I didn't found TDI
because AI became
interesting.

The origin of the insight

Skills were the answer before I knew I was asking the question.

My career started at SAP in 2006, in a role nobody had done before. The inside sales team was newly formed. Leaders knew they needed enablement — but to them, that meant product training. Get the sellers up to speed on what we sell, send them to the phones.

I saw it differently.

I built a call quality program around the full set of skills those inside sales people actually needed to succeed: product knowledge, yes, but also how to run the call itself, how to handle the conversation, how to perform. I used that same skills framework to develop a benchmark assessment — so when new cohorts were hired, we knew exactly where each group started and what they needed to learn. Call quality improved by 22%. Time-to-productivity dropped. The global sales team picked up the concept and ran with it. Services built the same model for all global consultants. Within a few years, every function at SAP had its own learning organization built on those principles.

I didn't set out to build a skills architecture. I built a program that worked — and the reason it worked was skills.

That insight has driven everything since.

When I led the Women's LEAP program years later, I was operating at a completely different scale — global leadership pipelines, succession strategy. The program overachieved its 25% women-in-leadership target, hitting 27% within the first year. SPARK, the MOOC I built as the foundation of the redesigned program, won a Brandon Hall Bronze Award for Best Unique or Innovative Learning and Development Program. I'm proud of what it accomplished.

But here's what I couldn't do: I couldn't surface skills data on those women. Not because I didn't want to — because I didn't have it. There was no skills layer to draw from. Sponsorship was asynchronous. Mentoring was intentional but manual. We found ways to surface capability the hard way, and it worked.

What I know now is that it could have worked so much better. A participant with a skills profile, with mapped adjacencies, with visible gaps that could be addressed — that would have been a fundamentally different experience for those women. More powerful. More portable. More real.

The alternative isn't just kinder. It's smarter. When skills data is real and current, organizations can move people proactively — preserve institutional knowledge, fill emerging gaps, redeploy before the crisis hits. Headcount stops being a secret currency when the skills are visible to everyone.

Cisco added operational fluency — a global team and the discipline of running a large learning function at scale. It reinforced what I already believed: programs fail when they're disconnected from the skills reality underneath them.

Founding TDI

The gap between what talent functions could be and what they actually are is still, in most organizations, enormous.

Not for lack of effort. For lack of the right infrastructure.

The system I've been building toward for two decades is CapabilityOS™ — fifteen components across five groups that individually deliver value and together compound it. Skills intelligence as the red thread. Workforce planning and talent supply built on top of it. Change architecture designed in from the start. Performance and development connected to real outcomes. Talent pipelines that are visible and trusted. And an HR technology and AI intelligence layer that makes the whole operating system run at the speed and scale modern organizations require.

For most of my career, building and sustaining that system at enterprise scale required more human capacity than most organizations could dedicate to it. AI changes that equation. Skills intelligence at scale requires AI to manage it — and AI transformation requires skills intelligence to succeed. The two have always belonged together. Now they can actually work together.

TDI exists to build that system for organizations that are ready to stop managing talent by instinct — and for organizations that just need one piece of it fixed right now. Both are the right starting point.