Your engineers have the AI tools but they do not have the training: Newsletter #002

Here is something I see in almost every engineering organization I talk to. The AI tools are there. Engineers are generating code, tests, documentation. AI is in the workflow. And yet, when leadership looks at the results, something is off. Velocity has not changed meaningfully. Quality is inconsistent. Some engineers are flying. Others are struggling. And no one can explain why.

The problem is not the tools. The problem is that the team was never trained to use them as a system.

Most engineering teams today are using AI across different stages of their software development life cycle. But without a shared framework, minimum engineering standards, and a structured adoption path, those experiments stay isolated. The tools work. The gains are real. But they do not compound. And when AI enters a system without the right foundation, it does not just fail to help. It scales whatever problems were already there.

I wrote about this in The Cognitive Leader because it is one of the most important lessons we learned during our own transformation:

"There is a difference between handing someone a tool and telling them to adopt it, and showing them how the tool changes the way they should approach their work. The first creates compliance. The second creates transformation."

That is why we built the AI-Ready Engineering Certification. Three levels. From foundational mindset to role-specific mastery. Designed to give every engineer on your team a shared language, a shared standard, and a clear path to grow into AI together, not in isolation.

We are opening limited free early access to this program for readers of this newsletter. If you want your team to stop experimenting in silos and start operating with consistency, this is where it begins.

From Franz Silva, CTO — Teravision Technologies

Prompt engineering is not a trick. It is a discipline.

I want to talk about something I see every day across our engineering teams and across client teams we work with. There is a perception that prompting AI is simple. You type something, you get something back, and you move on. If the output is not great, you try again with different words. Most engineers treat prompts the way they used to treat search queries. Fast, disposable, good enough.

That approach works when the stakes are low. It stops working the moment AI becomes part of how your team actually builds software.

When an engineer writes a prompt without structure, without context, without a clear definition of what they need, they get an output that looks reasonable but often misses the point. And because AI is confident in how it presents its answers, the engineer accepts it. That is where the real risk begins. Not in the tool being wrong, but in the engineer not knowing how to evaluate whether it is right.

Structured prompting changes that entirely. When we trained our teams to use frameworks like A.C.T., which stands for Action, Context, and Target, something shifted. Engineers stopped asking AI to do things and started telling AI how to think about the problem. The difference in output quality was immediate and measurable.

Here is what that looks like in practice. Instead of prompting "write unit tests for this function," an engineer using A.C.T. would prompt something like: "You are a senior QA engineer reviewing a payments module. The function handles currency conversion with rounding rules. Write unit tests that cover edge cases including zero amounts, negative values, and rounding discrepancies across three decimal currencies. Prioritize tests that would catch silent failures in production."

Same tool. Same engineer. Completely different result.

But here is the part that matters most. Structured prompting is not just about getting better outputs. It forces the engineer to think before they prompt. It requires them to understand what they are asking for, why they are asking for it, and what a good answer actually looks like before they see one. That thinking is the discipline. And that discipline is what separates an engineer who uses AI from an engineer who is truly AI-ready.

This is exactly what Level 1 and Level 2 of the AI-Ready Engineering Certification are built around. Not just learning how to prompt, but learning how to think with AI in a way that produces consistent, reliable, and accountable results across the entire team.

If your engineers are still treating prompts like search queries, you are leaving most of the value on the table.