ANS-100/P — Manufacturing & Mobility Sector Calibration Note

Drafted 30 August 2026 · Ratification not separately logged — see the note on the rulings index · written before the first manufacturing company was scored · Rater 1 · Companion to ANS-100/P v2.1 (frozen anchor set) · Sector: Tata Motors, Mahindra & Mahindra, Bajaj Auto, TVS Motor, Larsen & Toubro, Tata Steel, JSW Steel, UltraTech Cement

Ratification log: C11 ratified 30 Aug 2026 — Q32 scores only on disclosed third-party AI-product sales; scoreable for Info Edge and Zoho, N for the other six. Consumer-internet files final as drafted pending right-of-reply.

1. Why this is the biggest calibration jump yet

Every sector so far runs on software. This one runs on steel, cement, engines and factories — where AI appears as process control, predictive maintenance, computer vision on production lines, digital twins, and driver-assistance features in shipped products. The instrument's proxies (platforms, agents, model strategy, developer surfaces) will systematically under-read genuine industrial AI unless re-anchored. The LTTS lesson at full sector scale: score the sector's own evidence forms, or measure disclosure convention and call it nativeness.

Equally, this sector must not get free points. Classical process automation, PLCs, SCADA and robotic welding predate AI by decades. Automation is not AI. The rulings below cut both ways deliberately.

2. Proposed rulings (ratify individually: D1 ☐ … D10 ☐)

D1 — Q25 economic pair. Revenue per employee, read with the sector's own productivity metrics where disclosed: capacity utilisation, OEE (overall equipment effectiveness), yield/first-pass-yield, energy or specific-consumption per tonne, and unplanned-downtime hours. A1 unchanged: level 3 needs a number linking named AI systems to one of these. Commodity-price cycles, capex phases and demerger effects are named confounds.

D2 — Q27 is this sector's strongest scoring opportunity. Industrial firms publish quality and downtime metrics. Q27 admits: defect-detection rates from vision systems, predictive-maintenance-driven downtime reduction, yield improvement attributed to AI process control, and autonomous-operation shares in plants or mines. Excluded: conventional automation percentages, robot counts, and general digitisation figures — the D-analog of B2/C2.

D3 — "Automation is not AI" (the anti-free-points rule). PLCs, SCADA, MES deployments, robotic arms and ERP rollouts score nothing in D2 without evidence of learning, prediction, or perception — i.e., models trained on data. The file must state which of these the evidence shows. Industry-4.0 branding alone is level 1.

D4 — Q6/Q7 physical-autonomy anchor. Production autonomy means AI systems taking or directly triggering physical actions: closed-loop process control, autonomous inspection with reject authority, self-optimising lines, autonomous material handling. Advisory dashboards are level 2 at most.

D5 — Q33 product-embedded AI (the sector's distinctive question). For mobility firms, AI shipped inside the product — ADAS, autonomous features, connected-vehicle intelligence, battery-management AI — is first-class Q33 evidence when disclosed with volumes (vehicles shipped with the feature) or as a named revenue line. For materials firms, the analog is AI-derived product quality grades or customer-facing digital platforms.

D6 — Q31 exposure re-anchored. Manufacturing's AI exposure is not disintermediation; it is competitive cost displacement (rivals reaching lower cost curves via AI process optimisation) and, for mobility, product obsolescence (software-defined vehicles, autonomy stacks). Mitigation = disclosed AI-driven cost programmes, SDV/autonomy roadmaps with volumes, or proprietary industrial-data advantages.

D7 — Q18 in a safety-regulated setting. ISO 9001/14001/45001 and functional-safety standards (ISO 26262) are not AI governance — level 2 requires AI-specific governance; ISO/IEC 42001 or audited AI controls reach 3. But safety-critical AI validation evidence (ADAS validation regimes, model-release safety cases) counts toward 2–3 where disclosed with specifics: this sector has the strongest legitimate claim to safety-engineering-as-AI-posture, and the file must distinguish it from generic quality certification.

D8 — Q12/Q24 capital reality. Manufacturing discloses capex in detail. The A4 ≥5%-of-TTM-revenue line applies, but for Q24 level 3 the AI component must be identifiable within digital/technology capex; whole-plant capex is not AI investment. Edge/on-premise compute for factory AI is legitimate Q12 self-hosting evidence.

D9 — Entity rulings. Tata Motors: score the entity carrying the demerged passenger-vehicle/EV business, and state the demerger basis and date explicitly in the file header (the universe draft flagged this; resolve at pack assembly). M&M: auto and farm segments; Tech Mahindra is separately in the universe and its evidence must not cross over. L&T: the engineering/construction parent; LTTS and LTM are separately scored and excluded. Tata Steel, JSW Steel, UltraTech: standalone-plus-India-operations as reported.

D10 — Signal library additions. Integrated annual reports (India's manufacturing sector discloses well here), sustainability/BRSR reports (energy and yield metrics live here), plant-digitisation case studies with named technology partners, World Economic Forum Global Lighthouse Network designations (independent, audited third-party evidence of advanced-manufacturing deployment — first-class A7(c) grade where held), patent filings, and OEM technical papers.

3. What does not change

All v2.1 anchors, P-caps, evidence floor, 18-month window, A1–A10, ceiling arithmetic, evidence-language rule. Application, not amendment.

4. On ratification

Score in universe order: Tata Motors → M&M → Bajaj Auto → TVS → L&T → Tata Steel → JSW Steel → UltraTech. Expect lower D1/D3 across the board and higher D2/D6 than software sectors; the Lighthouse designations and OEE disclosures are where this sector's real evidence lives. Sector synthesis after all eight.

Published 9 September 2026 under CC BY-SA 4.0 as part of the Edition One replication material. Status changed from "proposed" to "ratified" on publication: the note was drafted on 30 August 2026 and written before the sector was scored; the ratification record for this set is described on the rulings index. No ruling text is changed by this publication.