The Automation Paradox

The promise of generative AI sweeping away our software clutter is alluring, yet it may fundamentally misunderstand how we use technology. While AI can create tools in minutes that once took hours or days, this doesn’t necessarily solve the core challenges of automation.

Beyond Simple Repetitive Tasks

Most companies today operate with a sprawling digital ecosystem—giant systems like SAP and Workday alongside hundreds of SaaS applications, custom scripts, and countless spreadsheets. The temptation to view AI as a universal solution is understandable, but it overlooks crucial aspects of how software evolves.

First, most people aren’t natural tool builders. A matrimonial lawyer focuses on legal strategy, not workflow optimization; an enterprise salesperson prioritizes client relationships, not sales enablement tools. Even when individuals recognize inefficiencies, they may lack the technical expertise or time to create solutions themselves.

The Hidden Nature of Problems

Many successful software applications address problems users didn’t even realize they had—or whose solutions weren’t immediately apparent. Take Excel templates: while helpful starting points, each one represents a potential company built on identifying and solving previously unrecognized needs.

This creates an opportunity for “forward-deployed engineers” who can identify automation possibilities others miss—similar to how tech enthusiasts once pointed out digital inefficiencies in their parents’ businesses.

The Human Element of Change

Even with clear solutions, implementation faces human hurdles. Many workflows span multiple departments, systems, and even regulatory frameworks. A brilliant idea for automating accounts payable requires buy-in from stakeholders across the organization—a process that can take months or years.

Software adoption exists on a spectrum from top-down (company-mandated ERP systems) to bottom-up (individual spreadsheets)—from institutionalized processes to improvised solutions. The most successful applications bridge this gap by addressing real needs in ways that fit existing workflows rather than requiring complete overhauls.

While generative AI lowers the technical barrier to creating automation tools, it doesn’t solve the fundamental challenge of identifying what problems need solving and ensuring those solutions are adopted across complex organizations.