Chloe Maraina is a visionary in the realm of business intelligence who believes that data is not just a collection of numbers but a living narrative waiting to be told. With a background that bridges the gap between complex data science and intuitive visual storytelling, she has spent years helping enterprises transform their raw information into strategic assets. As an expert in data management and integration, she has observed firsthand how the traditional boundaries of analytics are being dissolved by the rapid emergence of autonomous agents and conversational interfaces. Her perspective is particularly vital now, in 2026, as the industry moves away from static reports toward a future where intelligence is proactive, invisible, and deeply integrated into the very fabric of daily business operations.
The following discussion explores the seismic shifts currently reshaping the analytics landscape, focusing on the transition from traditional dashboards to agent-driven insights. Key themes include the rise of “headless” business intelligence, which allows data analysis to occur within third-party applications like Slack or specialized AI assistants, and the critical importance of building a robust “knowledge layer” to provide much-needed context to language models. We also delve into the changing competitive environment, where legacy providers are now facing off against ERP giants and AI startups, and the urgent need for the “DataFam” community to evolve their skills toward maintaining semantic depth and decision intelligence rather than just building visualizations.
As we move further into this era of agentic analytics, how is the fundamental interface of business intelligence shifting away from the traditional dashboards that defined the last decade?
For a long time, the dashboard was considered the holy grail of our industry, the ultimate destination where data was finally made visible and actionable. But we are seeing a profound shift where the dashboard is no longer the deliverable; it is becoming a byproduct of a much more fluid, conversational process. In the past, an analyst might spend weeks or even months crafting a perfect suite of reports, only for a business user to find them slightly out of date the moment a new question arose. Now, the interface is becoming a conversation where users can simply ask a question in natural language and receive a detailed analysis, supported by a visualization, in a matter of seconds. This change is visceral for many of us who grew up with the “DataFam” culture, as the artifact—the chart or the report—is no longer the end goal. Instead, the decision itself is the deliverable, and the analysis often happens entirely outside of a dedicated BI tool, such as when a team member shares a spreadsheet with an AI assistant and generates a insight-rich chart instantly. We are moving from a world of static, reactive observation to one of dynamic, proactive engagement where the software itself handles the instructions for reports and presentations that used to require manual labor.
With the introduction of “headless” BI capabilities, there is a concern that the analytics platform itself might become invisible to the end user. How do you balance the convenience of integrated insights with the risk of a vendor losing its brand presence?
The move toward headless BI is essentially an admission that analytics should never have been a separate discipline or a destination you “go to” to do work. By utilizing APIs and software development kits, we can now push the power of a platform like Tableau directly into the environments where people actually spend their time, whether that is Slack, Microsoft Teams, or even a custom ChatGPT interface. It is a bold bet, particularly since it means users might not even realize they are using a specific BI tool when they pull a chart into their collaboration feed. There is a real tension here because providing these capabilities gives customers a reason not to open their primary BI instance, which could lead to a perceived loss of value or “seats.” However, I believe the right direction is to turn that risk into revenue by ensuring that the semantic models and governance frameworks remain the foundational source of truth. If a vendor’s context and knowledge are being used to ground an agent in a different application, they are still providing the critical value that prevents AI from hallucinating or delivering misleading outputs. It is about being the “brains” of the operation even if you aren’t the “face,” and that requires a fundamental shift in how we measure success and brand loyalty in a world where the interface has largely disappeared into a universal chat experience.
Competition in the analytics space is no longer just among a few specialized vendors; we are seeing ERP providers and AI startups entering the fray. What does this mean for the “standard bearers” of the industry?
The barriers to entry in the analytics market have been significantly reduced, creating a barrage of new competition that feels almost overwhelming for traditional specialists. It is no longer just about beating a direct rival like Microsoft or Qlik; now, we are looking at data platform giants like Snowflake and Databricks, as well as major application providers like Workday, all offering their own integrated intelligence layers. Even AI-first companies like OpenAI are becoming direct competitors by providing chat interfaces that can handle sophisticated data tasks that used to require a dedicated analyst. This altered landscape creates a challenging environment where some of the problems BI vendors spent decades solving are now considered “solved” by more generalist tools. To stay relevant, the standard bearers must eat, sleep, and breathe analytics in a way that others don’t, pushing the boundaries of what a platform can do beyond simple query and response. They need to move into areas like decision intelligence and scenario planning—capabilities that we still don’t see enough of in the broader market—to ensure they aren’t just repackaging insights but are actually facilitating the complex evaluation of alternatives that leads to better business outcomes.
You’ve emphasized the importance of a “knowledge layer” as a differentiator for modern platforms. Why is context becoming the new currency in data management?
AI models are incredibly powerful, but they are often functionally blind without the specific context of an enterprise’s unique domain. This is why the launch of a dedicated knowledge layer—like the one we saw previewed during the webinar on August 24—is so critical to the future of agentic analytics. This layer acts as a data foundation, a semantic engine built on decades of modeling expertise that agents can tap into to understand what the information actually means. Without this, AI outputs are frequently misleading or outright incorrect because the model lacks the “connective tissue” of business rules, security protocols, and governance. When we head into major events like Dreamforce in San Francisco, which is held from September 15-17, the conversation is really about how this knowledge acts as the “unlock” for successful agentic systems. We are seeing a pivot where expertise in data analysis is being transformed into fuel for AI, ensuring that when an agent performs a task, it is doing so with a deep understanding of the enterprise’s history and constraints. It is about moving beyond just raw data to a state where the software possesses a proactive intelligence that surfaces insights users might never have thought to look for on their own.
The community of data professionals, often called the “DataFam,” is facing a period of intense transition. What kind of training and tooling do they need to remain effective as AI takes over the mechanical aspects of analysis?
There is a growing concern that simply providing new AI features isn’t enough to help the core community of analysts and data leaders who have built their careers on traditional BI tools. We need a new focus on training, tooling, and credentials that emphasize maintaining knowledge with the depth required for AI, rather than just learning how to re-package existing insights into a chat window. The role of the analyst is shifting from the person who creates the visualization to the person who curates the semantic model and ensures the AI is grounded in reality. This requires a much more profound understanding of data quality and governance, as the quality of the data is ultimately what determines whether an agentic system will succeed or fail in a production environment. We need to empower these professionals to be the guardians of context, the ones who can audit the AI’s logic and ensure that the “proactive intelligence” being surfaced is actually relevant to the strategic goals of the business. It is a transition from being a builder of artifacts to being a designer of intelligent systems, and that requires a level of technical depth and strategic thinking that goes far beyond the basics of SQL or dashboard design.
Looking at the current trajectory of the industry, what is your forecast for the evolution of analytics through the remainder of this decade?
I believe that for BI vendors to truly thrive from 2026 to 2028, they must learn to live and breathe in an agentic world by incorporating much deeper decision intelligence into their core offerings. We are currently in a phase where every vendor is moving in a similar direction, and there is a danger of stagnation if we don’t move past simple “chat with your data” features. The real growth will come from platforms that can handle complex scenario planning, allowing users to evaluate multiple alternatives and see the potential outcomes of their decisions before they make them. We will see the knowledge layer become a standalone revenue generator, where semantic models are sold and licensed to ground agents across the entire enterprise ecosystem, regardless of which development tools are being used. By the end of this decade, the most successful analytics platforms will be those that have successfully transitioned from being a “tool for looking at data” to being an “engine for taking action.” The risk of being marginalized by universal AI interfaces is real, but those who can provide the necessary context, governance, and scenario-based reasoning will find themselves more essential to the C-suite than ever before.
