Why one COO refuses to cut benefits to fund AI spending

“Software is another area where if you’re spending money on AI, companies should really be assessing: do we really need these legacy SaaS (software-as-a-service) platforms, or can we now build our own CRM (customer relationship management) tools, our own LMS (learning management system) tools?”

She also pointed to token usage as an underappreciated budget lever. Many AI platforms charge on a per-user basis, but additional features unlock consumption-based billing, and costs can escalate quickly without controls in place. Leadership, she said, has a responsibility to put token budgets in place on a per-team basis, using the controls available on the back end of most platforms.

“A lot of companies, because this maybe wasn’t on their radar last year, didn’t budget for 2026. Hence why they’re potentially scrambling and pulling from other places. It should not be coming from people’s compensation. That is sending totally the wrong message,” she said.

Building an AI-ready workforce

At Moxie, Connor has spent the past year rolling out AI tools across the agency. Her approach is notable for its deliberateness. Before selecting any platform, the team developed a comprehensive AI policy and security review checklist, consulting with chief technology officers and external advisors. No tool gets onboarded without passing a data security and privacy audit. It’s a level of rigor that many organizations are still catching up to, with studies finding that up to a third of employees use AI tools without IT oversight.

The training approach is just as deliberate. Research has found that poorly implemented AI rollouts can drive employees to quit, making the how of implementation just as important as the what. Rather than PowerPoint walkthroughs, Connor runs sessions where every employee shows up with their laptop open and works through the platforms in real time, building workflows and configuring personal settings as a group. The result, she says, has been unusually low resistance.