WSJ Tech News Briefing

Episode Title: Microsoft Sees Healthcare as Path to Independence From OpenAI
Date: October 10, 2025
Host: Julie Chang
Featured Guests: Sebastian Herrera (WSJ Reporter); Nicholas Elliott (Dow Jones); Chi Kin Lam (DBS Bank); Kelly Geer (FedEx)

Episode Theme & Overview

This episode explores how AI is transforming risk management in large organizations and delves into Microsoft's strategic move to expand its own AI capabilities—particularly in healthcare—in pursuit of greater independence from its current reliance on OpenAI’s technology. The episode offers perspectives from company executives and reporters, focusing on real-world AI applications, industry risks, and the shifting dynamics of AI partnerships and competition.

Key Discussion Points & Insights

1. AI for Risk Management in Global Companies

(00:19–04:56)

• Kelly Geer (FedEx):
  • Transcription in Investigations: AI is crucial for transcribing multilingual interviews across the 100+ countries in FedEx's EMEA region, significantly easing the reporting process for investigators not fluent in English.
    • "Our teams have saved so much time in going back and forth with the investigator, like, is this really how you want to say this?" (01:56)
  • AI Chatbots for Compliance: Legal teams (non-coders) developed an internal chatbot answering compliance policy questions via a decision tree and SharePoint. Future improvements aim to leverage more advanced AI features.
  • Customer-Facing AI Tool: A FedEx-developed bot assists customers with harmonized tariff schedules—vital for swift customs clearance—by interpreting described goods into accurate tariff codes and linking to US government resources.
    • "We're really proud of this tool...whatever item you're shipping into the US...this chatbot within this tool will help you describe the item more accurately..." (02:58)
• Chi Kin Lam (DBS Bank):
  • AI Agents & Governance: Stresses the need for governance policies that differentiate high-risk from low-risk AI use cases, citing automation of lunch ordering vs filing suspicious transaction reports.
    • "Filing a suspicious transaction report, you are effectively calling somebody a criminal. You really don't want the AI making that decision." (03:56)
  • Cost Control: Warns against overusing expensive large language model (LLM) calls when a simpler API or process automation may suffice.
    • "Don't always call the LLM because you pay for that...Your bill is going to be pretty high." (04:22)
  • Prompt Engineering Caution: Overengineering prompts increases risk and cost; sometimes, simple solutions are preferable.

2. Microsoft’s Strategic Push Toward AI Independence in Healthcare

(05:46–10:09)

The Healthcare Opportunity

• Julie Chang: Microsoft is collaborating with Harvard Medical School to improve Copilot's reliability in providing medical information, aiming to make user interactions more like consultations with healthcare professionals and to stand out in the consumer AI marketplace.

Differentiating Copilot vs. ChatGPT

• Sebastian Herrera (WSJ):
  • Microsoft’s goal: Deliver the most accurate, helpful medical advice compared to current chatbots, addressing high-profile incidents where ChatGPT gave poor or even harmful medical advice.
    • "Microsoft wants to be the opposite of that, and they're hoping to make Copilot something that people go and use, and if they can figure out the medical part of it, it could be really successful for them." (07:04)
  • A Stanford study revealed ChatGPT gave bad medical advice in approx. 20% of test cases, highlighting the need for more reliable AI.

Microsoft–OpenAI Relationship Status

• Sebastian Herrera:
  • Microsoft maintains that its partnership with OpenAI remains strong and contractual, but aims to reduce technical dependency by replacing OpenAI models in Copilot with their own, especially for healthcare-related workloads.
    • "They want to have their own models be used within their chatbot. Their goal is to have Microsoft models replace workloads for Copilot..." (07:54)

Internal AI Lab & Talent Acquisition

• Sebastian Herrera:
  • Microsoft’s internal AI lab, led by ex-Google executive Mustafa Suleiman, has seen rapid hiring, including talent from Google DeepMind and experienced clinicians for the new healthcare initiative.
    • "A lot of people that have joined...have been in the past six to nine months...They've recruited people from startups that are really experienced. They have clinicians on staff with some of this healthcare initiative." (08:38)

Challenges Ahead

• Sebastian Herrera:
  • Microsoft has a long history of failed or struggling attempts in health tech, and healthcare remains a highly regulated, technically complex field.
  • Copilot currently trails OpenAI and Anthropic in both usage and innovation.
  • Microsoft’s brand is strong in enterprise tech but lacks major consumer traction—making its Copilot-healthcare move a high-risk, high-reward bet.
    • "By no means is it promised for Microsoft to succeed here in any way." (09:56)

Notable Quotes & Memorable Moments

• Kelly Geer (FedEx) on AI Improving Internal Workflows:
  "Lawyers, they are not coders. And they got together and developed a chatbot that can answer questions related to our compliance policies." (02:09)

• Chi Kin Lam (DBS Bank) on AI Risk Management:
  "You really don't want the AI making that decision." (03:56)

• Sebastian Herrera (WSJ) on Microsoft’s Strategy:
  "Microsoft wants to have the most accurate and helpful medical advice, and that points people in the right direction." (06:45)

  "Not only is OpenAI their partner, but it's also their competitor." (09:02)

Timestamps for Key Segments

• 00:19–04:56: AI risk management at FedEx & DBS Bank—practical examples and governance
• 05:46–10:09: Microsoft’s Copilot strategy, partnership with Harvard Medical, internal AI lab expansion, and competitive/technical challenges

Summary Table

| Segment | Speaker(s) | Highlights | |---------------------------------|--------------------|---------------------------------------------------------------------------| | FedEx/DBS Bank AI Use | Kelly Geer, Chi Kin Lam | Multilingual transcription, compliance chatbots, agent risk policies, cost-control tips | | Microsoft & Healthcare AI | Julie Chang, Sebastian Herrera | Copilot’s healthcare push, OpenAI dependency, in-house talent acquisition, major hurdles |

This episode delivers practical insights into AI’s real-world applications in legal/compliance and gives a behind-the-scenes look at Microsoft’s high-stakes bid to reframe its AI strategy via the healthcare vertical. The themes of technological innovation, risk management, and tech industry competition are woven throughout, offering both tactical and strategic perspectives for listeners interested in the future of enterprise AI.