Giovanni Brees: The Zero-Employee Company

Giovanni Brees and Brian Soares

Giovanni Brees is a founder, product strategist, author, and speaker. An entrepreneur for more than 25 years, he has worked with founders, global brands, and Fortune 500 companies including Bain, P&G, and WordPress. Since 2010, his product launch and validation work helped bring 4,500+ products to market and raise over 500 million USD through crowdfunding. He now runs Meet Oscar and CiteEngine with no full-time employees, only AI agents, and is building KentoHQ. Author of The Zero-Employee Company and The Crowdfunding Quick Guide.

Episodes

  1. Aug 13

    QMD: The Logging Habit That Catches AI Hallucinations

    Giovanni Brees explains QMD, the highest-leverage logging habit for catching confident AI hallucinations before an agent acts on them. The three-line pattern makes AI agents expose the evidence behind every non-trivial decision, creating a practical safeguard for AI automation and an AI-first business. Listeners learn how Question, Memory, Decision logging works and why the order matters. The Memory line must identify the exact grounding used, such as an operating brief, world model, text log, queue status, or memory store result, before the Decision line records the action and reason. Giovanni Brees also shows how QMD makes decisions auditable, reduces debugging from reconstruction to direct inspection, and pairs with a mistakes log to turn failed grounding into precise prevention rules. In this episode: How an empty or vague Memory line reveals a guess dressed up as a factA worked example covering staging, human review, outward-action boundaries, and the review queueWhy recording evidence before acting changes agent behavior and lowers hallucination riskHow QMD fits into text-based AI memory today and supports long-term vector memory laterResources mentioned The QMD format (from the book's build kit):Q: the question the agent was answeringM: the context it actually used (files read, what the store returned)D: the decision and whyThe middle line catches hallucination: a decision with an empty or vague M was a guess. Links Website: giovannibrees.comBook: The Zero-Employee CompanyLinkedIn: linkedin.com/in/giovannibreesFrom the book The Zero-Employee Company by Giovanni Brees.

  2. Aug 11

    The Mistakes Log: How AI Agents Improve Themselves

    Giovanni Brees explains how a structured mistakes log gives AI agents a practical self-improvement loop without retraining the model. The episode shows how this layer of AI memory turns corrections into permanent guardrails for reliable AI automation and an AI-first business. Listeners learn the six-field schema: status, task, failure, root cause, fix, and prevention rule. Giovanni Brees details the "twice-equals-a-rule-change" threshold, where a repeated root cause is moved from the log into the agent's operating brief, changing the instructions it reads on boot. The system can run in a shared Markdown or plain text file, supported by immediate logging and a weekly review. In this episode: Why a correction without a specific prevention rule is documentation, not learningHow repeated root causes become permanent constraints in an AI agent's operating briefWhy the model remains frozen while its business-specific wrapper, instructions, and correction history improveHow to test prevention rules, replacing vague guidance like "be careful" with executable rules such as confirming a recipient's time zone before schedulingResources mentioned The mistakes-log entry (from the book's build kit):[ ] ERROR:- Task:- What failed:- Root cause:- Fix applied:- Prevention:Rule: the same root cause appearing twice from the same agent triggers a permanent rule change in that agent's instructions, not another note. Links Website: giovannibrees.comBook: The Zero-Employee CompanyLinkedIn: linkedin.com/in/giovannibreesFrom the book The Zero-Employee Company by Giovanni Brees.

  3. Jul 27

    The First Five Roles I Would Replace With AI Agents

    Giovanni Brees explains which five business roles he would replace with AI agents first, and why the order matters. His delegation ladder offers a practical path from basic AI automation to a zero-employee company where recurring work runs through systems. In this episode, Giovanni shows why founders should think in roles rather than tools when running a business with AI. He maps the first steps toward AI-first operations: correspondence, research, monitoring, content production, and outreach drafting. Each role in an AI-powered business starts with low-risk drafts and clear human approval, while irreversible money decisions, key relationships, legal work, and final taste calls stay human. Giovanni also explains how to define each role, set boundaries, and inspect its output before moving up the ladder. In this episode: Why Giovanni Brees recommends starting with an inbox agent that drafts replies for human approvalHow research and monitoring agents gather information, watch defined thresholds, and remain silent when no action is neededWhy content production and outreach drafting come later, with every important output reviewed before it reaches the outside worldWhich high-stakes decisions should never be delegated to AI agents, plus the first step for building a zero-employee company About Giovanni Brees Giovanni Brees is a founder, product strategist, author, and speaker. An entrepreneur for more than 25 years, he has worked with founders, global brands, and Fortune 500 companies including Bain, Procter & Gamble, and WordPress. Since 2010, his product launch and validation work helped bring 4,500+ products to market and raise over 500 million USD through crowdfunding. He now runs Meet Oscar and CiteEngine with no full-time employees, only AI agents, and is building KentoHQ. He is the author of The Zero-Employee Company and The Crowdfunding Quick Guide. Links Website: ⁠https://giovannibrees.com⁠Book, The Zero-Employee Company: ⁠https://www.amazon.com/dp/B0H72KCZ6Y⁠LinkedIn: ⁠https://www.linkedin.com/in/giovannibrees/⁠ From the book The Zero-Employee Company by Giovanni Brees.

  4. Jul 23

    Can One Founder Run an Entire Company With AI?

    Giovanni Brees answers whether one founder can run an entire company with AI agents. His answer is yes: he runs several through AI automation, with systems handling coordination while human judgment stays at the edge. This episode follows Giovanni through an ordinary day of running a business with AI, from scheduled jobs and structured handoffs to monitoring and approval gates. It explains why an AI-powered business requires the founder to stop acting as the doer or courier and become the architect and inspector of AI-first operations. The result is a zero-employee company where work continues without meetings, standups, or constant intervention. In this episode: Why building the system takes longer at first but removes repeated work laterHow a shared source of truth, clear ownership, logs, and structured handoffs coordinate AI agentsWhy silence is the default and agents contact the founder only when human judgment is requiredHow Giovanni Brees uses approval gates to stay at the boundary of the business instead of becoming its bottleneck About Giovanni Brees Giovanni Brees is a founder, product strategist, author, and speaker. An entrepreneur for more than 25 years, he has worked with founders, global brands, and Fortune 500 companies including Bain, Procter & Gamble, and WordPress. Since 2010, his product launch and validation work helped bring 4,500+ products to market and raise over 500 million USD through crowdfunding. He now runs Meet Oscar and CiteEngine with no full-time employees, only AI agents, and is building KentoHQ. He is the author of The Zero-Employee Company and The Crowdfunding Quick Guide. Links Website: ⁠https://giovannibrees.com⁠Book, The Zero-Employee Company: ⁠https://www.amazon.com/dp/B0H72KCZ6Y⁠LinkedIn: ⁠https://www.linkedin.com/in/giovannibrees/⁠ From the book The Zero-Employee Company by Giovanni Brees.

  5. Jul 21

    What Is a Zero-Employee Company?

    Giovanni Brees defines a zero-employee company as a business where AI agents fill operational roles while one owner designs the systems and judges their output. His approach to AI automation removes employees from the org chart without removing human oversight. This episode explains how Giovanni uses AI-first operations across several businesses, with a couple dozen agents handling work once assigned to teams. It covers why running a business with AI is now possible, how an AI-powered business replaces roles with connected systems, and why consequential work still requires human approval. Giovanni also explains the hidden costs of headcount, the limits of full autonomy, and why the model fits information-based work better than businesses that depend on human presence. In this episode: What "zero employees" means and why zero really means zeroHow AI agents handle defined roles, inputs, outputs, and handoffsWhy the owner remains the architect and inspectorHow Giovanni Brees keeps judgment and verification human About Giovanni Brees Giovanni Brees is a founder, product strategist, author, and speaker. An entrepreneur for more than 25 years, he has worked with founders, global brands, and Fortune 500 companies including Bain, Procter & Gamble, and WordPress. Since 2010, his product launch and validation work helped bring 4,500+ products to market and raise over 500 million USD through crowdfunding. He now runs Meet Oscar and CiteEngine with no full-time employees, only AI agents, and is building KentoHQ. He is the author of The Zero-Employee Company and The Crowdfunding Quick Guide. Links Website: ⁠https://giovannibrees.com⁠Book, The Zero-Employee Company: ⁠https://www.amazon.com/dp/B0H72KCZ6Y⁠LinkedIn: ⁠https://www.linkedin.com/in/giovannibrees/⁠ From the book The Zero-Employee Company by Giovanni Brees.

  6. Jul 21

    Giovanni Brees and The Zero-Employee Company

    Giovanni Brees built a zero-employee company around AI agents and AI automation after more than 25 years as an entrepreneur and founder. Today, Giovanni Brees runs several companies with zero employees, using connected agents for outreach, content, research, and operations while he designs the systems and inspects their work. This episode explains how Giovanni approaches running a business with AI and building an AI-powered business, including which roles should go to AI agents first and which decisions should stay with the founder. It defines AI-first operations in practical terms, examines trust when an agent can be wrong without embarrassment, and sets clear limits for the zero-employee company model. The podcast also covers his work since 2010, which took more than 4,500 products to market and helped founders raise over 500 million USD through crowdfunding. The episode introduces the book by Giovanni Brees and explains how the podcast uses the same AI systems it discusses. In this episode: What a zero-employee company is and how it worksHow AI agents handle outreach, content, research, and operationsWhich roles founders should delegate to AI firstWhy human inspection and judgment still matter About Giovanni Brees Giovanni Brees is a founder, product strategist, author, and speaker. An entrepreneur for more than 25 years, he has worked with founders, global brands, and Fortune 500 companies including Bain, Procter & Gamble, and WordPress. Since 2010, his product launch and validation work helped bring 4,500+ products to market and raise over 500 million USD through crowdfunding. He now runs Meet Oscar and CiteEngine with no full-time employees, only AI agents, and is building KentoHQ. He is the author of The Zero-Employee Company and The Crowdfunding Quick Guide. Links Website: https://giovannibrees.comBook, The Zero-Employee Company: https://www.amazon.com/dp/B0H72KCZ6YLinkedIn: https://www.linkedin.com/in/giovannibrees/ From the book The Zero-Employee Company by Giovanni Brees.

About

Giovanni Brees is a founder, product strategist, author, and speaker. An entrepreneur for more than 25 years, he has worked with founders, global brands, and Fortune 500 companies including Bain, P&G, and WordPress. Since 2010, his product launch and validation work helped bring 4,500+ products to market and raise over 500 million USD through crowdfunding. He now runs Meet Oscar and CiteEngine with no full-time employees, only AI agents, and is building KentoHQ. Author of The Zero-Employee Company and The Crowdfunding Quick Guide.