Library review — the flag index.

Open flags

• Source-PDF printed anomalies preserved verbatim in the imported text: two mis-numbered/mis-copied table captions, and a repeated subheading (page 5 reuses "Common signs of governance immaturity" over bullets that actually list benefits). Fixed 2026-08-28 in the approved hygiene pass.
• No publication date in frontmatter — undated items sort to the end of the shelf.

Open flags

• No publication date in frontmatter — undated items sort to the end of the shelf.

Open flags

• No publication date in frontmatter — undated items sort to the end of the shelf.

Open flags

• No publication date in frontmatter — undated items sort to the end of the shelf.
• Context: the legacy blog twin of this paper was dropped in the migration (this paper owns the topic); the near-dup 'AI Impact Activation' paper was deleted 2026-07-07.

Open flags

• MIGRATION PLACEHOLDER: abstract only — needs the source PDF (then the gs/sips cover recipe) and full text; blocked externally (BACKLOG §2). Shows 'PDF coming soon' and a typographic cover meanwhile.
• Go-public gate EXECUTED on the go-public branch: flipped draft:true because the paper is still a stub blocked on its source PDF (BACKLOG §1). When the finished paper lands, revert to draft:false and restore its /library bullet in public/llms.txt.

Open flags

• ROI ranges withheld per the no-unsourced-stats rule (inventory -10-25%, lost sales -15-30%, forecast accuracy +10-30pts, OTIF +5-15pts, 3-8x returns over 18-36mo) — the source case numbers trace to a file named 'A2go-Generic-Case-Study-based-on-JBS.docx', i.e. synthetic. Stephen decides if any range returns in hedged form (scope question raised 2026-07-15 alongside the TACO/JBS verification).
• OKR example targets (15% finished-goods reduction in 12 months; turns 5x to 7x) are illustrative hypotheticals from the source — confirm comfort presenting them as an example.

Open flags

• Ships with zero first-party evidence: approve ONE hedged engagement example (the beef supply-chain pricing collaboration from the retired quick-wins post is the strongest candidate) or consciously accept argument-only framing. Do not reinstate the vendor-blog percentages or unsourced case bullets.
• Verify ADIP shipped-state phrasing matches reality: 'built to work alongside your ERP', agents that 'monitor conditions and support specific planning decisions'.
• McKinsey-Oxford large-scale-IT-projects reference is qualitative and linked to the mckinsey.com root — source a stable deep link if preferred.

Open flags

• Names Databricks positively as an example platform — confirm comfort. Table-caption headings are numbered 2-5 but the first table is unnumbered; normalize if desired.

Open flags

• "the 'Slowness Tax' alone drains up to 5% of annual revenue" — 'Research estimates' with no named source
• "A2go customers have reduced Master Production Scheduling cycles from 18 hours to 15 minutes" — unnamed-customer results claim needing verification
• Results table ("Forecast Error Reduction 30–50%", "Lost Sales Reduction up to 65%", etc.) presented as "outcomes from deployed systems" with no verifiable sourcing
• "A2go's ADIP platform deploys purpose-built agents in 3–6 months, with measurable ROI typically visible within 6 months" — unsourced results/timeline claim
• "BCG reports that agentic systems already accounted for 17% of total AI value in 2025, rising to a projected 29% by 2028" — specific figures cited to BCG without a linked source
• Markup defect preserved from WordPress: 'They manage outcomes.The Compounding Advantage' — a lost h2; restored as a heading 2026-08-28 in the approved hygiene pass.

Open flags

• "One industrial manufacturer reduced master scheduling time from 18 hours to 15 minutes... Inventory costs dropped by 30 percent. The cash cycle improved by 25 percent. The return on investment reached 400 percent." — unnamed customer results claim, no source; needs verification.
• "Deployment timelines... are typically three to six months from kickoff to initial deployment, with measurable return on investment achievable within six months of go live." — unsourced performance/timeline claim.
• "More than 80 percent of AI initiatives fail." — attributed loosely to IndustryWeek but no link/citation; verify origin.
• "The result was a reduction in data transmission volume of more than 90 percent" (Hitachi) — third-party case cited without a link to the source article.
• "Ford collaborated with IBM to deploy computer vision and Edge AI for real time vehicle body inspection across multiple production plants" — named-company claims (Ford/IBM, Hitachi, NVIDIA) repeated from IndustryWeek with no outbound citation links.

Open flags

• "TACO, the $400 million industrial manufacturer, reduced its production scheduling process from 18 hours across 24 spreadsheets to 15 minutes, with a 400 percent return on investment." — named customer + specific results claims needing verification
• "JBS, the $8 billion global protein processor operating across 37 facilities and 27 regulatory regimes, deployed A2go..." — named customer claim needing verification
• "He estimated that around 94 percent of operational problems could be traced back to the system itself" — attributed to Deming but no citation/link
• "some researchers estimate upward of 90 percent, are rooted in systemic issues" — vague unsourced statistical claim
• "delivered in weeks, not months or years" / "can begin delivering meaningful results... within weeks" — time-to-value results claims with no supporting evidence

Open flags

• "explore how companies like TACO have achieved 30% inventory cost reductions and 25% improvements in cash cycle times" — names a customer and specific results claims needing verification
• "We have heard that 95% of AI pilots fail to deliver measurable results." — specific statistic with no cited source (likely the MIT NANDA study, uncited)
• "Forty percent of our agentic AI projects got canceled." — numeric claim presented as a quoted fear with no source (likely Gartner's 40%-by-2027 prediction, uncited)
• "begin generating value in weeks rather than waiting 12 to 18 months" / "first value is typically achievable in weeks, not months or years" — time-to-value claims with no supporting evidence
• "These are the documented results of companies that have already made this transition." — asserts documented results (master scheduling in minutes, etc.) with no documentation cited
• Legacy closing CTA kept verbatim: 'Visit a2go.ai ... companies like TACO have achieved 30% inventory cost reductions and 25% improvements in cash cycle times' — rewrite or cut (self-referential CTA + unverified named-customer numbers).