Brian Madden

Everyone’s worried about the wrong AI security risk

While many fear that AI will leak corporate secrets into its training data, the true risk lies in "execution" rather than "exfiltration," as AI agents can perform unintended or har…

The invisible 80%—what corporate-led AI transformations can’t see

Corporate-led AI initiatives often fail because they only address the visible outputs of work, overlooking the "80%" of essential, unwritten expertise held by employees. Real trans…

Workers don’t want to build automations. They want to delegate.

While AI automation studios offer impressive capabilities, the author argues they are merely a stepping stone because most knowledge workers prefer delegating tasks rather than des…

AI will be THE interface to knowledge work. Here’s how we’ll get there.

The article argues that AI is evolving from a simple assistant into the primary interface for knowledge workers, acting as a centralized gateway to existing enterprise applications…

IT admits workers control AI. Workers admit they use it to leave at 5.

Survey results from Microsoft Ignite reveal that AI adoption is being driven by workers seeking personal efficiency rather than through top-down corporate mandates. Because AI is b…

Everyone wants to provide your AI. Nobody wants to help you manage it.

As knowledge workers increasingly rely on a fragmented array of AI tools from various vendors, attempting to standardize on a single AI platform has become an impossible task. Inst…

Will AI need to operate your legacy desktop apps or is direct file manipulation enough?

After observing Claude edit an Excel file directly without using the Excel application, the author posits that we are entering a "post-app" era where AI bypasses software interface…

AI just created 10,000 accidental citizen developers in your company. Welcome to the post-application era!

The rise of AI is enabling an explosion of custom, ephemeral "micro-apps" created by non-technical employees, making traditional IT methods of managing individual applications obso…

The bitter lesson of workplace AI: Stop engineering, start enabling

Applying "the bitter lesson" to the workplace, the author argues that simple, worker-driven AI adoption consistently outperforms complex, IT-engineered enterprise solutions. Rather…

If AI is normal technology, boring infrastructure is your best strategy

The article argues that AI will follow the incremental adoption pattern of "normal technologies" like electricity or the internet, taking decades rather than months to fully transf…

Worker-led AI isn’t shadow IT. It’s shadow strategy.

Instead of dismissing worker-led AI use as "shadow IT," businesses should view it as "capability adoption at the edge" where employees are actively prototyping future workflows. To…

Everyone’s wrong about why enterprise AI is failing

While large enterprise Gen AI pilots often fail to deliver measurable ROI, the author argues that a significant transformation is already occurring as individual employees use cons…