Why Digital Workers (AI Agents) Will Soon Be Your Newest (and Fastest-Scaling) Hires
And why you should think of them more as people than as traditional software!
Most leaders still treat AI like software. They fund it like a tech project. They manage it like infrastructure. Read on for a summary of what you need to know, but checkout my full whitepaper (paywall link below) for several industry examples and lessons learned for technology and HR partners!
Thinking of AI agents as software that needs to be perfect when your human workforce isn't even close to perfect is the wrong way to really get value out of your investment.
Lately I've been discussing the concept a lot that AI agents — especially those built on LLMs — should be treated like a new class of worker: digital colleagues. These aren’t traditional programs. They’re knowledge workers in disguise. They learn, adapt, get coached, make mistakes, improve, and (here’s the kicker) can be replicated instantly once they get good. And the real deal is that they do these things more like people than like programs meaning the traditional ways of programming and "fixing bugs" are soon to be out the window.
I’m seeing a shift emerge: Forward-thinking companies are treating Agentic AI like people — onboarding, training, and even managing them like junior associates. The results? Big impact, fast.
Here’s what you need to know:
AI Agents Aren’t Software — They’re Your New Hires. They require onboarding, context, and coaching — just like a new team member...and this means your code of conduct, your position on ethics, your framework for achieving business results responsibly, etc.
They Learn Faster Than Humans. Humans might take 6–12 months to ramp. AI agents? A few weeks to a couple months with the right supervision. GIC’s insurance bot went from 19% accuracy to 95% in 60 days.
They’re Not Perfect. (Neither Are Humans.) The risk isn’t that AI gets it wrong. It’s that when it does, it does it at scale. Build guardrails, not illusions. Focus on monitoring over delaying role out until the results are perfect.
They Get Better — and They Don’t Forget. Every lesson learned becomes a permanent upgrade — instantly replicable across your org. Let's be honest, that's not the same when training new associates.
They Find Faster Ways to Do the Work. In one case, an AI-driven operations team doubled profitability by optimizing processes humans hadn’t even considered.
The Opportunity: We’re not deploying AI tools. We’re scaling a new kind of workforce. That means we need a new leadership playbook — one built on hybrid teams, AI coaching, ethical oversight, and constant iteration.
Companies like Lemonade, Klarna, and Morgan Stanley are already there. Their AI agents aren’t experiments; they’re outperforming humans in speed, accuracy, and cost — and improving daily.
What You Can Do Now:
Start treating AI onboarding like employee onboarding. Assign managers and track performance.
Design hybrid workflows: AI handles volume, humans handle edge cases and strategy.
Create new roles: AI trainers, coaches, and product owners who manage digital labor.
Build in governance. Errors scale. So should accountability.
📌THE POINT IS: The next generation of operational scale won’t come from hiring more people. It’ll come from managing a fleet of high-performing, self-improving AI agents — like you would your best employees.
And when they’re ready, you don’t hire one more — you clone them.
Let’s stop thinking that AI needs to be perfect before it can be deployed. The idea is to mix in AI Agents with new human hires to start training them side-by-side. Soon you'll see the AI Agents outperform your human workforce on operational tasks while opening the door for new human roles such as AI monitors, AI coaches, AI manager, etc. Let’s start thinking like workforce architects.