AI Shifts Consulting Roles from Coding to Context

For many technology professionals, the daily work of building software used to define their career. Now, the bottleneck has shifted. According to Justin Johnsen, a lead technical architect at KPMG, the primary challenge is no longer writing code by hand. Instead, it is gathering the specific context and data that artificial intelligence needs to produce useful results.

This change means that engineers are spending more time understanding business needs and refining intent before any development begins. The goal is to act as a bridge between raw data and actionable insights, ensuring that AI tools do not just generate text, but create functional assets for clients.

Context becomes the primary asset

AI models are powerful, but they are only as good as the information they receive. Johnsen explains that much of his work involves collecting the right signals to help AI reason across complex datasets. In one recent project, his team used this approach to create a risk scoring system for a client. By aggregating data from various sources, the AI could assess how risky a specific business partner might be.

This method allowed the client to deliver a complex application in about one month, a process that might have taken much longer with traditional coding. The transparency of this method was a key factor. Because the process relied on clear data and context, the client could see exactly how the solution was built, which increased their trust in the final product.

From handoff to ongoing collaboration

Traditional software projects often end with a simple handoff. A vendor builds the system, delivers it, and leaves. This can leave client teams struggling to maintain or adapt the new tools. Johnsen describes this as a dynamic where the knowledge is lost once the project concludes.

The new model involves staying on-site and working directly with the client’s team. This allows for real-time coaching and knowledge transfer. Instead of just receiving a finished product, the client team learns how to use AI effectively and how to maintain the system. This approach ensures that the capability stays within the organization, rather than being tied to a single external engineer.

The trade-off for human judgment

There is a catch to this efficiency. By automating the repetitive tasks of coding, the role of the engineer changes significantly. The trade-off is that humans must now focus more on high-level decision-making and problem definition. If the context provided to the AI is flawed or incomplete, the output will be useless.

As reported by GN technics/ai (en-US), this shift requires a different set of skills. Engineers must now combine technical depth with strong business understanding. The ability to communicate clearly and define the problem correctly is now more valuable than the ability to write complex code lines. This move prioritizes human judgment over mechanical execution.