The AI Reckoning

You’ve heard it on the golf course, in team meetings, and all over LinkedIn: adopt AI or get left behind. The messaging seems relentless. The urgency feels manufactured. But the tools themselves? They rarely deliver on their promises for businesses. That’s the part nobody says out loud. It’s Zeno’s paradox writ large: you keep closing half the distance to the target, but you never quite arrive. And the data is beginning to confirm it.

AI initiatives often promise to reduce overhead while increasing output, but many businesses are experiencing the opposite. Forrester’s 2026 Future of Work report found that 55% of employers regret replacing workers with AI. Major brands like IBM, Salesforce, and Google have been adding people back in redefined roles, and one in three companies that restaffed ended up spending more on rehiring than they saved from the original cuts.

It turns out, we need human creativity and thinking in our business workflows, where accountability, decision-making, and trust are required. Some aspects of work should not be automated or outsourced to robots.

This is often where the breakdown happens. Last year, MIT reported that U.S. businesses have collectively poured $35–$40 billion into AI initiatives, and 95% have seen zero measurable return on that investment. The 5% of businesses seeing strong results from AI initiatives? They succeed by targeting a single, specific pain point, not by trying to transform everything at once.

They also applied AI to the right part of their work processes. Most AI tools work best for administrative tasks and repetitive functions, but “more than half of the funds spent on AI projects tried to use the technology for sales and marketing, two areas that the researchers say still need human involvement and have a lower ROI [with AI tools].” Sales and marketing are still relationship-driven parts of the business, and they need people who understand people to drive those initiatives. That doesn’t mean AI can’t be used, but if businesses expect to transfer these roles to AI agents, they might not see the returns they expect.

At Fidelis, AI is integrated into many of our tools and workflows, but we maintain a human-robot-human process that maintains the integrity of our work and the relational trust with our clients.

A few questions we ask ourselves as a team about AI:

• Are we chasing AI adoption for results or fear of being left behind?
• Where on our team is AI actually creating leverage and supporting our goals, and where is it creating a false sense of progress?
• What would it look like to pair real human expertise with the right tools, instead of solely relying on one or the other?
• Are we measuring output, or just activity?
• Is our workflow with AI still aligned with our business goals and core values?