The 'Morna, coding!' Attitude: What if We Embrace AI as Our Collaborative Partner?

Recently, I came across discussions where developers openly admit to no longer writing code manually, fully delegating programming to AI tools. Some even celebrate the end of traditional coding, describing it as a tedious task now replaced by AI. I find this attitude a bit frustrating—not because I oppose AI, but because it risks turning coding into a black box where understanding and knowledge get lost.

In Norway, many small and medium-sized businesses with 10-50 employees face rising costs for software development and automation. Hiring specialists can easily cost over 1000 NOK per hour, while administrative work like bookkeeping or integration management still runs around 400-600 NOK hourly. Companies using systems like Tripletex, Fiken, Vipps, and various government APIs often struggle with costly, complex workflows that require maintaining both domain knowledge and technical expertise.

Now, imagine if instead of fully outsourcing coding to AI, we leveraged AI as a productivity partner—handling routine, low-risk tasks while we focus on the critical, innovative parts needing deep understanding. What if AI could manage standard integrations, data scraping, or notification bots, freeing us to build mental models and solutions that truly fit the business domain? This would reduce complexity and risk, and keep knowledge accessible.

This CAN be done by combining tools like n8n for workflow automation, integrating with Norwegian APIs such as Tripletex or Vipps, and using AI APIs like Claude or Azure OpenAI to generate and optimize code snippets. I often split codebases into two categories: low-risk code that follows standards which AI can handle, and high-risk, business-critical code that I model carefully in my mind. AI accelerates the first category, while I remain deeply involved in the latter, ensuring system integrity and domain alignment.

This approach fits well for Norwegian SMBs aiming to optimize costs and productivity without losing control over their core systems. It’s less suitable for enterprises requiring large-scale custom systems or those unwilling to invest time in understanding the AI-assisted workflows. The goal is not to eliminate coding but to evolve how we code—making AI a powerful collaborator rather than a mysterious black box.

So, what if we stopped fearing AI as a replacement and started treating it as a tool to amplify our expertise? How would that change the way your company approaches software development and automation?