November 2025 Summaries

Stepping into the role of Head of Data requires a significant shift in mindset from being an individual contributor (IC) focused on technical tasks to a leader driving business strategy and impact. Newly promoted data leaders often struggle to detach from hands-on coding and instead need to prioritize strategic decision-making, team delegation, and fostering stakeholder relationships. The success of a data leader is not measured by the volume of technical outputs but by the business impact those outputs have. It is crucial for data leaders to align their efforts with business goals, cultivate supportive stakeholders, and communicate the value of data initiatives in terms that resonate with business needs. This transition involves learning new skills, such as managing relationships and expectations, which are vital for effectively leading a data team and delivering impactful solutions.

Carlos Aguilar, Hex's Head of Product, argues that traditional checklist-based evaluations for AI analytics tools are ineffective, emphasizing that success depends on understanding context management and real-world user interaction rather than merely comparing features. As conversational analytics begins to unlock self-service capabilities for business users, Aguilar suggests that data teams should lead evaluations by considering both end-user and data team experiences. This involves testing tools with real users and questions, improving context, and monitoring responses to ensure accuracy and relevance. The evaluation process should focus on how well a tool manages context, improves over time, and integrates into existing workflows, rather than relying on simplistic feature comparisons. Aguilar highlights the need for a thorough evaluation approach to determine if a tool will be effective within an organization, advocating for an understanding that conversational analytics can transition from an experimental phase to operational use.

Hex Technologies has become an essential tool for data teams and product managers in various industries, allowing them to leverage AI and data analytics for actionable insights. With features like the Notebook and Threads agents, users can predict customer churn, track product adoption, and align business performance seamlessly. For example, Mark Boongaling at aXcelerate developed a customer risk-scorecard app that predicts churn by integrating data from multiple sources, while Jack Christensen uses Hex to build data apps that help measure the success of new product features without the need for engineering support. At Jampack, David DyTang utilizes Hex to create data applications for tracking revenue, event pacing, and traveler behavior, empowering smarter business decisions. Hex's ability to combine SQL, Python, and visualization tools in a single platform allows users to transform complex data into meaningful reports and applications quickly, enhancing data-driven decision-making across organizations.

Katie Bauer, Hex's Head of Data, emphasizes the importance of balancing speed with accuracy in data delivery, arguing that while traditional data workflows have focused heavily on precision, they often fail to meet the speed expectations of stakeholders. She highlights the evolving role of AI-powered agents in transforming data teams' ability to quickly process and respond to natural language queries, thus allowing stakeholders to engage more dynamically with data. By leveraging AI, data teams can iterate rapidly, providing immediate, albeit initially imperfect, answers that can be refined over time, fostering a cycle of continuous learning and trust-building. This shift towards agentic analytics, as seen with Hex’s Threads capabilities, is poised to redefine the way businesses access and utilize data, marking a potential new era where data teams can scale their operations and strengthen their role as trusted partners.

Rachel Herrera explores the distinction between data apps and traditional dashboards, using her experiences at Hex Technologies to illuminate the shift in data visualization and interaction. She describes her initial confusion, rooted in a decade of experience with dashboards, and how data apps offer a more dynamic and interactive experience compared to static dashboards. Data apps allow for deeper exploration and direct action on data, as exemplified by Hex's integration of SQL, Python, and visualization in a single platform, enabling users to input data, update databases, and immediately see the results. This approach contrasts with traditional dashboards, which often require multiple tools for comprehensive data interaction and analysis. Herrera emphasizes the importance of understanding this evolution, encouraging readers to consider whether a project might benefit from being developed as a data app rather than a dashboard, highlighting the potential for enhanced user exploration and workflow integration.

AI-driven self-service analytics agents can either enhance or undermine a team’s expertise, largely depending on the quality of context engineering. Context refers to the specific business knowledge imparted to AI agents, enabling them to understand the unique data environment, distinguish between similarly named metrics, and deliver accurate and trustworthy results. This process shifts the data team's role from repeatedly answering questions to building a "context lake," a repository of business knowledge that AI can draw from. The approach involves curating existing data resources into a structured framework that AI can utilize to provide consistent, reliable insights. Effective context engineering combines various tools, such as warehouse metadata, endorsements, rules files, and semantic models, to guide AI behavior. This strategy fosters a virtuous cycle where business users can confidently self-serve, allowing data teams to focus on strategic initiatives rather than repetitive queries. By progressively improving context rather than seeking perfection, organizations can scale their expertise and enhance the value of AI analytics.