# Industrial examples from external sources

What current visual and document systems demonstrate, and what an enterprise agent still needs.

Author: Murali Sid (https://linkedin.com/in/muralisid)
Source: https://www.agenticarchitectureskills.com/research/industrial-examples
Published: 2026-09-17
Updated: 2026-10-15
Licence: CC BY-SA 4.0, https://creativecommons.org/licenses/by-sa/4.0/
Cite as: Murali Sid. Industrial examples from external sources. Agentic Architecture Skills. https://www.agenticarchitectureskills.com/research/industrial-examples

---


These examples identify documented building blocks and vendor-reported applications. They do not establish the business results of the proposed agent designs in this guide. Sources were checked on 17 September 2026.

## Video investigation [#video-investigation]

NVIDIA's [video-search and summarization blueprint](https://build.nvidia.com/nvidia/video-search-and-summarization/blueprintcard) documents video analysis, search, summarization, and question answering. This supports considering a visual evidence service in a larger investigation. Identity, work-order permissions, coaching policy, and verified outcomes remain separate enterprise responsibilities. Vendor documentation.

## Visual workflows [#visual-workflows]

[Roboflow Workflows](https://docs.roboflow.com/workflows) describes configurable computer-vision processing for images and video. A fixed detection and alert pipeline may be sufficient for a known event. An adaptive agent should be added only when choosing the next evidence-gathering step improves the result. Vendor documentation.

## Document evidence [#document-evidence]

LandingAI describes [visual grounding for document extraction](https://landing.ai/llms/visual-grounding-and-auditability-how-landingai-ade-makes-every-extraction-defensible), including locating extracted information on the source page. That can help a reviewer inspect an extracted field. It does not establish that a document is authentic, current, or sufficient to satisfy a standard. Vendor documentation.

## Vegetation planning [#vegetation-planning]

AiDASH reports [satellite-assisted vegetation management with National Grid](https://www.aidash.com/resource/aidash-transforming-vegetation-management-for-national-grid-using-satellite-analytics-and-ai/). Treat the account as a vendor case study. It is relevant to inspection prioritization, but any extension to adaptive investigation or autonomous work requests needs separate evaluation.

## What to compare before adopting [#what-to-compare-before-adopting]

Check the task, permitted input data, required resolution and freshness, edge or cloud constraints, interfaces, total cost, and human review. Ask whether the system detects, extracts, follows a fixed workflow, or chooses its next action. Those are different capabilities.

Explore proposed [safety coaching](/use-cases/safety-coaching), [vegetation inspection](/use-cases/vegetation-inspection), and [compliance evidence](/use-cases/compliance-evidence).