First, AI has brought a massive increase in the productivity of software developers. Every software developer I know claims to be at least twice as efficient, and sometimes ten times more efficient, since the launch of ChatGPT and GitHub Copilot. And the productivity boost is not slowing down. New tools like Cursor AI are emerging, which take things to the next level by understanding the developer’s codebase and integrating a chat window into the development environment. This allows the programmer to chat with their code and give high level instructions such as “write unit tests for the entire code base”.
Second, the advent of AI agents marks a pivotal juncture for enterprises that necessitates custom development. With AI transformation underway, a new layer of software is emerging to integrate AI agents into companies’ business processes. These agents will operate autonomously over long periods of time, automating workflows from end to end, making strategic decisions, allocating resources, and steering business processes, all the while communicating in natural language with their human counterparts. Because this layer of software is so intimately interwoven into the company’s business processes and workflows, it will need to be heavily customized for each enterprise.
The emergence of ERP systems in the 80’s and SaaS systems in the 2000’s led companies to swallow the bitter pill of modifying their business processes in order to match the software systems they were able to buy. As companies begin their AI transformation, given the plummeting cost of software development and the need for tightly integrated software solutions, they will more often choose to build.
