Supply Chain, Retail, and Consumer Analytics

ARA7 · Evidence review for team discussion · 27 October 2026

What is supported

Retail ordering and checkout systems support an industrial application discussion. Generic business computing is broader than this area.

Disclosed wholesale and warehousing components show substantial application hosts. The latest wholesale parent is suppressed; retail employment is large but nearly flat. Retained industry context and the patent observations describe different populations.

Scope and boundary check

The unchanged screen finds 998 narrow and 1,070 broad current-CPC candidates, versus 926 and 1,000 at issue. These are overlapping screening counts among utility grants with Arkansas inventors dated 2015–2025, not validated area totals or uncertainty intervals.

Primary-document examples

• US10028599B2: A checkout station coordinates input/output conveyors, sensors, scanning and payment terminals. Shows retail systems invention, without establishing a forecasting method, a university partnership or willingness to share store data.
• US10001973B2: A project-testing system estimates resource/cost work units and reports changes during testing phases. The general business-computing prefix extends beyond retail, consumer analytics and supply-chain operations.
• US12100387B2: An ordering interface uses speech tokens, learned features, product classifiers and purchase history to identify requested products. A concrete retail analytics subject. Its 2024 grant is outside the historical AI overlay ending in 2023 and cannot be treated as a match to that older model.

Decision to take next

Ask a retailer/distributor to define an ordering decision, usable item/location data and a frozen comparator.

Confidence and remaining checks

Source identities, event dates, input hashes and the reviewed technical passages are checked. Area-wide membership remains provisional. Patent ownership, current team readiness, facility/data access, partner demand and commercial outcomes require separate evidence. Low patent counts do not disqualify an area.

Review method: abstract and first independent claim for three purposefully/hash-selected documents per area. This is a bounded analytical review, not a statistical accuracy estimate, external expert sign-off or legal opinion. The seven growth areas and proposed commitments remain in place.