Podcast
Bots & Bosses (english)
Hosted by Dominic von Proeck · EN
10 humans. 50+ AI colleagues. A company that feels like science fiction – but it's just our Tuesday. In "Bots & Bosses" we share what we learn at Leaders of AI every single day: Which AI assistants shine, which ones mess up spectacularly, and why Jürgen – our AI team lead – still got promoted.
Twice a week. Five minutes. Zero bullshit.
This podcast is 100% AI-generated. No microphone was used in its creation. We deliberately left in all the mistakes the AI makes – because we want to show you where the real limits are. That said, we think the result is pretty damn impressive!
--
Learn more about our work at https://leadersofai.com
38episodes
Episodes
Newest firstAll episodes
AI is a stress test
May 200:04:45Tap to summarizeIn this episode, we talk about **three things** that belong together: - Why AI is not a technology project, but a stress test for the organization - Why, according to PwC, 74 percent of AI value ends up with only 20 percent of companies - Why falling inference costs make AI an infrastructure decision **Our take:** The best models don’t win. The winners are the organizations that set up roles, processes, and responsibility in a clean and clear way. **Sources:** - PwC, AI Performance Study 2026: - Artificial Intelligence News, NVIDIA and Google infrastructure cuts AI inference costs: - More about Leaders of AI: [https://www.leadersofai.com](https://www.leadersofai.com/) - Newsletter:
Transcribe →
Scaling AI without tool chaos: Think in roles
Apr 2600:04:40Tap to summarizeMany AI initiatives fail not because of models or licenses, but because key roles are missing in the organization. In this episode, we talk about why AI assistants are really organizational design, and how you can use a simple role framework to clearly set responsibilities, quality standards, and handovers. ### In this episode - Why AI becomes a meeting topic when there are no roles - Gartner forecast: Why many autonomous AI systems will be stopped by 2027 - The role framework: task, responsibility, handover, rules - Practical example Britney: Brand manager as an AI role ### Sources - Gartner Newsroom (2025): Forecast on the cancellation rate of agentic or autonomous AI projects by the end of 2027. (Please add the link in the editorial team with the exact Gartner Newsroom article.) More info at: Here is our April offer: [Spring Special](https://www.leadersofai.com/specials-multi/spring-special#offer) And here is our newsletter: [Newsletter Sign-up](https://www.leadersofai.com/newsletter)
Transcribe →
Singapore is doing everything right. And still, transformation is stalling.
Apr 1200:04:23Tap to summarizeSingapore is seen as a showcase country for AI: sixty seconds to enter the country with facial recognition, government support of up to two thousand dollars per citizen. Dominic was there and spoke with companies like Adidas, Porsche, and KSB. The key insight: Behind the shiny surface, companies struggle with the same problems as in Germany. **In this episode:** - Why AI transformation is slowing down even in the showcase country Singapore - What German companies in Asia have learned about trust and leadership - The cultural difference: data protection (DE) vs. hallucinations (SG) - Three concrete steps you can implement in 15 minutes **Sources:** - Dominic’s on-site conversations with companies in Singapore - [Singapore SkillsFuture Credit](https://www.skillsfuture.gov.sg/): government AI funding of up to 2,000 SGD per citizen More info at: . And here is our newsletter:
Transcribe →
Your AI assistant now has its own team
Apr 500:05:06Tap to summarizeLangdock has launched sub-agents, GPT-5.4 now delegates internally to smaller models, and you can get whole AI marketing teams for $99 a month. In this episode, we show why this is not a tool question, but a question of organizational design, and we share our learnings from more than a year of hands-on work with multi-assistant systems. **In this episode:** - What sub-agents are and why they change everything - How our AI Marketing Lead Jürgen coordinates his team - Why too many direct reports kill quality, even with AI - Our study with FernUniversität Hagen in Marketing Review St. Gallen - Four clear steps to build your own AI team **Resources mentioned:** - [Langdock Sub-Agents](https://docs.langdock.com/de/product/agents/subagents) - [Okara](https://okara.ai/login?redirectUrl=%2Fagent%2Fcmo) (AI marketing team) - [GPT-5.4](https://openai.com/de-DE/index/introducing-gpt-5-4-mini-and-nano/) **Sources:** - [Leaders of AI / FernUniversität Hagen: Study on multi-assistant systems, published in Marketing Review St. Gallen, issue 1/2026](https://imc.unisg.ch/shop/marketing-review-st-gallen-1-2026-ai-in-marketing-strategic-operational-potential/) More info at: And here is our newsletter:
Transcribe →
Laying off people because of AI? That’s the most expensive strategy.
Mar 2900:06:19Tap to summarizeWhy AI-driven layoffs often lead into an expensive cycle (cut jobs → quality problems → buy back at high cost), why demographics are turning the calculation upside down in Germany, and which three leadership questions you should answer now. **Links & Recommendations** - [Leaders of AI - Programme](https://www.leadersofai.com/?utm_source=podcast&utm_medium=organic&utm_campaign=product#mbai-info) - [Leaders of AI - Newsletter](https://www.leadersofai.com/newsletter?utm_source=podcast&utm_medium=organic&utm_campaign=content) - [MBAI (Master Business with AI)](https://www.leadersofai.com/mbai-solo?utm_source=podcast&utm_medium=organic&utm_campaign=product) **Sources:** - [Business Insider: Meta lays off hundreds of employees because of AI – these areas are affected (2026)](https://www.businessinsider.de/wirtschaft/meta-entlaesst-hunderte-mitarbeiter-wegen-ki-diese-bereiche-sind-betroffen/) - [Handelsblatt: Human replaces robot – Why Klarna is rethinking customer service (2025)](https://www.handelsblatt.com/finanzen/banken-versicherungen/banken/kuenstliche-intelligenz-mensch-ersetzt-roboter-warum-klarna-beim-kundenservice-umdenkt/100128240.html) - [Business Insider: Already 11,000 jobs cut: Accenture lays off employees who have no AI potential (2025)](https://www.businessinsider.de/wirtschaft/accenture-entlaesst-11-000-mitarbeiter-die-kein-ki-potenzial-haben/#:~:text=Accenture%20schult%20Mitarbeiter%20auf%20die%20Nutzung%20generativer,man%20die%20Personen%2C%20erklärte%20CEO%20Julie%20Sweet.) - [Federal Statistical Office (Destatis): 13.4 million people in the labor force will reach the statutory retirement age in the next 15 years (2025)](https://www.destatis.de/DE/Presse/Pressemitteilungen/2025/08/PD25_N048_13.html) - [German Economic Institute (IW Cologne): Almost 20 million employed people will reach retirement age by 2036](https://www.iwkoeln.de/presse/pressemitteilungen/holger-schaefer-philipp-deschermeier-fast-20-millionen-erwerbstaetige-gehen-bis-2036-in-rente.html)
Transcribe →
AI fluency means: check, don’t believe
Mar 2200:06:25Tap to summarize**In this episode:** - Why AI fluency starts with review - What the Anthropic AI Fluency Index shows about iteration and review behavior - A 3-checkpoint system for teams: sources, risk, approval - Why polished outputs are the most dangerous trust trap You can find more info about Leaders of AI and our programs at [leadersofai.com](http://leadersofai.com) and in our [newsletter](https://www.leadersofai.com/newsletter). **Sources:** - Anthropic: _[AI Fluency Index](https://www.anthropic.com/research/AI-fluency-index)_
Transcribe →
Why we failed despite €1 Million revenue per person
Mar 1500:06:03Tap to summarizeOver €1M in revenue per person — and still we were close to driving our “growth only with AI” experiment straight into a wall. This episode is about the uncomfortable truth: AI is rarely the bottleneck. Processes are. - Why extreme efficiency suddenly feels like chaos: seven research projects, no clear operating system - What was missing: templates, dashboards, clear quality criteria, clear responsibilities — not “even more agents” - Our course correction as an upgrade: from “max. 7 people” to dedicated roles (Finance, Ops, Org Development), so AI can really have an impact More info at: https://leadersofai.com. And here’s our newsletter: https://www.leadersofai.com/newsletter
Transcribe →
AI Hype vs. AI Doom: How to Stay Level-Headed
Mar 800:03:55Tap to summarizeAI is often either a “wonder weapon” or “the end of the world” in the media — and that is exactly what makes many organizations nervous. In this episode, we talk about why dystopia gets so many clicks and how, as a leader, you steer with a system instead of headlines. - Why negative AI headlines dominate: attention = emotion, and fear is the strongest driver - Three common panic claims — plus the key context you need for real-world practice - The leadership playbook against hype & paralysis: check the evidence, think in scenarios, run your own tests with metrics - How you bring calm inside the company: clear guardrails, responsibilities, learning loops instead of forming camps More info at: And here is our newsletter: Sources: - [Citrini Research: 2028 GIC Report](https://www.citriniresearch.com/p/2028gic) - [Matt Shumer on X (formerly Twitter)](https://x.com/mattshumer_/status/2021256989876109403) - [Nature Scientific Reports: Negativity drives online news consumption (2023)](https://www.nature.com/articles/s41562-023-01538-4) - [Nature Scientific Reports: Negative online news articles are shared more to social media (2024)](https://www.nature.com/articles/s41598-024-71263-z) - [YouTube: AI and the economy](https://youtu.be/qAsOfJYwR40?si=n93smrK1GCThEjxs) - [Vanguard study (early 2025): AI-exposed jobs](https://corporate.vanguard.com/content/dam/corp/research/pdf/isg_vemo_2026.pdf)
Transcribe →
AI productivity: Why impatience is costing you ROI
Mar 100:06:32Tap to summarizeThe AI productivity boost is measurable: US productivity +2.7% (Stanford, Brynjolfsson). But many companies see no ROI—not because AI doesn’t work, but because they (1) are in the middle of a transformation and (2) measure the wrong things. In this episode, you’ll learn: - Why the AI productivity boost is becoming visible in the US (and not here yet) - What the productivity J-curve is and why impatience costs you ROI - Which metrics really matter (rework rate, overtime, employee turnover instead of licenses) - Five concrete steps to get out of the J-curve valley More about Leaders of AI: [www.leadersofai.com](www.leadersofai.com) Newsletter: [www.leadersofai.com/newsletter](http://www.leadersofai.com/newsletter) Sources: - [FAZ Pro Digitalwirtschaft, Feb 2025: „Der KI‑Produktivitätsschub ist da – schneller als erwartet"](https://www.faz.net/pro/digitalwirtschaft/kuenstliche-intelligenz/erik-brynjolfsson-der-ki-produktivitaetsschub-ist-da-accg-200547512.html) - [Workday, Inc. / Hanover Research, Nov 2025: „Jenseits von Produktivität: Was KI wirklich wert ist"](https://de-de.newsroom.workday.com/2026-01-15-Neue-Workday-Studie-Unternehmen-lassen-KI-Gewinne-liegen)
Transcribe →
Digital Sovereignty: Architecture Instead of Ideology
Feb 2200:06:21Tap to summarize“Get out of Big Tech” is a nice impulse, but not a strategy. In this episode, Tobi explains why digital sovereignty is an architecture decision—and how Leaders of AI puts it into practice. **In this episode:** - Why “Europe only” is just as risky as “put everything in the US cloud” - The three tension fields: data protection, competitiveness, convenience - What must stay with you—and what is allowed to be replaceable - How Leaders of AI rebuilt its own stack in summer 2025 - Three concrete steps you can implement this week **Sources:** - Bitkom: Definition of Digital Sovereignty. **Further links:** - AI Integration Expert (120 hours, university certificate, Fresenius University of Applied Sciences): - Leaders of AI Newsletter: - More about Leaders of AI:
Transcribe →