Changes for page POC Specification

Last modified by Robert Schaub on 2025/12/24 10:00

From version 1.1
edited by Robert Schaub
on 2025/12/23 18:19
Change comment: Imported from XAR
To version 3.2
edited by Robert Schaub
on 2025/12/24 09:59
Change comment: Renamed back-links.

Summary

Details

Page properties
Content
... ... @@ -4,24 +4,67 @@
4 4  
5 5  **Core Philosophy:** Prove the AKEL automation works before building production features. Each POC phase validates critical technical capabilities.
6 6  
7 +{{warning}}
8 +**Critical Innovation in POC1 Context-Aware Analysis:**
7 7  
10 +POC1 tests whether AI can understand that an article's overall credibility is not simply the average of its individual claim verdicts. An article with mostly accurate facts can still be misleading if it draws unsupported conclusions.
11 +
12 +See [[Article Verdict Problem>>Test.FactHarbor.Specification.POC.Article-Verdict-Problem]] for full discussion.
13 +{{/warning}}
14 +
15 +
8 8  == POC Phase Navigation ==
9 9  
10 -* **[[POC1 Requirements>>Test.FactHarbor.POC.POC1.WebHome]]** - Core Workflow with Quality Gates
11 -* **[[POC2 Requirements>>Test.FactHarbor.POC.POC2.WebHome]]** - Robust Quality & Evidence Deduplication
12 -* **[[Beta 0 Requirements>>Test.FactHarbor.POC.Beta0.WebHome]]** - User Testing & Production Readiness
13 -* **[[V1.0 Requirements>>Test.FactHarbor.POC.V10.WebHome]]** - Public Launch Requirements
18 +* **[[POC Requirements>>Test.FactHarbor.Specification.POC.Requirements]]** Complete POC1 & POC2 specification (v4.0)
19 +* **[[POC Summary>>Test.FactHarbor.Specification.POC.Summary]]** Executive overview
20 +* **[[Article Verdict Problem>>Test.FactHarbor.Specification.POC.Article-Verdict-Problem]]** Why context-aware analysis matters
14 14  
22 +== Roadmap Navigation ==
15 15  
16 -== Specification Documents ==
24 +* **[[POC1 Roadmap>>Test.FactHarbor V0\.9\.100 incremental.Roadmap.POC1.WebHome]]** Core Workflow with Quality Gates
25 +* **[[POC2 Roadmap>>Test.FactHarbor.Roadmap.POC2.WebHome]]** Robust Quality & Reliability
26 +* **[[Requirements Roadmap Matrix>>Test.FactHarbor.Roadmap.Requirements-Roadmap-Matrix.WebHome]]** Phase assignments
17 17  
18 -* [[POC Requirements>>FactHarbor.Specification.POC.Requirements]]
19 -* [[POC Summary>>FactHarbor.Specification.POC.Summary]]
20 -* [[POC1 Specification (Detailed)>>FactHarbor.Specification.POC.Specification]]
28 +== Key Specifications ==
21 21  
30 +**POC1 Goal:** Prove AI can extract claims and evaluate them with context-aware analysis
22 22  
23 -== Related ==
32 +**POC1 Output:** 5 components
24 24  
25 -* [[Roadmap>>FactHarbor.Roadmap.WebHome]] - Implementation roadmap and timeline
26 -* [[Requirements>>FactHarbor.Specification.Requirements.WebHome]] - System requirements
34 +1. Analysis Summary (context-aware, 4-6 sentences)
35 +2. Claims Identification (3-5 claims)
36 +3. Claims Verdicts (verdict + confidence + reasoning)
37 +4. Article Summary (optional)
38 +5. Usage Statistics (cost tracking)
27 27  
40 +**POC1 Philosophy:**
41 +
42 +* Context-aware analysis (article ≠ sum of claims)
43 +* Cost efficiency tracking (viable at scale)
44 +* Fail-fast learning (test critical capabilities)
45 +
46 +**POC1 Success Criteria:**
47 +
48 +* AI quality ≥70% (including context-aware assessment)
49 +* Cost per analysis <$0.05 (target)
50 +* Minimal manual editing (<30%)
51 +
52 +**POC1 → POC2 Decision Gate:** GO / NO-GO / ITERATE
53 +
54 +See [[POC Requirements>>Test.FactHarbor.Specification.POC.Requirements]] for complete specification.
55 +
56 +
57 +== POC vs. Full System ==
58 +
59 +|= Feature |= POC1 |= POC2 |= Full System
60 +| Scenarios | ❌ Deferred | ✅ Multiple | ✅ + Persistence
61 +| Evidence Display | ❌ None | ✅ With sources | ✅ + Quality scoring
62 +| Quality Gates | 🟡 4 simplified | ✅ 6 comprehensive | ✅ + ML validation
63 +| Database | ❌ Stateless | ✅ PostgreSQL | ✅ + Redis + S3
64 +| User Contributions | ❌ Read-only | ✅ Enabled | ✅ + Reputation
65 +
66 +**Development Time:**
67 +
68 +* POC1: 2-4 weeks
69 +* POC2: 6-8 weeks (additional)
70 +* Beta: 4-6 months (additional)