Vertical Bar

At Vertical Bar, the core challenge was straightforward but non-trivial: teams had rich data in NetSuite, but extracting answers still required someone who understood both the business context and the underlying schema. My work focused on building an AI-driven analytics agent that could bridge that gap.

The agent ingests NetSuite transaction data and maps it to a question-first interface. Users can ask things like “Which customer orders are at risk of delay?” or “Where are we losing the most time in our fulfillment flow?” Behind the scenes, the system translates these questions into structured queries, joins the right records, and returns explanations along with charts—rather than just raw tables.

A key Builder principle in this project was aligning business language with data language. Instead of forcing users to think in terms of tables, fields, and joins, the agent was designed around real operational concepts: lead time, backlog, exceptions, bottlenecks. The job of the system is to handle the “how”—query construction, aggregation logic, and visualization—so operators can stay focused on “what decision to make.”

This work has been an ongoing exploration of how AI agents can sit on top of complex ERP systems and act as an analytic co-pilot: continuously reducing friction, surfacing issues earlier, and helping teams see their processes not just as data—but as flows that can be improved.