BossAgentsQianlin Brain · 2-Minute Demo Video Script (L1 Deterministic Routing, Reproducible)
All three instructions follow L1 rules directly (auto_execute=true), independent of LLM inference. Verified on 2026-08-03:
S1 → 3 BOMs + 4 parts, Medium; S2 → Full scan of 1235 Parts, 5739 issues, 3.3s; S3 → 20 audit logs.
S1 execution will automatically write to governance audit, feeding directly into S3, forming a closed loop.
I. Pre-Recording Checklist
| # | Check Item | Verification Method |
|---|---|---|
| 1 | Server port 3006 is normal | https://ylxt.chat/api/change/stats returns 200 |
| 2 | Part P-1001 and its 3 associated assembly data are complete | ASSY-A001/A002/A003 exist |
| 3 | Quality dashboard data is initialized, governance audit logs ≥ 20 | /api/quality/alerts + /api/audit-logs |
| 4 | Browser dark theme, address bar ylxt.chat visible throughout recording | Adjust before recording |
Pre-warm: Execute 'Query the impact scope of P-1001' once to ensure the path is warm (server had hanging issues before, pre-warming avoids this).
II. Second-by-second Storyboard Script (Basic Version)
| Time | Visual & Action | Voiceover / Caption | Core Proof |
|---|---|---|---|
| 0:00-0:15 | Black screen opening → Type ylxt.chat in browser → Homepage displays digital agent matrix, showing "62 agents online" | 【Voiceover】This is not a demo, but a real, live product you can experience anytime. 43 industrial digital agents are ready for direct, no-login experience. | Product is live, not fictional |
| 0:15-0:55 | Go to AI Assistant → Enter 'Query the impact scope of P-1001' → Wait 1-3 seconds for real response → Output 3 BOMs, 4 affected parts | 【Subtitle】Scenario 1・BOM Change Impact Analysis; Complete full-link impact analysis in 10 seconds | L1 rule direct pass-through, full-link risk deduction capability |
| 0:55-1:35 | Enter 'Check Part data quality' → Return five-dimensional quality score → Switch to data asset dashboard → Display quality warning red/yellow/green lights | 【Subtitle】Scenario 2・Panoramic Governance of Data Assets; Automatic scanning with thousands of rules, five-dimensional scoring, proactive risk warning | Govern while doing business, data assetization capability |
| 1:35-2:00 | Enter 'View recent governance records' → Display 20 audit logs → Flash digital employee matrix → End with freeze-frame of product domain | 【Voiceover】All governance operations are traceable, 43 digital employees run on the client side, data is fully controllable | Closed-loop governance is auditable, compliant and trustworthy |
III. Three Core Demo Instructions
| Scenario | Conversation Command | Routing Level | Expected Output |
|---|---|---|---|
| BOM Change Analysis | Query the impact scope of P-1001 | L1 Rule Engine | 3 BOMs + 4 affected parts |
| Data Quality Inspection | Check Part data quality | L1 Rule Engine | Five-dimensional quality score + problem classification statistics |
| Governance Record Traceability | View recent governance records | L1 Rule Engine | 20 audit logs |
Test Details (2026-08-03):
- S1: chat 93ms (L1 passthrough) + execute 488ms → Affected BOMs:3 / Affected Parts:4 / Severity:Medium
- S2: chat 851ms + execute 5.2s → totalChecked:1235 / issuesFound:5739 / autoFixable:4407 / manualFix:1332 / 3.3s
- S3: chat 79ms + execute 57ms → 20 audit logs
Known Minor Issue: In S1 response, affectedParts[].name may be ?????? (SCSAI Chinese encoding). Highlight item_number P-1001 during recording to avoid.
IV. Visual Upgrade Version - Second-by-Second Script (No change to workflow, only visual enhancement)
| Duration | Screen & Action | Voiceover / Subtitles | Optimization Notes |
|---|---|---|---|
| 0:00-0:15 Opening | Black screen with white text opening→Type ylxt.chat in browser→Homepage displays digital employee matrix, revealing "62 agents online" | [Voiceover] This is not a demo, but a live, ready-to-experience real product. [Subtitle] 43 industrial digital employees, experience directly without login | Maintain authenticity, reinforce perception of 'finished product, not prototype' |
| 0:15-0:55 Scene 1 | Enter AI Assistant→Input 'Query the impact scope of P-1001'→Keep 1-3 seconds of real waiting→After output, dynamically highlight the association topology lines from P-1001 to 3 BOMs and 4 parts, with relationship graph visualization expanding | [Subtitle] Scene 1・BOM Change Impact Analysis [Voiceover] Previously, this required a process engineer 2 days of manual ledger checking; now, 10 seconds. [Subtitle] 739k relationship graphs, N-hop reasoning completed in seconds | Added dynamic visualization of relationship graph, making abstract capability visible; added human-machine comparison to emphasize efficiency value |
| 0:55-1:35 Scene 2 | Input 'Check Part data quality' → Five-dimensional scoring presented with dashboard animation, needle sweeps from low to high score zone, problem categories pop up simultaneously → Switch to data asset dashboard → Display quality warning red/yellow/green lights | 【Caption】Scene 2・Zero-Cost Governance of Data Assets 【Voiceover】In the past, data governance was a separate project; now, it's embedded in every daily operation. 【Caption】Rule engine with tens of thousands of rules, automatic verification and repair | New dashboard animation visually demonstrates "governance effect"; voiceover highlights the core difference of "zero cost, invisibility" |
| 1:35-2:00 Closing | Input 'View recent governance records' → Display 20 audit logs → Quick flash of digital employee matrix → Closing freeze frame with domain name and slogan | 【Voiceover】All operations are auditable and traceable, with 100% data control. 【Closing Caption】Qianlin Brain・Empowering the Phosphorus Chemical Industry with Its Own Industrial AI Agent | Reinforces the trust points of auditability and high security, concluding with industry value. |
All special effects can be achieved through post-production editing; recording only captures real system operations.
V. Visual Specifications for Fallback Plans
| Scenario | Handling |
|---|---|
| On-site network anomaly | Switch to pre-saved screenshot demonstration |
| Badge in the upper right corner of the screen | Semi-transparent gray small text "Simulated Environment Demo", font size no larger than 1/20 of the screen, not interfering with main content |
Demonstrates professional rigor while avoiding judges mistaking it for a system failure.
