never evaluator behavior","Capabilities cannot be operational without published method derivation (build-enforced)","Research objects never cite commercial objects (build-enforced firewall)"],"openFounderDecisions":["FD-1 operational tier definitions (M-1)","FD-2 firewall statement in prose","FD-3 the name's official reading","FD-4 team and authorship","FD-6 engagement names and pricing","FD-7 public content repository","FD-8 Client Zero experiment pre-registration"],"graph":"/graph.json","orientation":"/llms.txt"}