AI Pulse by Inblix

Philips' 70,000-employee AI playbook skips the hype

OpenAI Blog · Jul 12, 2026 · 3 min read · Read original article →

Curated by the Inblix editorial team


Featured image for article: Philips' 70,000-employee AI playbook skips the hype

Philips, a 134-year-old healthcare tech giant, isn’t treating AI literacy as a one-off workshop. It’s a full-blown cultural rewiring, and Head of Data Science & AI Engineering Patrick Mans detailed exactly how they’re pulling it off across a massive, 70,000-person workforce. The strategy rests on a disarmingly simple framework: Toy → Tool → Transformation. “You start playing with it, then you start working with it—and from there, you start innovating with it,” Mans explained. The initial spark wasn’t some top-down mandate from a corner office. Philips discovered employees were already using OpenAI tools privately, so the company decided to channel that organic curiosity into legitimate business workflows rather than fight it.

The rollout moved on two tracks to build unstoppable momentum. Top leadership got hands-on training first, not so they could recite AI buzzwords, but to actively model usage and signal that this wasn’t a passing fad. Simultaneously, a company-wide challenge invited anyone to propose concrete use cases, creating a groundswell of demand. Providing access to ChatGPT Enterprise poured fuel on that fire. But for a company operating under crushing regulatory scrutiny in patient monitoring and diagnostics, unbridled experimentation isn’t an option. The trust-building sequence was deliberate: start with low-risk internal admin tasks, let people build muscle and confidence in controlled sandboxes, and bake formal responsible AI principles—transparency, fairness, human oversight—into the process before anything gets near a patient-impacting workflow.

The primary target right now is soul-crushing administrative burden. Mans shared a sharp, infuriating example that crystallizes the mission: a clinician who spent 15 minutes saving a life and then had to spend the next 15 minutes typing up the paperwork. In that wasted quarter-hour, another patient could have been treated. That’s the problem Philips is using AI to solve, and it’s a far cry from flashy, futuristic robotics. It’s about reclaiming stolen time.

Philips is now pushing past individual productivity hacks into workflow-level automation and agent-supported processes, all under a clear AI policy. The endgame is profoundly human, not technological. Mans put it bluntly: “We want to deliver better care for more people. AI is one of the most powerful tools we have to do that.” For a company that’s been around since the 19th century, this isn’t about chasing the new new thing. It’s a pragmatic, trust-first blueprint for making a massive, heavily regulated organization genuinely AI-fluent, and it works precisely because it starts with people admitting they’re just playing with the tools.

💡 Key Takeaways

  1. Philips is moving its 70,000 employees along a "Toy → Tool → Transformation" curve, deliberately channeling existing personal use of AI into sanctioned business innovation.
  2. Executive training was hands-on and designed to model behavior, not just rubber-stamp an initiative, creating a permission structure for the rest of the company.
  3. The company is deliberately restricting AI to low-risk internal workflows first to build institutional trust before letting it impact regulated, patient-facing clinical environments.
  4. Philips' immediate strategic focus is narrow and pragmatic: using AI to eliminate administrative documentation that directly steals time from clinicians interacting with patients.
  5. The next phase moves from individual productivity gains to automating entire workflows and agent-supported processes, all governed by a clear, pre-existing responsible AI framework.

Keep reading: See related articles below for more coverage on this topic.

Get smarter about AI

The sharpest AI news, curated daily. Delivered free to your inbox.

Learn more

Glossary terms

← Back to all articles