Paul Lewis is the Chief Technology Officer at Pythian, where he helps organizations use cloud, AI, and data together to drive business transformation. With 17 years as a CTO in banking and a decade as global CTO for Hitachi Vantara, Paul brings a unique perspective on enterprise technology challenges. From bullet trains to nuclear power plants to everyday business applications, he’s seen what works and what doesn’t when implementing AI at scale.
About the Episode
The AI hype is real, but so is the gap between what’s possible and what’s practical for most businesses. Paul cuts through the noise with intellectual honesty about where AI actually stands today. He breaks down the two different “AI races”—the edge cases that tech companies showcase versus the enterprise reality of budgets, regulations, and real teams.
We discuss Paul’s framework for evaluating AI readiness and the critical difference between prompting and prompt engineering. Paul shares the strategic insight leaders need to make smart AI investments without getting burned by the hype.
Key Moments
Career Evolution and Digital Transformation
- From mediocre developer to technology leader [2:20]
- Moving from project management to strategic thinking [8:30]
- How the CTO role have evolved in the last decade [9:30]
AI Reality Check and Implementation Framework
- The two different AI races: tech companies vs. enterprise reality [12:00]
- Why intellectual honesty is crucial in AI education [13:00]
- Five-part AI readiness framework for businesses [14:00]
Data Governance and Enterprise Concerns
- The black box problem with AI lineage and accountability [25:20]
- Why data quality remains the core challenge [28:40]
- Personal vs. enterprise AI usage recommendations [30:30]
Practical AI Applications and Use Cases
- Breaking down 400 ideas into actionable categories [16:50]
- People productivity vs. process productivity applications [21:30]
- The difference between prompting and prompt engineering [38:00]
Top Insights
- There Are Two AI Races Happening – Tech companies showcase edge cases without real constraints, while enterprises face real budgets, teams, and regulatory boundaries. Understanding this gap is crucial for realistic AI planning.
- Start With Your Data House in Order – The same data quality problems that plagued BI 15 years ago and machine learning 5 years ago still exist today. Fix data governance before implementing AI solutions.
- Most “AI” Ideas Are Actually Analytics – When organizations brainstorm AI use cases, 350 out of 400 ideas typically turn out to be traditional business intelligence needs, not true AI applications.
- Prompt Engineering Is Software Development – The difference between basic prompting and prompt engineering is like the difference between asking questions and writing code—it requires structured thinking and constraints.
- AI Accuracy Isn’t Enterprise-Ready Yet – Free AI tools operating at 67% accuracy aren’t suitable for production environments where the answer always has to be correct due to reputational and legal risks.
Notable Quotes
- “It’s very easy to read the news and say, I’m incredibly behind, when the reality is there are two races—the technology companies trying to get to edge use cases versus enterprises saying, what are we actually ready to implement?”
- “AI is the microwave of language. I wouldn’t use a microwave to build a gourmet meal, but I would use it to defrost chicken.”
- “Would you ever put anything in production that’s 67% good? No. But that’s what we’re dealing with in most AI implementations today.”
- “You still can trick a GPT. You can convince it that it’s wrong by providing your own opinion, and that’s not helping you—that’s making your knowledge worse.”
Resources
Industry Experts
- Paul Lewis: LinkedIn | Instagram | Facebook | X
- Amanda Natividad on LinkedIn
- Tyler Lane on LinkedIn
Companies and Organizations
Tools and Platforms
Ready to Cut Through the Hype?
Paul’s framework proves that successful AI implementation is about honest assessment, solid data foundations, and strategic thinking. Most organizations aren’t ready for the AI applications they think they want, but they can build real value with the right approach.
At Session Interactive, we help mid-market companies navigate digital transformation with the same honesty Paul brings to AI strategy. We’ll help you separate the signal from the noise and build data-driven strategies that actually work.
Ready to get serious about digital marketing in the AI age? Let’s talk.
