# Podcast > Prompt & Proper: the German-language podcast about artificial intelligence in companies. Every two weeks with Romano Roth and Steffen Ochsenreither. Source: https://romanoroth.com/en/podcast/ **Prompt & Proper** is my podcast about artificial intelligence in companies, hosted together with Steffen Ochsenreither from Swiss Post. Every two weeks, in German, hands-on and without the marketing speak. We take one real problem per episode and work through what actually moves the needle: use cases, organization, feedback loops, and the hard decisions in between. Listen and subscribe on [Spotify](https://open.spotify.com/show/033SMEcWCZWibLOoMG0qMB), [Apple Podcasts](https://podcasts.apple.com/us/podcast/prompt-proper/id6792786850) or [Amazon Music](https://music.amazon.com/podcasts/b8a56367-a595-4865-8e73-97c706fa4743). All episodes and links also live at [promptandproper.ai](https://promptandproper.ai/). ## In this section (10) - [Who Is Afraid of AI? Agents, Permissions and the Boring Topic of Governance](https://romanoroth.com/en/podcast/e10-wer-hat-angst-vor-ki/), 2026-10-14: In episode 10 of Prompt & Proper, Steffen Ochsenreither and I talk about the AI labs warning against their own models. The real risk sits elsewhere: in agents that are allowed to do more than anyone keeps track of. We cover containment, agent inventories, human in the loop under load, and whether you can hand responsibility to an agent. - [AI FinOps: Why Falling Token Prices Make Costs Explode](https://romanoroth.com/en/podcast/e09-ai-finops/), 2026-09-29: In episode 9 of Prompt & Proper, Steffen Ochsenreither and I talk about AI FinOps. The price per token keeps falling, yet the bill keeps growing, because agentic workflows consume far more than the price drop can absorb. We cover the right model for the right task, why the model choice has to be automated, why cost alone is no KPI, and why companies will soon buy tokens the way they buy electricity. - [Dark Software Factory: When Nobody Reads the Code Anymore](https://romanoroth.com/en/podcast/e08-dark-software-factory/), 2026-09-16: In episode 8 of Prompt & Proper, Steffen Ochsenreither and I take the dark factory concept from manufacturing and apply it to software development. Five stages of AI coding, the point where the factory analogy breaks, my own experiment running a full V-model, and the shift that actually matters: the specification becomes more important than the code. - [How to AI-Transform a Company: Vision and Leadership First, Technology Last](https://romanoroth.com/en/podcast/e07-wie-man-ein-unternehmen-ki-transformiert/), 2026-09-02: In episode 7 of Prompt & Proper, Steffen Ochsenreither and I take on the transformation itself. The most common question is still the tooling question, yet no tool transforms an organization. The usual maturity frameworks start with technology, which puts them upside down. Plus my AI maturity matrix, the transformation framework we run in practice, and the real lever: transformative leaders who go ahead as beacons. - [AI-First or AI-Enabled? Why the Name of Your AI Strategy Comes Second](https://romanoroth.com/en/podcast/e06-ai-first-oder-ai-enabled/), 2026-08-26: In episode 6 of Prompt & Proper, Steffen Ochsenreither and I walk through the terminology zoo: AI-First, AI-Native, AI-Enabled, AI-Ready, AI-Shaped, AI-Augmented. Why AI-First wins at C-level and still sends the wrong signal, why AI-Augmented fits software development best, and why companies want the label in the first place. - [The PoC Trap: Why Over 90 Percent of AI Pilots Never Reach Production](https://romanoroth.com/en/podcast/e05-die-poc-falle/), 2026-08-19: In episode 5 of Prompt & Proper, Steffen Ochsenreither and I dig into the PoC trap. Classically around 70 percent of all proofs of concept got stuck before production, with AI the rate sits beyond 90 percent. Technology is the smallest of the problems. What is missing is the strategic goal, the data quality, the production foundation and a culture that carries experiments into daily work. - [Is AI Killing 40 Percent of All Jobs? Why the Headline Is a Marketing Pitch](https://romanoroth.com/en/podcast/e04-killt-ki-40-prozent-aller-jobs/), 2026-08-12: In episode 4 of Prompt & Proper, Steffen Ochsenreither and I take the 40 percent headline apart. Yale, Brookings and the KOF find no AI-driven mass job loss, and most of the layoffs started six months before ChatGPT shipped. The structural shift behind it matters more: broader roles, teams of 3 plus minus 2, and new bottlenecks at review and at the very start of the chain. - [AI and Jobs Beyond Developers: A Broken Process Stays a Broken Process](https://romanoroth.com/en/podcast/e03-jobs-jenseits-der-entwickler/), 2026-08-05: In episode 3 of Prompt & Proper, Steffen Ochsenreither and I look at every job outside engineering. In management, legal and office work, AI could cover 60 to 90 percent of the tasks, yet actual usage stays below 30. The reason is rarely technology. It is processes nobody owns. - [Agentic Coding and Token Costs: Why the 10x Claim Is a Myth](https://romanoroth.com/en/podcast/e02-agentic-coding-und-token-kosten/), 2026-07-29: In episode 2 of Prompt & Proper, Steffen Ochsenreither and I do the math on what agentic coding actually delivers: 20 to 30 percent across the whole value chain, not 10x. Plus the economics behind it, from 50 dollars of tokens per developer hour to what happens when the bubble bursts. - [The Giant Use-Case Fallacy: Why AI Projects Fail Before They Start](https://romanoroth.com/en/podcast/e01-der-riesen-use-case-irrtum/), 2026-07-20: In the first episode of Prompt & Proper I talk with Steffen Ochsenreither about why the hunt for one giant use case kills so many AI projects, how to start small with a real purpose instead, and why the biggest bottleneck is not the technology but the organization.