Moving Up the Stack: Analytics Engineering in the Age of Agents

Why It Matters

AI‑driven agents are rapidly redefining data workflows, making automation a competitive necessity and reshaping the skill set of analytics engineers.

Key Takeaways

• dbt reaches over three million daily downloads
• Agents generate more than 50% of new Hex cells
• dbt MCP server usage grows 40% month-over-month
• Analytics engineers replace manual SQL with automated pipelines
• AI agents reshape data roles, requiring advanced analytical skills

Pulse Analysis

The rise of "moving up the stack" began with dbt’s open‑source transformation layer, which turned repetitive SQL scripting into a maintainable, version‑controlled pipeline. By abstracting data models into reusable code, dbt freed analysts to focus on experimentation and business insight, driving adoption that now exceeds three million daily downloads. This shift not only boosted individual salaries but also created a community of analytics engineers who view automation as a pathway to strategic impact rather than a threat.

Today, generative AI agents are the next catalyst. Platforms like Hex report that over half of newly created cells are generated by agents, and dbt’s Managed Cloud Platform (MCP) server sees 40 % month‑over‑month growth as enterprises embed AI into their data stack. Companies such as Ramp are deploying "agentic analysts" to query data, generate visualizations, and even suggest business actions autonomously. These tools extend the dbt paradigm by providing natural‑language interfaces that translate intent into SQL, dramatically accelerating the time from question to insight.

For professionals, the imperative is clear: evolve from manual query writers to architects of AI‑augmented data ecosystems. Mastering prompt engineering, model governance, and the integration of semantic layers will become core competencies. Organizations that invest in upskilling their analytics engineers and building robust agent‑ready pipelines will capture higher ROI, while those that cling to legacy processes risk falling behind in an era where data‑driven decisions are made at machine speed. The window for this transition is narrowing, making proactive adoption essential for sustained competitive advantage.