YUAN transforms complex video sources into production-ready edge intelligence.
Establish YUAN's Video-Native Edge AI category in the US market and increase qualified design-in opportunities, PoC-to-production conversion, and design wins.
74
Direct · adjacent · ecosystem database6
Core marketing messages to validate200
ISV · Camera OEM · Robotics · SI · AOI7 / 7
Stage gates A–GInbound-led demand for capture & edge platforms
Product pages, SEO topic cluster, engineering content, and developer assets capture existing demand for video capture, Jetson platforms, and the NexVDO SDK.
ABM-led design-in motion for named accounts
Solution kits, reference architectures, and LinkedIn + outbound plays target 200 named accounts toward architecture reviews and design wins.
What the strategy can already prove
Market benchmark seed
The 30-company seed and 74-company scorecard remain intact; ten official-source expansion packs add 360 unique-domain evidence records across industrial cameras/connectivity, frame grabbers/recording, spectral/thermal imaging, stereo/depth/calibration, visual inspection, retail vision topology, roadway sensing/control, robotic perception, physical AI, rugged industrial edge compute, deterministic timing/TSN/industrial networking, precision agriculture, production inspection, laboratory/digital-pathology imaging, wearable vision, industrial/fleet/rail/construction/care/energy/security/medical vision, sports video, maritime/heavy/aerial autonomy, inspection robotics, professional-video transport and edge ecosystems without altering the seed. A deterministic quality audit confirms complete core and source-type fields, 59 dated primary sources, 20 separate dated current-signal supplements, 16 medium-confidence refresh records and 41 entity-hierarchy heuristic flags. A 360-URL point-in-time health audit ends with zero 404/410 sources and eight transport-level manual retries; 81 high-fit rows remain in the undated refresh queueYuan public evidence
Six public capability claims verified; chip-level, deterministic-performance and Re-ID ownership claims still need internal proofInternal benchmark data
Latency, stream count, frame-drop, thermal, utilizationNamed customer proof
Case studies require customer authorizationMessaging hypothesis
Six pillars now map to segment offers, permitted claims and a 14-item proof backlog; customer language remains unvalidated in-marketABM named-account universe
Canonical T05 universe now contains 200 unique-domain North American accounts: 169 United States and 31 Canada, with exactly 50 accounts in each user-defined vertical tier (Surveillance / Smart Infrastructure, Medical Imaging / Surgical Video, Defense / Dual-Use Autonomy, Robotics / Physical AI). All rows retain company-level official evidence, inference state, unknown current trigger, no contact PII and research_required readiness. A 205-URL point-in-time audit reports 185 reachable, 2 browser-verified after locale/WAF transport differences, 13 restricted, zero known 404/410 and 5 manual retries; the original 100-account CSV is preserved as a v1 archiveApollo current-client account audit
Historical read-only company-level snapshot covers 4,112 Current Client records, the original 100-account exact-name audit, six overlapping keyword pools, 30 decision rows and a separate 30-company official-evidence pack. A consolidated 48-company master has 24 research_now, 14 validate_account_structure, 6 ecosystem-only, 1 evidence-validation and 3 do-not-develop rows. The new 200-account North America universe was crosswalked only against this prior snapshot and has 7 matches (Eagle Eye Networks, Motorola Solutions, Forterra, Locus Robotics, RightHand Robotics, OSARO and Symbotic); 193 net-new accounts are explicitly not_checked_in_live_apollo, not treated as absent or non-customers. Zero current programs or outreach-ready accounts were confirmed. No Apollo MCP endpoint was exposed in this run, so no live query, PII export, CRM write or outreach occurred; the specification-only read-only MCP contract remains the next authorized pathAgentHarness market and Apollo run manifest
Machine-readable manifest records the immutable seed hash, 360-company market database, 48-company Current Client master, 24 research-now decisions, 10 market-to-Apollo decision insights, connector limitation, safety boundaries, next execution order and SHA-256 hashes for 32 core deliverables; the 39-URL supplemental current-signal health audit has zero known broken or retry-required sourcesBilingual ABM content backlog
300 unique English / Traditional Chinese ideas are mapped evenly across five segments, stages and pillars with formats, personas, conversion paths, priority scores and evidence gates; idea status does not authorize publicationMeasurement and attribution contract
A specification-only 21-event taxonomy covers website, account readiness, design-in pipeline and recognized revenue with server-side de-duplication, MQL/MQA/SQL definitions, account-level influence views, PII prohibitions and Phase 2 implementation gatesStage gates A–G
Source Audit+Claim Ledger
Competitor Deep-Dive Scorecard
Messaging Architecture
Website IA+Page Briefs
Content Operating System
ABM Account Model
Measurement+Attribution
Latest research documents
Every agent run lands its findings here as versioned markdown — the dashboard reads them directly from the repo.
Phase 1 — Research & Strategy Document Site
Next: Phase 2 — GA4 / HubSpot / Apollo / LinkedIn integrations for ABM tracking