CoM Products

Clouds On Mars

Clouds on Mars Products is a podcast about what's actually happening at the intersection of technology and business: AI systems, data platforms, automation, and the real decisions behind them. Each episode, host Marcin Kubica (Head of Product at Clouds on Mars) sits down with guests and experts to explore the products and solutions being built for industries like FMCG, retail, healthcare, pharma, and real estate, as well as the broader technical and architectural challenges every modern organization has to face. Let your data make an impact.

  1. Oct 2

    Is Your Semantic Model Ready For An AI Agent?

    An AI agent on an unprepared semantic model gives answers that are confident, clear and wrong. The fix starts with the model, not the prompt. Have a data, analytics, automation or AI challenge you'd like us to dig into? Get in touch, we'd love to hear about it. https://cloudsonmars.com/contact/?utm_source=podcast&utm_medium=description&utm_campaign=ai-ready-models Marcin Kubica sits down with Power BI World Championship finalist Michał Jezik to break down how to prepare semantic models for AI agents, from naming to verified answers. See why duplicate fields and technical column names confuse an agent, how the AI data schema cuts its input context, why descriptions should stay within 200 characters, and how to spend one week making a model agent-ready. This podcast is provided for general information purposes only and does not constitute legal, tax, financial or other professional advice. Nothing herein should be relied upon as a substitute for advice from a qualified professional. Views expressed are the speakers' own. Product features, availability and licensing details may change over time, and any client results shown are illustrative only and are not a guarantee or promise of similar outcomes. © Clouds on Mars. All rights reserved. This content is the property of Clouds on Mars and may not be reproduced or distributed without permission. #PowerBI #MicrosoftFabric #SemanticModel #AIAgents #DataModeling #CloudsOnMars

  2. Oct 2

    Your Fabric Bill Is High Because You Use Too Little

    Have a data, analytics, automation or AI challenge you'd like us to dig into? Get in touch, we'd love to hear about it. ⁠https://cloudsonmars.com/contact/?utm...⁠On most platforms you overpay by using too much. On Fabric, you can overpay by using too little. Marcin Kubica sits down with business unit lead Jakub Tamas to unpack Microsoft Fabric administration at scale, from capacity cost to tenant governance. See how one client was running around 50% more Power BI Pro licenses than it needed, why a 30% average capacity usage is not a result to celebrate, how the 40% reservation discount changes capacity sizing, and what an automation layer adds on top of FUAM.

  3. Aug 25

    More Computing Power Won't Fix a Broken Power BI Model

    One Clouds on Mars client scaled a Power BI model to 300 scenarios and a billion-row table. The report still loaded in a few seconds, exactly as the business asked.Marcin Kubica sits down with BI Competency Center Lead Radosław Górecki to unpack why Power BI reports and semantic models slow down, and how to fix them.Hear how Radosław breaks down citizen development gone wrong, the metric drift that makes two reports disagree on the same number, the tools behind a Power BI refactor assessment, and a model that scaled to 300 scenarios and a billion-row table without losing performance.Have a data, analytics, automation or AI challenge you'd like us to dig into? Get in touch, we'd love to hear about it. https://cloudsonmars.com/contact/

  4. Aug 7

    Why Getting Your Own SAP Data Out Feels So Hard

    One Clouds on Mars client released millions of euros in cash just by connecting their SAP warehouse data to Microsoft Fabric. Marcin Kubica sits down with CTO Estera Kot to break down SAP to Microsoft Fabric integration, from architecture choices to real business use cases. Hear how Estera breaks down SAP Datasphere mirroring versus OData extraction, when the new BDC Connect connector makes sense, and real customer wins from faster financial closes to freeing up warehouse cash.Have a data, analytics, automation or AI challenge you'd like us to dig into? Get in touch, we'd love to hear about it. https://cloudsonmars.com/contact/ Subscribe for more episodes of Clouds on Mars Products #SAP #MicrosoftFabric #DataIntegration #ERP #DataPlatform #CloudsOnMars

  5. Jul 30

    Stop Paying Per User: How Microsoft Fabric Changes Power BI Licensing

    Most companies treat their Power BI Premium renewal as a formality. Microsoft is retiring it, and the switch to Fabric changes more than your invoice.Marcin Kubica sits down with BI Consultant Alesia Abrazova to break down why organizations are migrating from Power BI Pro/Premium to Microsoft Fabric, and how to do it without wasting your budget. See why the old Premium license is being phased out, the real difference between paying per user and paying per usage, and a practical 4-stage framework for planning your own migration.

  6. Jul 21

    The Market Doesn't Need Another Carbon Calculator. It Needs Better Data.

    ESG reporting isn't disappearing, it's maturing, and most companies don't need another carbon calculator. They need better data.Marcin Kubica sits down with ESG expert Wojtek Pastyrczyk to show Vega, our emissions reporting tool built natively on Microsoft 365. See a live demo of scope 1 and 2 data entry, audit trails, and Power BI dashboards, and hear how AI-powered invoice processing can automatically turn a PDF utility bill into GHG Protocol-aligned CO2 data. Have a data, analytics, automation or AI challenge you'd like us to dig into? Get in touch, we'd love to hear about it.Talk to Wojtek about Vega and ESG reporting: https://www.linkedin.com/in/wojciech-pastyrczyk/Talk to Marcin about your business challenge:https://www.linkedin.com/in/marcinkubicaquality/Follow Clouds on Mars on LinkedIn: https://www.linkedin.com/company/clouds-on-mars/?viewAsMember=trueSubscribe for more episodes of Clouds on Mars Products#ESGReporting #SustainabilityData

  7. Jun 23

    AI Assistant for Project Managers: From Meeting Notes to Real Follow-Ups

    After eight back-to-back meetings, a project manager still needs to write up notes from all of them. Then update JIRA, draft follow-up emails, and check what action items are still pending. The admin is the job - not the project. This episode covers PM Manager - an AI assistant built specifically for project managers, designed to handle 60-70% of the administrative work that currently consumes most of their day. Marcin Kubica and Aleksandra Gepert, a project manager with 14 years of experience, discuss what makes the PM role genuinely hard - and why generic AI tools like Copilot or ChatGPT don't solve it. The assistant is context-aware, plugged into the project manager's actual environment, and learns how each person works over time. You'll take away: Why general-purpose AI tools fall short for project managers: they lack project context and require manual effort to useWhat four core integrations (calendar, documentation, email, task management) look like when filtered by project contextHow auto-generated meeting minutes, gap detection, and inconsistency flagging work in practiceWhy the assistant acts as a coach for less experienced PMs and an accelerator for seasoned onesWhat it means for organizations: visibility into project health and into who their strongest players areKey topics covered: The admin burden on project managers: why context switching across multiple projects creates constant overheadHow PM Manager auto-generates meeting minutes and flags action items, gaps, and inconsistencies from transcriptsAutomatic task creation in JIRA based on what was discussed - not what someone remembered to logThe difference between a broad AI copilot and an assistant that knows your project, your role, and your permissionsWhat "learning to work with AI" means in practice: how your communication style shapes the quality of outputsProject managers are not administrators. The admin work expanded because there was no better option - someone had to close the loop. PM Manager doesn't make project management easier by simplifying the work. It removes the parts that shouldn't have been the PM's job in the first place, so they can focus on what actually moves a project: relationships, decisions, and unblocking teams. Watch full episode: https://www.youtube.com/watch?v=jI46Dq211DI&list=PL053kgqQv_fspiVbwCStY8rTcjldz4DtO&index=6 Website: https://cloudsonmars.com Follow the host: https://www.linkedin.com/in/marcinkubicaquality/

  8. Jun 23

    Why Data Quality Fails Silently - and How GODI Fixes It

    One of Clouds on Mars's clients had over 5,000 business rules to validate incoming data. Bad data kept coming in anyway. The rules were checking format. Nobody was checking context. This episode covers GODI - Governed Operational Data Improvement - a framework built to catch data errors that pass technical validation but are wrong in the context of how a business actually operates. Marcin Kubica and Aleksandra Gepert walk through a real case: a procurement team entering a 300-day lead time for a component that always arrives in 14 days. The system passed it. Planning triggered emergency orders. Production was at risk. GODI is built to catch exactly this kind of error - early, before it cascades. You'll take away: Why structural validation rules are never complete - and why they become outdated as business changesHow GODI's four layers work: from deterministic business rules, through historical context, to category comparison and closed feedback loopWhat three levels of user response (accept, correct, escalate) feed back into the ML model and make it sharper over timeWhy organizational readiness matters more than technology in an implementation like thisWhat 90% recommendation accuracy and 3% false positive rate looks like in practiceKey topics covered: The difference between format-correct data and contextually correct dataHow a single data entry error (300 days vs. 14 days lead time) can trigger emergency shipments and production holdsGODI's four-layer detection model: business rules, historical context, category comparison, learning loopWhy escalation cases - where the operator isn't sure - are the strongest learning signal for the modelPost-mortem culture vs. built-in learning: why lessons from incidents rarely make it back into the systemData quality tools tend to focus on rules. GODI focuses on judgment. The difference is that rules catch what someone anticipated in advance. GODI catches what didn't fit the pattern - even when no rule was written for it. Over time, with human feedback at every step, the system gets better at distinguishing what's genuinely wrong from what just looks unusual. That's a different kind of reliability than any static rule set can provide. Watch full episode:https://www.youtube.com/watch?v=NZbJr-jTK-o&list=PL053kgqQv_fspiVbwCStY8rTcjldz4DtO&index=5 Website: https://cloudsonmars.com Follow the host: https://www.linkedin.com/in/marcinkubicaquality/

See All (13)

About

Clouds on Mars Products is a podcast about what's actually happening at the intersection of technology and business: AI systems, data platforms, automation, and the real decisions behind them. Each episode, host Marcin Kubica (Head of Product at Clouds on Mars) sits down with guests and experts to explore the products and solutions being built for industries like FMCG, retail, healthcare, pharma, and real estate, as well as the broader technical and architectural challenges every modern organization has to face. Let your data make an impact.

Information

  • Creator
    Clouds On Mars
  • Years Active
    2026
  • Episodes
    13
  • Rating
    Clean
  • Copyright
    © Clouds On Mars
  • Show Website
    CoM Products