The typical large American company operates with hundreds or thousands of software tools, including ERP systems like SAP and Workday, numerous SaaS applications, workflows, scripts, automations, and even massive spreadsheets. Companies often lack visibility into their software inventory and spending, yet face endless repetitive tasks. While AI promises to automate most of these tasks in minutes without coding, the reality is more complex.

AI can generate tools rapidly, making automation accessible to non-engineers, but most workers are not tool builders. Lawyers, salespeople, and other professionals focus on their core duties, not on optimizing software. Solutions like Excel templates or Claude for X offer assistance but fall short of transforming workflows.

The concept of the 'forward-deployed engineer'—a builder who identifies automation opportunities unseen by end-users—highlights a deeper issue: many problems are hidden, bundled, or not recognized as problems until they're automated. Successful software companies often undergo multiple failed attempts before finding the right approach. Simply making coding easier doesn't solve the core challenge of identifying the need for a tool in the first place.