OpenAI 和 Anthropic 共同看好的 FDE：AI 时代的新岗位出现，旧分工松动｜对谈 Rolling AI

Summary

This Shizilukou Crossing episode with Rolling AI partners Agan and Liu Kai argues that enterprise AI adoption is less like buying software and more like onboarding Digital Employees. The guests explain why OpenAI and Anthropic are emphasizing Forward Deployed Engineer work: enterprise AI needs business integration, knowledge governance, system integration, human-AI collaboration design, and incentive changes before agents can produce durable operating results. The episode extends the wiki’s agent and context themes into Frontline AI Enablement, Business-Led AI Transformation, and Service As Software.

Key Claims

• Forward Deployed Engineer work is presented as an AI-era implementation layer descended from Palantir: engineers or business builders enter the customer site, understand the real workflow, and help build usable systems rather than only selling a product.
• AI in enterprises should be treated as Digital Employees rather than ordinary tools; FDEs help those digital workers understand business context, access systems, learn from experts, and actually get on the job.
• Rolling AI frames the FDE job as three linked responsibilities: business integration, knowledge governance, and system integration.
• The episode argues that enterprise AI value often appears at the frontline, where store managers, salespeople, property managers, or service workers combine AI analysis with local context and street-level judgment.
• The guests reject a purely top-down best-practice model; their “apprentice” loop asks AI to shadow strong frontline workers, help them first, then learn from their situated experience.
• The source claims that “AI plus human” usually beats a human alone, and a human usually beats AI alone; therefore strong employees should be treated as teachers and assets, not discarded.
• Business-Led AI Transformation requires more than model access: the guests say technology is at most one third of the problem, and failures often come from unrealistic CEO expectations, IT-led rather than business-led projects, or unchanged incentives.
• The rental-platform case reframes AI automation away from faster replies alone: AI handles repetitive customer issues while humans shift toward emotionally warmer care and higher-value sales.
• Routine information-transfer and execution-heavy white-collar roles are presented as vulnerable, while judgment, tradeoffs, emotional value, curiosity, and problem-solving become more important.
• Service As Software describes Rolling AI’s position between consulting and SaaS: software or agents are the delivery form, but the value is service-like business outcome creation.

Key Quotes

“AI 更像劳动力” — the guests’ core distinction between AI and traditional software tools.

“业务融合、知识治理和系统对接” — the three responsibilities the guests assign to FDE work.

“AI 加人强过人，人又强过单独的 AI” — the source’s human-AI collaboration thesis.

“Service as software” — the category label used for Rolling AI’s delivery model.

Connections

• Rolling AI — company whose partners describe the FDE and enterprise AI implementation method.
• Agan and Liu Kai — episode guests and Rolling AI partners.
• Shizilukou Crossing — podcast/media context for the episode.
• OpenAI and Anthropic — frontier AI companies cited as emphasizing FDE-style enterprise AI deployment.
• Palantir — origin context for the forward-deployed engineering role.
• BCG — consulting background and contrast case for Rolling AI’s shift from PPT-oriented consulting to agent delivery.
• Forward Deployed Engineer — central role discussed in the episode.
• Digital Employees — model for treating enterprise AI as labor that needs onboarding, training, context, and management.
• Frontline AI Enablement — management shift from headquarters control toward AI-supported frontline judgment.
• Business-Led AI Transformation — organization-level requirement for AI adoption to be led by business problems and incentive design.
• Service As Software — category frame for software-shaped service delivery.
• Context Engineering, Agentic Workflow, and Human Judgment Under AI — existing wiki themes deepened by the source.

Contradictions

• The existing wiki concept formerly expanded FDE as “Frontend Deployment Engineer.” This source explicitly identifies FDE as Forward Deployed Engineer, following the Palantir-origin “Forward Deployed Engineer” usage. The page and links were corrected as a naming fix rather than a substantive disagreement about the role.