I had a conversation recently that stopped me mid-sentence.
A senior developer, someone with over a decade of Business Central experience, said to me: "I watched a junior on my team build in two hours what would have taken me a week. So, what exactly am I for now?"
It was not self-pity. It was a genuine question. And it is one most have been wrestling with too.
The Speed Illusion
Let me tell you what I have seen over the past year. Developers using AI assistants to stand up entire BC extensions in a single afternoon. Consultants generating documentation, test plans, even architecture notes across multiple projects simultaneously. One colleague described running six parallel workstreams using different AI chat sessions, each handling a different role: developer, translator, tester, architect, technical writer, project manager. None of them were real people.
What used to take weeks now takes hours. That is not an exaggeration. That is Tuesday.
And if you are someone who has spent 15 years building expertise in this ecosystem, it is natural to feel a bit rattled by that. Because the thing you were rewarded for, the ability to write solid code and solve tricky problems, is suddenly available to anyone with a subscription and a decent prompt.
But Here Is What I Keep Coming Back To
Speed is not the same as quality. And output is not the same as outcome.
I recently watched a team vibe-code a complete solution to replace a buggy ISV app. The junior developer had the first version done in a day. Impressive. But after testing, they discovered the entire design was wrong. Not the code. The design. The approach. The architecture. So, a senior stepped in, rethought the problem, simplified the data model, eliminated unnecessary complexity, and the junior rebuilt it in a few hours with the new design.
The second version was not just faster. It was fundamentally better. And the reason it was better had nothing to do with AI. It had everything to do with someone who had seen enough implementations to know what "simple" actually looks like.
That is the gap.
What AI Cannot Replace (Yet)
I have been a Technology Consultant long enough to know that the most expensive problems in any project are never about writing code. They are about making the wrong decisions early and discovering it late. It is one of the reasons every project and product at Equerra is led by a principal consultant. When you are designing solutions for complex businesses, particularly in industries like food and beverage where traceability, compliance, and integration all intersect, the cost of a bad architectural decision is enormous. You need experienced people making those calls from day one.
Which API approach do you use when Business Central gives you three options and all of them have limitations? When should you push logic into Azure Functions versus keeping it in AL? When does a "quick fix" create technical debt that will cost you ten times more in two years?
These are not coding questions. They are judgment calls. And they come from doing the work, getting burned, recovering, and building a mental library of "I have seen what happens when you go down that path."
AI can generate a solution. It cannot yet tell you whether that solution is wise.
The Real Shift Is Not Code. It Is Role.
Here is what I think is actually happening. The value chain in our industry is being compressed. The mechanical parts of building software, typing code, wiring up pages, writing boilerplate, those are collapsing toward zero cost. What remains expensive is the thinking that happens before and after the code.
Before: What should we build? Why? What are the trade-offs? What will break?
After: Does this actually work in production? Does it scale? Does it fit the customer's real process, not just the one they described in the requirements doc?
If your identity as a professional is "I write code," you are going to have a rough few years. But if your identity is "I make good decisions about how systems should work," then you are more valuable than ever. Because there is now more code being generated that needs someone to evaluate whether it should exist at all.
The Junior-Senior Paradox
Something interesting is emerging. Juniors are adopting AI faster than seniors. That makes sense. They have less to unlearn. The workflow shift from "I type everything" to "I direct an AI and review the output" is smaller for someone who never got deeply attached to the old way.
But here is the paradox. Juniors who skip the foundational learning, who never understand why the code works, end up producing more output with less understanding. They are fast but fragile. The moment something goes wrong in a way the AI has not seen before, they are stuck.
Seniors who resist AI entirely are slow but increasingly irrelevant.
The sweet spot is the experienced professional who embraces AI as an amplifier for their judgment.
Not a replacement for it.
So, What Are You for Now?
You are the person who knows the difference between a solution that works and a solution that lasts. You are the person who can look at an AI-generated architecture diagram and spot the three assumptions that will blow up in production. You are the person who has enough scar tissue from past projects to know that the "elegant" approach is sometimes the dangerous one.
In a world where everyone can build, the person who knows what to build and what not to build becomes the most important person in the room.
That is what you are for.
Regards,
Tharanga Chandrasekara
