The agent proliferation problem

You've bought the licenses. You've made the announcement. And a few months later, you're staring at a workforce either ignoring the tools entirely or churning out AI-generated fluff that nobody's reading. The ROI? Nowhere to be found. 

The problem isn't AI. It's how organisations are rolling it out. 

In this How I AI episode, Neo and I share a recording from a live webinar we ran on why AI rollouts fail and what it actually takes to set yours up for success. We cover the most common mistakes we're seeing on the ground, the research that should change how you think about training, and the question every leadership team needs to answer before spending another dollar on licenses. 

Neo and I cover: 

• Why an unclear AI strategy at the leadership level creates confusion across teams, with some departments using the tools freely and others too uncertain to try. 

• The leadership credibility problem: when CEOs and senior leaders don't use AI themselves, the rest of the organisation takes it as a signal that they shouldn't either. 

• Why buying licenses without training can actually reduce productivity, producing a flood of low-quality, AI-generated work that slows people down rather than speeding them up. 

• The surprising finding from HBR research that the average time saving from AI is just 2.5%, and why vanity metrics like daily logins and token usage tell you almost nothing about real value. 

• The agent proliferation trap: why having everyone in your organisation building agents independently leads to duplicated, inconsistent, and unowned tools that create more confusion than clarity. 

• The cultural barrier of AI fear, and why people won't genuinely adopt a tool they believe is there to replace them. 

• AI brain fry, a distinct form of cognitive fatigue caused by constant context-switching and vigilance when managing multiple AI tasks simultaneously. 

• Why training to tasks rather than features is what actually shifts behaviour, and the research-backed minimum dose of five hours of hands-on training required to see any meaningful benefit. 

• The case for training close to license rollout to prevent bad habits forming before people know how to use the tools well. 

• A three-stage framework for AI adoption: access, literacy, and leverage (individual then organisational), with literacy positioned as an ongoing capability rather than a one-off event. 

• Why planning where time savings will actually go matters as much as achieving them, otherwise people simply fill the space with more work. 

Connect with Neo Aplin on LinkedIn (https://www.linkedin.com/in/neoaplin/) and via inventium.ai (https://inventium.ai), where he leads Inventium's AI training and upskilling work with organisations and teams. 

My latest book The Energy Game is out on July 7, 2026. You can order a copy here: https://amzn.to/48ID29M  

Connect with me on the socials: Linkedin (https://www.linkedin.com/in/amanthaimber)  

Instagram (https://www.instagram.com/amanthai)  

If you are looking for more tips to improve the way you work and live, I write a weekly newsletter where I share practical and simple to apply tips to improve your life. You can sign up for that at https://amantha.substack.com/  

Visit https://www.amantha.com/podcast for full show notes from all episodes.  

Get in touch at amantha@inventium.com.au 

Credits:  

Host: Amantha Imber 

Sound Engineer: Martin Imber 

See omnystudio.com/listener for privacy information.