Paul Slater helps professionals, leaders, and organizations navigate the human side of AI transformation. He is the author of The AI-Ready Human: Your 90-Day Program to Stay Relevant as Technology Transforms Work, host of the Humanity Working podcast, and founder of Paul Slater Advisory.

Paul spent nearly two decades at Microsoft, leading global digital transformation initiatives — including defining strategy for the company's Life Sciences business — and authored more than 20 books and courses for senior technologists. He has contributed to AI think tanks at Harvard, Duke, and Arizona State University, and briefed Fortune 500 executives and national governments worldwide. He is currently an analyst-relations thought leader at Adobe, where the frame problem behind this episode — "how do I stay on top of what 70+ industry analysts are saying?" — is the one he ended up vibe coding a solution for.

The 12-hour build. Paul's opening story: needing a hybrid of a CRM, a CMS, and a news feed for his Adobe work — a "CIA-profile-every-analyst" system Salesforce doesn't cover — and building it end-to-end in Claude Code in about twelve hours. His advice: if you're not technical, get a bit technical. If you're technical, get more technical. Then build the thing that fits your work like a glove.

The barrier to coding has collapsed — knowing what to build has not. Paul's clearest reframe: the how-to-code is gone. What remains is the ability to think holistically enough to imagine the system that would solve the problems you have. If you can do that, everything else isn't that hard. This is the muscle worth building.

The standardization paradox in regulated industries. Matt's frame: life sciences needs standardized foundational systems it doesn't quite have — and yet if those foundations were solid, everyone could accelerate the individualized layer on top. So can AI let us leapfrog around the missing standardization? Sometimes yes; sometimes the messy reality is the whole reason your custom tool is valuable.

Products are an artifact of cost, not need. Paul's epiphany: the explosion of products since 1990 isn't a reflection of need — it's a reflection of how cheap it became to bring a solution to market and wrap it in sales and marketing. Every product then accumulates functionality to justify its existence, until — Matt's line — the product becomes the problem.

What actually needs to be centralized. A relatively small amount of data must live in highly regulated, locked-down systems. Everything else — the messy, unstructured, contextual layer that supports the art of the work rather than the science of it — has been artificially wedged into expensive applications where it never belonged. Paul's clinical-trial example: a huge amount of real-world Phase IV signal is already out in the world; you don't need to lock it down at the point of first inspection, only when you build a hypothesis on it.

Why software companies have to reinvent themselves. With AI writing an increasing share of software (Paul references Anthropic's public numbers on internal AI-written code), the moat under any SaaS resting purely on software gets very thin, very fast. Paul's bet: winners like Veeva don't get displaced — they reinvent themselves from "product for these roles" into a role-enablement layer that uses their unmatched understanding of how the work is done to build self-forming, naturalistic support for each user.

The Star Trek IV point. In the film, Scotty tries to talk to a 1980s Mac and Bones hands him the mouse like it's a microphone. The joke: the future is one where you talk to the computer and it responds. That future has arrived — and it means the same tool should present four different faces to four different people doing the same job, adjusting to how each of them actually thinks and works.

The ugly middle ground on pricing. We're paying legacy subscriptions plus token consumption plus occasional old licenses — the worst of every pricing model at once. Paul's read: pick a lane. His preference is everything is a token — like electricity. Leave the lights on all night, pay more. Cleaner accountability, cleaner incentives.

Local models vs. the data center. The recent Dell / Microsoft / NVIDIA announcements around local AI processing hint at a different future — where most inference runs on the user's expensive workstation and only escapes to the cloud for the heavy lift. That may be the shape of the cost curve that finally makes token economics work.

Nothing about work is fit for purpose. Paul's summary line: technology infrastructure, organizational infrastructure, information infrastructure — none of it is where the puck is. On alternate mornings that terrifies him ("we're all going to be out of a job") and thrills him ("that is a lot of work for a lot of people"). Either way, it's the biggest reset since the PC.

The AI-Ready Human and the third framing. Paul's book is designed to meet people wherever they're at — from wall push-ups to elite AI users — and centered on the evergreen part: the human. In work with creatives at Adobe he's landed on a third way to think about AI, beyond productivity-enhancer and quality-improver: AI as medium. Things that couldn't have been created any other way — the Beatles' Now and Then, the short-form work of surrealist artists — are the flipside of AI slop. Same technology, different intent.

"I built something way more useful than any commercial software at all for my job in twelve hours."

"You can literally vibe code the perfect workflow support system for whatever job you have."

"The product becomes the problem."

"You wind up being held back by the very solutions that you're paying for."

"Almost every aspect of how work is structured and how work is done is not fit for purpose."

"AI can do a poor job of mimicking things humans would do, but it might do a really good job of creating things humans cannot."

Life sciences, Reg Ops, and Quality leaders trying to see around the corner of the SaaS-and-AI shift; enterprise architects and IT strategists rethinking buy-versus-build; software vendors serving regulated industries; and anyone wondering what "AI-ready" means for their own job in the next twelve months, not the next five years.

Guest & Work

Tools & Platforms Discussed

Concepts & Cultural References

Transcript provided by Otter.ai.

Operations Utopia - Where Regops, Innovation, Technology, and Execution Meet.

Disclaimer: This podcast reflects only the opinion of the podcaster and guests and does not reflect those of their organizations, system vendors, or service provider

Original show theme "Little Sammy" by Matt Neal