The Promise of Agentic Analytics

In its latest Looker product announcement, Google unveiled “agentic analytics”—a term meant to capture the evolution of AI-powered business intelligence—at the core of its updated Looker platform. Through a blend of Gemini’s conversational AI and the Looker semantic layer, Google presents a future where business users interact with data as naturally as they use their favorite apps. With Looker Agents, users can access and analyze data, build visualizations, and even execute Python-based logic through natural language prompts—all while reducing technical dependencies.

Google’s premise starts with a challenge that has burdened data teams for ages, what Google identifies as the “over reliance on small teams of data experts.”

This reliance creates an endless, burdensome and frustrating backlog of analytics requests. I am quite familiar with this issue in my work leading and guiding analytics teams. The solution typically requires a mix of business-aware prioritization, effective organization of data teams across the conflicting imperatives of centralization and decentralization of resources and responsibilities (per the Insights Supply Chain and the Data Org Matrix), and strategic alignment between data functions and the business. Tooling deficiencies almost never top the list of demands.

Of course, Google’s focus in its product announcement is on tooling. Thus, enter Google’s latest solution to this age-old tension: Looker Agents.

Looker Agents are AI-powered interfaces that allow users to embed custom analytical assistants directly into chat applications, enterprise software, or bespoke workflows. These agents can retrieve data, run advanced queries, and integrate seamlessly with other systems to facilitate timely, contextualized decision-making. According to Google, this advancement shifts business intelligence from traditional dashboards to fluid, dialogue-based insights, bringing analytics to the point of decision-making.

The vision genuinely sounds compelling. Yet if this new era of agentic analytics must do more than improve interfaces, automate queries, and democratize the generation of insights. This new era still faces deep, structural challenges in organizations with data professionals —chief among the challenges, the methodologies of analytics leadership.

Beyond Tools: The Ongoing Role of Data Professionals

From my perspective as an experienced analytics practitioner, I welcome tools like Looker Agents that promise to reduce friction and accelerate workflows. I can easily imagine reducing my time spent navigating syntax rules, troubleshooting LookML, searching documentation, and organizing a collection of promising insights into an inspirational narrative. However, I also appreciate that AI tooling does not (yet?) replace strategic thinking, comprehensive data governance, or collaborative insight development. The AI tooling supplements the human foundation of high-quality analytics and insights.

In my critique of Looker’s promotion of conversational analytics, I argued that while AI interfaces can democratize access to data, they do not remove the need for data professionals. Rather, they shift the focus of our roles—from builders of dashboards to designers of insights-delivering ecosystems. Looker’s new capabilities continue this trend. These tools will empower data professionals to spend less time on manual tasks and more time curating data models, guiding business users, and aligning analytics strategy with organizational goals.

Analytics Leadership and the Insights Supply Chain

To frame this evolution, I use the concept of the Insights Supply Chain. This framework describes how organizations convert raw data into actionable insight through interconnected functions: data engineering, governance, modeling, analysis, and storytelling. When this chain functions properly, guided by capable analytics leadership, it delivers clarity and alignment. When this chain breaks—when insights are generated without context or models are used without governance—the data ecosystem can suffer from the very backlogs and frustrations that Google seeks to remediate with Looker Agents.

However, tooling alone cannot fix a broken supply chain. Analytics leadership needs to integrate emerging technologies like AI into a broader ecosystem of knowledge management and cognitive diversity. AI is a hyperscaler—it can extend the reach of analytics, but it must be guided. Knowledge management organizes the environment. Cognitive diversity fuels it with varied perspectives and interpretations that challenge groupthink and expand the solution space.

Unlocking Value Through Cognitive Diversity

Cognitive diversity—defined as differences in thinking styles, problem-solving strategies, and experience—is especially critical in today’s complex data environments. When organizations foster this diversity and embed it in analytics workflows, the results can be transformative. In this sense, AI becomes another facet of cognitive diversity that complements human cognitive diversity.

The real opportunity of agentic analytics lies not just in eliminating the dependency on analysts, but in creating systems where AI, knowledge management, and cognitive diversity interact constructively. Business users may interact with agents, but data professionals must still design the systems, define the metrics, and steward the governance structures that make those interactions meaningful.

Conclusion: From Insights to Impact – the On-Going Central Need for Data Professionals

Through this context, Looker Agents represent an important next step. They can reduce technical friction, improve access to insights, and bring data to the forefront of decision-making. But without thoughtful analytics leadership, they risk becoming sophisticated tools in service of superficial insights. AI can even produce a surplus of insights that humans lack the ability to prioritize, scrutinize, and effectively implement. Analytics leadership must maintain a strategic framework for effectively channeling all this analytics energy.

Agentic analytics will not make up for poor analytics leadership—but in the hands of thoughtful leaders who understand the interplay of technology, people, and process, they will unlock new levels of organizational intelligence and impact.