Monday, November 2, 2026 | 7 mins read11/02/2026 | 7 mins read
Connecting AI to enterprise data is not enough; organizations must also provide business context, preserve expert judgment, and ensure decisions remain governed and traceable. Through a fictional offshore exploration scenario, North Basin, this article illustrates how AI can connect external signals, enterprise knowledge, operational data, and human expertise through a six-step learning loop: See → Understand → Decide → Act → Remember → Improve. The goal is to retain evidence, corrections, approvals, and findings so that knowledge becomes reusable across the organization. Ultimately, context improves decisions, memory preserves learning, and governance keeps both accountable. True institutional intelligence means reusing previous judgment without assuming yesterday's conclusions are automatically valid today.
Tuesday, October 13, 2026 | 7 mins read10/13/2026 | 7 mins read
Enterprise AI needs more than access to data—it needs business context. The same term, such as “oil production,” can mean different things to operations, refining, sales, and finance. Without governed definitions, relationships, metrics, ownership, and permissions, AI agents risk producing confident but inconsistent answers. I call this inefficiency the “Context Tax”: duplicated logic, conflicting KPIs, repeated reconciliation, wasted compute, and declining user trust. Microsoft, Snowflake, and Databricks are addressing this challenge through evolving semantic and ontology capabilities, with Microsoft also connecting business data to organizational knowledge and collaboration context. Leaders should start with one business domain, identify conflicting definitions, establish ownership, and test agents against real-world questions. Enterprise AI is often limited not by a lack of data or model intelligence, but by a lack of business context. Governed semantics can reduce ambiguity, improve explainability, and enable more consistent, trustworthy decisions.
Sunday, October 11, 2026 | 8 mins read10/11/2026 | 8 mins read
The article explores how to navigate career uncertainty by focusing on what we can control, preparing for what we cannot, and protecting our mental clarity during difficult times. It introduces the idea of <<Magneto helmet>> as a personal metaphor for maintaining focus, listening thoughtfully, and making deliberate decisions without shutting others out. Rather than letting uncertainty, external pressure, or fear dictate our actions, we should prioritize meaningful outcomes, continue learning, build relationships, and seek support when needed. Asking for help is not a weakness but a courageous skill that fosters connection, growth, and well-being. The article also offers practical advice: be specific about the help you need, approach the right people, communicate respectfully, express gratitude, and define a personal manifesto to guide future decisions. Ultimately, while we cannot control every circumstance, we can choose how we respond and take the next meaningful step.
Tuesday, October 6, 2026 | 6 mins read10/06/2026 | 6 mins read
This edition explores how Data and AI professionals can move beyond technology-first conversations to create measurable business outcomes. Drawing on consultative selling, it connects curiosity with Challenger, Value Selling, and MEDDPICC to clarify why change matters, quantify value, and navigate decisions. It also introduces Nemawashi as a thoughtful approach to stakeholder engagement and consensus-building. The central message is simple: technical expertise delivers greater value when combined with curiosity, measurable outcomes, stakeholder alignment, and the humility to revise assumptions.
Thursday, October 1, 2026 | 17 mins read10/01/2026 | 17 mins read
After several days at FabCon Barcelona, Pablo Junco sees Data Agents 2.0 as a major evolution—from simply talking to data toward understanding the business behind it. The key shift is context engineering: combining instructions, descriptions, examples, Topics, semantic models, Ontology, Graphs, and unstructured content so agents can retrieve the right business context before generating queries. Advanced NL2SQL/NL2DAX, smarter routing, observability, and richer sources move Data Agents closer to context-aware virtual analysts. Fabric IQ adds business meaning, while Work IQ could bring organizational context—what Pablo calls “Your Company IQ.” The goal is not just better answers, but trustworthy, governed, explainable intelligence that helps people and AI make sense of enterprise data. And this journey extends beyond analytics: Fabric Apps can turn governed data into production applications and workflows.
Tuesday, September 29, 2026 | 5 mins read09/29/2026 | 5 mins read
At FabCon Europe in Barcelona, Microsoft unveiled major Fabric innovations, reinforcing the platform’s evolution from analytics toward intelligent, governed action. OneLake continues its momentum with 76B daily interactions, 13M shortcuts, and 300%+ YoY storage growth, alongside expanded mirroring support. Five highlights stood out: Fabric Apps, enabling applications directly on trusted data and semantics; Fabric Observability, providing end-to-end visibility across code and agents; Data Engineering Agents, shifting engineers toward orchestrating intelligent systems; the Database Hub, bringing database discovery, management, and AI-powered operations into Fabric; and enhanced governance, including Fabric Policies, Purview DLP, security, and capacity management. The broader message is clear: enterprise AI requires more than intelligence—it needs governance, security, operational control, and sustainable economics. The journey is Data → Context → Intelligence → Agents & Apps → Action → Value.
Tuesday, September 22, 2026 | 10 mins read09/22/2026 | 10 mins read
This article explores the shift from AI that answers to AI that acts. As AI Agents gain more autonomy, the challenge is no longer only capability—it is trust. The article argues that Agentic AI makes Data Governance more important, because poor or misunderstood data can lead not only to bad answers, but to bad actions. Trusted AI requires more than governed data. Agents need business context: understanding entities, relationships, semantics, signals, and how the business actually works. The article connects Data Governance, business semantics, Fabric IQ, Responsible AI, observability, permissions, and human control into a foundation for Trusted AI Agents. The central idea is: Capability will move AI forward. Context will help it understand. Governance will keep it grounded. Trust will determine how much autonomy we give it.
Tuesday, September 1, 2026 | 9 mins read09/01/2026 | 9 mins read
This edition explores why the question should not be Databricks or Fabric? but rather What business outcome are we trying to achieve, and where is the best place to achieve it? The article compares architectural considerations across data integration, warehousing, BI, governance, real-time analytics, AI and agents, automation, and observability. Databricks offers strong engineering, lakehouse, streaming, and AI capabilities, while Fabric provides an integrated SaaS experience across Data Factory, OneLake, Power BI, Real-Time Intelligence, and AI. The key message: organizations do not need to force every workload into one platform. A combination of technologies may deliver greater business value. Ultimately, architecture should be business-centric, not platform-centric—choosing the right tool for each requirement rather than simply using what is already available.
Monday, August 17, 2026 | 7 mins read08/17/2026 | 7 mins read
Databricks is often positioned as the answer to cloud vendor lock-in—and its multi-cloud architecture provides a compelling portability advantage. But cloud portability is not the same as platform portability. As organizations adopt Unity Catalog, Databricks-specific pipelines, AI services, Apps, governance, and operational processes, dependency can shift from the cloud provider to Databricks itself.Microsoft Fabric presents the opposite trade-off: deeper alignment with Microsoft and Azure, but potentially less analytics-stack complexity through a highly integrated SaaS platform. The real question isn't which platform has “less lock-in.” It is where an organization is willing to accept dependency in exchange for business value. The article also proposes a practical discipline: assess migration cost, complexity, and time-to-production every year. You may never migrate—but understanding your exit cost is one of the best ways to measure how dependent your architecture has become.
Tuesday, August 4, 2026 | 8 mins read08/04/2026 | 8 mins read
Edition #15 introduces the concept of Scalable Agentic Value: the ability to transform isolated AI pilots into enterprise-wide capabilities through four essential pillars—Governance, Multi-Agent System (MAS) orchestration, FinOps, and Responsible AI. Rather than focusing on model quality, the article argues that AI success depends on operating discipline, organizational readiness, and leadership alignment. It explains why enterprises need clear ownership, trusted data, controlled experimentation, and a control plane (such as Microsoft Agent 365) to govern identity, policies, observability, and costs across AI agents. The edition also explores the shift from assistive AI to autonomous agentic operations, emphasizing bounded autonomy, human accountability, and ethical oversight. It concludes that organizations gain lasting competitive advantage not by adopting AI faster, but by institutionalizing learning, governance, and measurable business outcomes through scalable agentic systems.
Tuesday, July 28, 2026 | 6 mins read07/28/2026 | 6 mins read
Microsoft Fabric Apps (currently in Preview) transform Fabric from an analytics platform into a foundation for building business applications powered by trusted data. Instead of exposing users to reports and dashboards, Fabric Apps deliver guided, persona-based experiences that combine data, AI insights, and recommended actions. RayFin, Microsoft's open-source framework, accelerates development with reusable components, security, and deployment automation. The article explains when to use Lakehouse versus SQL Database, recommends two enterprise architecture patterns—read-only analytics and transactional applications—and clarifies how Fabric Apps complement, rather than replace, Power BI Apps and Power Apps. Although still in preview, Fabric Apps represent a major evolution for Microsoft Fabric, enabling organizations to move from simply delivering insights to creating AI-powered business experiences built on governed data.
Monday, July 27, 2026 | 4 mins read07/27/2026 | 4 mins read
A year after a major reorganization changed my role from people manager to Principal Solution Engineer, I realized that resilience is not about protecting a title—it's about protecting your mindset. For years, I believed that working harder, overdelivering, and becoming indispensable would create security. This experience taught me otherwise. Like Magneto's helmet shielding him from psychic attacks, we all need to protect our minds from doubt, negativity, and the constant noise around us. Instead of focusing on what I lost, I chose to learn, reinvent myself, invest in my family, and stay open to new opportunities. Today, I no longer define myself by an organization chart or job title, but by the value I create, the integrity I uphold, and my commitment to continuous growth. True success isn't your position—it's the person you become through adversity.
Subscribe onLinkedIn to get the latest Pablo Junco's perspectives and insights to move from messy data to measurable outcomes — governed platforms that power agentic AI.
If you’re a Chief Data Officer (CDO), a data leader, or simply someone who believes in the power of preparing data for AI—you’re already a Data Massagist.
Whether you have an idea, a challenge, or just want a fresh perspective, let’s connect. I’m always open to collaborating, learning, and helping others move forward.
