What 101 Career Changers Taught Us About AI-Powered Guidance

For six months, we ran WAI inside Orange Digital Center's reskilling programs. Three cohorts (AI Essentials, Data Analytics, and a broader community of ODC alumni) used the platform alongside their instructor-led training. We expected to learn about activation rates and engagement metrics. We ended up learning more about people.

The confidence gap is real

The Data Analytics cohort surprised us. These were participants with the most concrete technical skills. They'd completed rigorous training in Python, SQL, Power BI. They knew their stuff. And yet, the recurring theme in their conversations wasn't "what tools should I learn next?" It was imposter syndrome. How do I position myself when I don't have formal experience? Am I good enough to apply for this role? Will they take me seriously?

The platform became a space to work through these questions without judgment. Not because AI is better than humans at this (it's not) but because it's available at 2 AM when the doubt kicks in, and it doesn't get tired of hearing the same fear rephrased five different ways.

Two realities in the same room

The cohorts revealed something we keep seeing: people in reskilling programs aren't one population. They're at least two.

About 43% were employed professionals looking for advancement or a pivot into higher-potential fields. They had jobs, stability, and a clear sense that they needed to level up. Their questions centered on strategic positioning: how do I translate what I already know into a new domain?

The other 45% were in active transition: unemployed, returning from career breaks, or just starting out. Their questions were different. More urgent. Less about optimization, more about viability. Can I actually do this? Will anyone hire me without traditional credentials?

Same classroom. Same curriculum. Completely different internal conversations. A platform that adapts to both can surface these differences in ways that help facilitators intervene more precisely.

The gap between interest and action

Across all cohorts, we noticed a pattern: participants were strategically aware. They understood that AI and data skills matter for the future. They'd signed up for the courses, done the modules, engaged with the content. But when it came to translating that into concrete next steps (building a portfolio, applying for roles, articulating their value) many were stuck.

There's a tension between knowing what matters and knowing what to do about it. Technical training addresses the first part. It often leaves the second part to chance.

The finding that changed how we think about deployment

Here's the number that matters most to us: 44% of participants who received only a digital invitation became active users. That's solid for an AI platform. Industry benchmarks for sustained engagement in digital coaching tools typically land between 30-50%.

But participants who got a 20-60 minute in-person introduction during their group sessions? 78% activation.

Same platform. Same features. Nearly double the engagement.

This isn't a story about AI replacing facilitators. It's a story about AI extending them. A short human touchpoint at the beginning (explaining the context, demonstrating the value, answering the first few questions) unlocks sustained autonomous usage afterward. The platform carries the load between sessions. But the human opens the door.

For anyone designing workforce development programs: don't skip the handoff. It's not overhead. It's infrastructure.

What we're taking forward

ODC taught us that career guidance at scale isn't just about making a good product and hoping people use it. It's about understanding where people actually get stuck, which is often not where you'd expect. The Data Analytics cohort didn't need more technical content. They needed someone (or something) to help them believe their skills were valid.

We're continuing the partnership into 2026, with an expanded approach: WAI for deep coaching in longer programs, ZVAC for rapid vocational profiling in shorter interventions, and VERTA for AI literacy across the board. Different tools for different moments in the journey.

But the core lesson stays the same: technology works best when it meets people where they are. And where they are is usually more complicated than a skills gap.