The Gap Nobody's Talking About

Here's a number that should concern every engineering leader: 80% of your people want to use AI at work. But only 12% received any kind of formal training in 2024. And while executives talk about building "AI-native organizations," most haven't figured out what that actually means or how to get there. The result is millions in potential productivity gains left on the table while competitors quietly pull ahead.

What You'll Learn

In this article, I'll share the talent transformation approach that's worked across my consulting engagements. You'll get a practical framework for upskilling your existing team without disrupting operations, plus a complete rethink of how to hire in an AI-augmented world. They're patterns I've extracted from working with engineering teams in EdTech, consumer electronics, and e-commerce over the past two years.

The Two Paths (You Need Both)

As a leader, you've got two options for building AI capability. You can reskill your existing talent, which is cheaper and doesn't blow up your organizational headcount. Or you can fight for external talent who already have AI experience, which is expensive and hard to get. Ideally you do both, of course.

Transforming Your Existing Team

Look, most internal AI training fails because it ignores psychology. Your people are feeling two things simultaneously: fear that AI will displace their jobs, and genuine desire to work with it and become more efficient. Generic mandates like "start using ChatGPT every day" don't address either emotion. They just create confusion followed by resistance.

What actually works is the champion model. In a recent engagement with a 200-person engineering organization, we identified 2-3 enthusiastic experimenters in each team. Not managers. Ground-level engineers who were already playing with AI tools on their own time. We gave them three things: dedicated experimentation time (10 hours weekly), permission to fail without consequences (of course within reasonable limits), and a structured way to share learnings with their peers (workshops and dedicated Slack channels).

The results after 90 days: those champions' teams showed 35% higher AI tool adoption than teams that received only top-down training. And the quality of AI usage was dramatically better. People weren't just copying and pasting from ChatGPT. They were building custom workflows, creating internal prompt libraries, and catching inadequate AI output before relying on it.

Peer influence beats top-down mandates every time. When someone on your team shows you a workflow that saved them 3 hours, you pay attention in a way you never would during mandatory training.

But one training cycle isn't enough. The tools change quarterly. The techniques evolve monthly. Your champions need ongoing time allocation and your organization needs continuous learning loops. And as people become more capable with AI, you're raising the bar on what "good" looks like. Someone using AI effectively can deliver what used to take a senior engineer half the time. That's the new baseline.

Hiring Has Fundamentally Changed

When you do hire externally, forget everything you knew about technical interviews.

Traditional hiring is broken. Candidates are using AI on your take-home assignments. And frankly, they should be. That's how real work gets done now. LeetCode-style interviews test skills that matter less and less. Resumes take 10 seconds to generate with AI, so they're nearly meaningless as signals.

What matters now is how candidates use AI to solve problems. Not whether they can write a sorting algorithm from memory.

I've started recommending a different interview approach to my clients. Instead of "implement this data structure," ask candidates to solve a complex, ambiguous problem with AI tools available. Watch their process. Do they know when to trust AI output and when to verify? Can they break down a problem in a way that makes AI coding assistants effective? Do they have systematic workflows, or do they just type into an input field of ChatGPT and hope?

Red flag: "I just use ChatGPT for everything." That tells you they haven't developed judgment about AI limitations.

Green flag: They show you custom tools, explain their verification process, and can articulate why certain approaches work better for different problem types. One candidate I helped evaluate had built a personal toolkit with specialized prompts for debugging, code review, and architecture decisions. She could explain exactly when each tool failed and how she compensated. That's the skill you're hiring for in 2025 and onwards.

You're not looking for narrow specialists anymore. You're looking for T-shaped generalists. Some people call them Product Engineers. I call them people who can learn faster than the technology changes. Because in six months, the specific tools will be different. The meta-skill of rapidly adopting and evaluating new AI capabilities is what matters over time.

Your 90-Day Talent Transformation Plan

Start this week: Identify your potential champions. Look for engineers who are already experimenting, asking questions about AI tools in Slack, or building side projects with Gen AI. You probably know who they are.

Week 2-4: Give those champions formal permission and time. Ten hours weekly minimum. Create a simple structure for them to document what works and what doesn't. Set up a bi-weekly sharing session where they demo to their teams.

Week 5-8: Revise your interview process. Design at least one AI-assisted problem-solving exercise. Train your interviewers on what good AI usage looks like versus superficial usage.

Week 9-12: Measure and adjust. Track adoption rates across teams. Compare champion-led teams against others. Identify what's working and double down.

Series Navigation

Previous:

• Part 1: Why Your 3-Year AI Roadmap Becomes Obsolete in 6 Months
• Part 2: Stop Hiring: Get 3x Output From Your Current Dev Team
• Part 3: The Only Way to Prove Your AI Investment Is Worth It

Next in this series:

• Part 5: From Zero to Production-Ready AI in 6 Months

Let's Connect