October 2025 Summaries

Connor McIntire discusses the transformative impact of using Hex’s AI conversational data-chat interface, Threads, in accelerating the onboarding process for new sales leaders. Within just eight days of his new sales leadership role, McIntire was able to gain a deep understanding of his team's performance, which traditionally would have taken months. Threads enabled him to access and analyze sales performance data without needing extensive technical skills, allowing him to make informed strategic decisions and engage in meaningful one-on-ones with his team. By leveraging Threads, McIntire quickly transitioned from a new leader to a data-driven strategist, identifying key account opportunities and enhancing team productivity and pipeline generation. The AI tool not only facilitated better decision-making and strategic planning but also encouraged a cultural shift within the team, promoting proactive data analysis and strategic proposal development by team members themselves.

Hex Technologies has launched a new Slack integration and Model Context Protocol (MCP) support for its Hex Agent, enabling users to access data insights directly in their existing workflows. The Slack integration allows team members to ask data-related questions by mentioning @Hex in any channel, with the agent providing answers, explanations, charts, and links to relevant Hex projects. This integration is available to Explorer+ seats on Team and Enterprise plans for US multi-tenant stacks, while the MCP server extends Hex's capabilities to code editors and AI tools like Claude Desktop and Cursor, facilitating seamless data analysis without custom API integrations. These innovations aim to make AI-powered analytics more accessible and transparent, allowing users to collaborate effectively and make informed decisions.

Data teams have long aimed to facilitate self-service analytics, but traditional methods often resulted in bottlenecks and frustration due to slow response times. The emergence of AI technologies has shifted this dynamic by enabling team members to independently conduct analyses and answer questions using natural language processing, although it introduces risks of potential inaccuracies. Despite the inherent unpredictability of large language models (LLMs), they empower users to quickly obtain insights, reducing dependency on data teams and allowing these teams to focus on more complex strategic issues. To support this AI-driven self-service model, there is a need for well-structured data systems, comprehensive documentation, and clearly defined semantic layers to ensure consistency and accuracy. This evolution highlights the data team's role as architects of reliable systems rather than mere gatekeepers of information, ultimately broadening the impact of data across the organization and encouraging more strategic use of analytics.

Hex Technologies has enhanced its Notebook Agent by utilizing error clustering to better understand and address the agent's reasoning processes, rather than just its errors. Initially relying on feedback and internal testing, the team realized the need for a more structured analysis as the agent's adoption grew. By clustering similar errors and using natural language labels, they identified key areas for improvement, particularly in handling Python and SQL errors, cell timeouts, and vague error messages. These insights led to refinements such as improving the agent’s comprehension of Hex tools and altering its response to cell timeouts, thereby enhancing its ability to recover from errors gracefully. This approach not only optimized the agent's performance but also facilitated the creation of an internal tool that continuously updates with new data, helping engineers prioritize future enhancements. The development process itself was aided by the Notebook Agent, which wrote queries and analysis code, demonstrating the tool's capability in real-world applications.

The blog post from Hex Technologies discusses the development challenges and insights gained while building the Notebook Agent, an AI tool designed to assist in data analysis within their platform. It highlights the unique difficulties faced by analytics agents compared to coding agents, emphasizing that data exploration requires judgment and cannot be unit-tested like code. The post explains that data objects can be unpredictably large and lack predefined abstractions, making it difficult to summarize without losing critical information. The team tackled these challenges by implementing context engineering principles such as structuring the context as a computational map, setting token boundaries, and explicitly communicating limitations to the agent. These strategies improved the agent's efficiency and understanding of the analytical workflow, ultimately enabling it to navigate data flows more intelligently. The blog underscores that the key to effective analytics agents lies in representing complex environments in ways that facilitate AI reasoning, rather than just fitting more data into a context window.

The blog post by Ben Rogojan critiques the prevalent trend of being "data-driven" in business, arguing that an excessive focus on data can lead to decision-making paralysis and inefficiency. Rogojan suggests that companies often become bogged down in building data infrastructure and perfecting analytics tools, which can distract from their actual business goals. As an alternative, he advocates for an "outcome-obsessed" or "decision-driven" approach, where data is used to inform rather than dictate decisions, and emphasizes the importance of setting decision thresholds, avoiding vanity metrics, and fostering a culture of fast feedback and action. He also highlights the necessity of bridging the communication gap between data and business teams to ensure data is effectively used to drive outcomes. The post concludes by stressing that while data can provide valuable insights, it should not be the sole driver of decisions, and a cultural shift is needed to integrate data into a broader business strategy.

Marketing pipeline forecasting can transition from chaos to clarity by moving from spreadsheets to data apps, as experienced by Bryanna Clancy. Initially relying on spreadsheets for forecasting led to disjointed and unreliable predictions, causing panic and reactive measures when numbers fell short. However, adopting a data app like Hex transformed the approach by allowing seamless data ingestion from CRM systems, utilization of up-to-date conversion rates, and the ability to model specific time-to-convert patterns. This shift enabled accurate, dynamic forecasts that reduced organizational stress and eliminated the need for emergency interventions. The app's evolving capabilities, including integration with AI tools for generating diverse forecast scenarios, further enhanced forecasting precision and organizational confidence. As a result, the focus shifted from reactive measures to strategic planning, enabling data-driven decisions and fostering a proactive marketing environment.

Hex Technologies, originally conceived as a notebook for data analytics, has evolved into a comprehensive platform that challenges traditional categorizations in the analytics domain. Initially, the company grappled with defining itself within existing categories like notebooks or business intelligence (BI) tools, as these labels seemed too restrictive or carried outdated connotations. Instead, Hex has positioned itself as a versatile solution that integrates workflows for deeper analysis, data science, and data applications, enhanced by its agentic Threads feature, which leverages natural language processing for intuitive user interaction. This evolution reflects a broader industry shift towards integrated, self-serve analytics platforms that transcend traditional BI limitations, aiming to redefine how data-driven insights are generated and utilized. As Hex continues to develop, it remains focused on reshaping the analytics landscape with innovative tools that promote collaboration and insight generation, while acknowledging that its approach may not suit those simply seeking to replace existing BI functionalities.

Hex is a platform that enables data teams to create dynamic and impactful data applications, as showcased in various real-world examples. Jonathan Paquette from OpenSpace utilizes Hex and Mapbox to visualize global adoption of AI-powered construction technology, providing a crucial tool for sales and leadership. Jenn Schilling at Albert.io leverages Hex to automate the creation of school-level PDF reports, significantly reducing manual effort and enhancing accuracy, while integrating with other tools like Airtable. Rachel Whaley from RedCircle uses Hex to develop an interactive app that ranks podcast partners, simplifying data presentation and saving engineering resources. These cases illustrate the versatility and efficiency of Hex in transforming data into actionable insights across different industries.

In the rapidly evolving field of AI product development, the author reflects on the lessons learned from past experiences, emphasizing the need to adapt to shifting model capabilities rather than relying on complex engineering solutions that may become obsolete. The text discusses the importance of understanding the unique value propositions of AI products and being agile in adjusting roadmaps to leverage new model capabilities. The author recounts specific challenges faced, such as the initial failure of the Notebook Agent due to premature development and the blind spots created by focusing on "pretty good" solutions rather than exploring new opportunities. Key strategies now include validating features with early user engagement, identifying capability shifts in models, and the willingness to abandon projects that rely on clever hacks to compensate for current model limitations. The narrative encourages celebrating failures as part of the learning process and highlights the exciting potential of building AI products during technological advancements.

Agentic analytics is emerging as a key trend in the data industry by leveraging AI agents to transform the way analytics tasks are accomplished, promising enhanced depth, breadth, and speed of insights. Unlike traditional data tools that focus on single-layer dashboards, agentic AI operates iteratively, enabling more profound analysis and causal understanding by reflecting on previous steps and planning subsequent ones. This shift democratizes access to analytics, allowing more individuals within organizations to obtain insights quickly and easily, thus fulfilling the long-held dream of self-service analytics. Although deeper analyses can take time, the overall process is faster due to reduced back-and-forth and the ability of agentic systems to function asynchronously, notifying users upon task completion. As AI begins to reshape the data landscape in 2025, teams are encouraged to embrace these changes with an open mind to fully harness the potential of agentic systems.

Ramp, a company utilizing AI to enhance its data operations, has integrated Hex's Notebook Agent into its processes to streamline and optimize data analysis and reporting. As emphasized by Ricky Meyers, Ramp's Analytics Engineering Manager, AI is essential in transforming data teams from mere reporters to strategic guides. Hex's AI capabilities, particularly the Notebook Agent, allow Ramp's team to quickly build analysis frameworks, conduct exploratory analysis, and synthesize findings for easy distribution, significantly reducing the time from idea to insight. The success of these AI implementations is attributed to a robust semantic layer that provides the agent with comprehensive context and documentation, enabling efficient analysis tailored to specific audiences. While primarily used by technical and quasi-technical teams, Hex's tools also empower users who possess only basic SQL knowledge to perform self-service analytics. Ramp's approach, guided by the philosophy of "AI or die," focuses on incremental improvements, aiming to enhance the company's operations by 1% daily through AI-driven insights and innovations.

Threads is an innovative AI-driven conversational analytics tool developed by Hex, designed to empower non-technical users to engage with data using natural language while ensuring trustworthy and quality responses. Unlike traditional standalone AI chatbots that struggle with context and integration, Threads leverages Hex's existing framework and semantic models to provide a seamless experience where users can ask questions and receive detailed, actionable insights. The tool prioritizes governed data, drawing on semantic metadata and allowing users to follow its step-by-step reasoning process, which enhances both auditing and the usefulness of the outputs. Threads is integrated with Hex's platform, allowing data teams to manage data access and maintain the quality of AI responses, thereby broadening their impact within organizations. As a public beta, Threads is available to Explorers+ on Team and Enterprise plans, with extended AI usage limits during the early access period, facilitating self-service analytics and collaboration across data teams and business users.

Hex Technologies has launched a major update focused on enhancing data analytics through AI with the introduction of Threads, a new conversational analytics interface designed for both technical and non-technical users. Built on Hex's existing infrastructure, Threads integrates seamlessly with users' workspaces to leverage context, making data analysis more accurate and accessible. By using Anthropic's Sonnet 4.5 model, Threads improves the speed and capability of handling complex analytics tasks. Additionally, it allows business users to generate insights via natural language and collaborate with data teams through notebooks for deeper analysis and validation. Hex's commitment to interoperability is evident in its support for the Open Semantic Interchange initiative, enabling more flexible and portable semantic models. The platform also continues to advance its no-code data visualization tool, Explore, and has improved the functionality of its Notebook Agent, facilitating easier deep-dive analysis and dashboard creation. With all these new features live, Hex aims to bridge the gap between self-service data access and analytics, inviting users to explore its capabilities across various devices, including mobile.

Explore, a visual data analysis tool developed by Hex, has evolved significantly since its launch, providing users the ability to analyze data without writing code. Initially designed to allow Hex users to visually interact with data from a warehouse without SQL queries, Explore now supports semantic models, enabling analyses with governed metrics for reliable insights. Recent updates have enhanced its functionality with features like top n value specification, aggregate calculations, pivot table totals, and the ability to drill into underlying data segments. These enhancements allow both technical and non-technical users to perform detailed data exploration and visualizations efficiently. Users can now define reference lines in charts and filter data based on aggregated measures, akin to SQL's "HAVING" clause, all within a no-code environment. These developments aim to make Explore more powerful and accessible, with plans to integrate these features into chart cells for broader application within Hex's data tools.