GGW Voices is an ongoing collaboration between CDC Gaming and Global Gaming Women featuring commentary and insight from women in the gaming industry.
AI-built products are like a house built overnight. The shell goes up fast, but the wiring and plumbing aren’t done and nothing is connected. It’s a building, not somewhere anyone can live.
Building has become commoditized. AI can develop and ship something fast. What it can’t supply is the human context around it: how a team actually operates, what it will and won’t trust, and what breaks when you drop something new into an existing way of working. Agents are missing that context, and only a person who understands the whole system can provide it.
It’s also why a product with a hundred new features isn’t automatically a better one. If the people on the receiving end can’t understand it or fit it into their day, it doesn’t matter how fast it was built.
This cuts against how most of us were taught to build a career. At school you pick a path. At university you pick a specialty (I chose business, then had to narrow down to marketing). Going deeper has always been the safe advice. But AI can now be a specialist in many areas at once, so depth alone is no longer what sets people apart.
The evidence points the same way. Adoption, not technology, is now the bottleneck: change management and workflow redesign rank above the technology itself as the main constraints. McKinsey has suggested that for every $1 spent on developing and deploying a genAI solution, businesses should spend around $3 on change management. And it pays off. Prosci’s research found that projects with excellent change management are up to seven times more likely to achieve their objectives.
Hiring is moving in the same direction. A Harvard Business School study found that since generative AI’s public launch, U.S. job postings for roles built around structured, repetitive tasks have fallen 13%, while demand for roles involving human-AI collaboration has grown 20%.
I’ve seen this from both sides. I spent years inside CRM teams at Paddy Power and Betfair, so I know where the bottlenecks sit. For the last two years I’ve been at Xtremepush, which works with more than 220 iGaming clients, and I’ve worked closely with several of the biggest.
What I’ve learned is that the steps in a campaign were never the problem. Research, data, creative, compliance, QA and reporting all have to happen. It’s the handoffs between them where the days disappear. That’s what we built XpertOS around: a governed, AI-native operating system for CRM teams, where agents handle the legwork across the whole campaign cycle and marketers move into a strategic oversight role, steering the work and deciding what goes out.
But a big part of my job is the human bit: onboarding clients, helping them understand how it works and showing them where it adds value. I can do that because I’ve been on the receiving end.
That’s why I think the people who matter most over the next few years won’t necessarily be the best builders. They’ll be the ones doing the wiring and plumbing: understanding a whole system well enough to connect new capability so people can actually use and trust it.
Depth still matters, but the edge is knowing how the pieces connect. So if you feel you should be specializing harder, my takeaway is that going broader is a strength.
