Stenobird

Episode

986: Building Hardware is Hard but AI Agents Help, with Kishore Subramanian

Podcast
Super Data Science: ML & AI Podcast with Jon Krohn
Published
Apr 24, 2026
Duration seconds
1783
Processing state
processed
Canonical source
https://www.podtrac.com/pts/redirect.mp3/chrt.fm/track/E581B9/arttrk.com/p/VI4CS/pscrb.fm/rss/p/traffic.megaphone.fm/SUPERDATASCIENCEPTYLTD5532445330.mp3?updated=1777027157
Audio
https://www.podtrac.com/pts/redirect.mp3/chrt.fm/track/E581B9/arttrk.com/p/VI4CS/pscrb.fm/rss/p/traffic.megaphone.fm/SUPERDATASCIENCEPTYLTD5532445330.mp3?updated=1777027157
JSON
/v1/public/podcasts/super-data-science/episodes/986-building-hardware-is-hard-but-ai-agents-help-with-kishore-subramanian
Markdown
/podcast/super-data-science/986-building-hardware-is-hard-but-ai-agents-help-with-kishore-subramanian.md

Actions

  • POST https://stenobird.com/v1/public/podcasts/super-data-science/episodes/986-building-hardware-is-hard-but-ai-agents-help-with-kishore-subramanian/transcription-requests
    Idempotently request low-priority transcript generation for this episode.
  • GET https://stenobird.com/podcast/super-data-science/986-building-hardware-is-hard-but-ai-agents-help-with-kishore-subramanian.md
    Read the agent-friendly Markdown representation of this episode resource.

Summary

Hardware manufacturing faces massive costs when errors are discovered late in the lifecycle, unlike software which can be patched. This episode explores how AI agents and Agentforce 360 enable 'shifting left' to catch compliance and quality issues during the design phase.

Topics

  • Product Lifecycle Management
  • AI Agents
  • Hardware Manufacturing
  • Agentforce 360
  • Quality Management Systems
  • Enterprise AI
  • Predictive Maintenance
  • Software Engineering

Highlights

  • Main idea: AI agents can act as automated reviewers for engineering change orders to prevent expensive hardware recalls
  • Practical takeaway: Use Agentforce 360's reasoning engine to automate complex tasks like breaking down problems into actionable sub-tasks
  • Failure mode: In hardware, unlike software, you cannot simply push a patch, making early-stage quality assurance critical
  • Main idea: Leveraging existing platforms like Salesforce provides the necessary security, governance, and data plumbing for enterprise AI
  • Practical takeaway: To deploy AI, first identify which manual tasks can be automated and ensure you have the high-quality data required to support them

Chapters

  1. 1:00 Introduction to PLM and QMS: An overview of Product Lifecycle Management and Quality Management Systems for physical goods like medical devices and consumer electronics.
  2. 5:25 The High Cost of Hardware Errors: Comparing the cost of fixing errors in software versus the massive financial impact of errors in physical manufacturing.
  3. 7:40 Introducing Propel One AI: A deep dive into how Propel uses AI to manage product information and engineering changes.
  4. 14:35 Building on Agentforce 360: How Propel leverages Salesforce's infrastructure for data security, governance, and the Atlas reasoning engine.
  5. 21:05 Framework for AI Deployment: A methodology for breaking down problems into tasks and determining which parts are suitable for AI automation.
  6. 25:20 Mindfulness and Engineering: A discussion on how yoga and meditation can enhance creativity and cognitive performance in AI development.