The Post-Pilot Reality: Why Your AI Investment Is Not Scaling
The conversation around artificial intelligence in the C-suite has officially shifted. We are no longer debating whether to adopt AI. We are now confronting the uncomfortable reality of what happens when we do.
The data published this week paints a stark picture of an enterprise landscape that has rushed into deployment without the infrastructure to support it. According to KPMG’s Q1 2026 AI Quarterly Pulse, AI agent deployment has jumped from 11% to 54% of enterprises in just two years. BCG’s AI Radar confirms that 65% of CEOs now rank accelerating AI among their top three priorities.
Yet, despite this surge in investment and deployment, the returns are not materialising. BCG’s research reveals that 74% of companies are still struggling to achieve ROI from their AI programmes, with only 26% able to scale AI effectively. McKinsey’s latest State of AI survey echoes this, finding that nearly two-thirds of organisations have not yet begun scaling AI across the enterprise, and only 39% report any EBIT impact at the enterprise level.
The gap between deployment and value is widening.
Adoption is no longer the hard part. Getting a return on it is. And most C-suite leaders I speak to know it, even if they are not saying it out loud yet.
The Agentic Web: Why Your Next Hire Might Need Their Own Browser
Nvidia’s CEO Jensen Huang recently painted a picture of the future that should make every C-suite leader pause. He estimates that his company’s 75,000 employees will eventually work alongside 7.5 million AI agents. That is a staggering ratio of 100 to 1. š«¢
He is not alone in this vision. McKinsey found that 62% of organisations are already experimenting with AI agents. But there is one major issue that most leaders are missing: these new-age workers are not designed to work on our internet.
Right now, many agents are hitching a ride on their owners’ email, computer, or Slack channel. They are guests in a human-centric digital world. But for the industry to mature and for these agents to actually deliver the autonomous value we expect, they will eventually need their own tools, credentials, and compute.
They need an internet built for them.
Building the Agent Layer
The global AI agent market was valued at just $5.43 billion in 2024, but Precedence Research expects it to reach $236 billion by 2034, a compound annual growth rate of roughly 46%! That kind of growth requires serious infrastructure, and dozens of startups are already hard at work building it.
Take AgentMail, which recently raised $6 million to give AI agents their own email inboxes, allowing them to send, receive, and parse emails autonomously. Or AgentPhone, which hands them a phone number to call and text from. Platforms like Kite and Sponge are letting agents make secure payments with their own wallets. And Browserbase, recently valued at $300 million, gives them a headless browser to freely roam the web, read data, and execute tasks without human intervention.
Even major institutions are getting involved. MIT’s Media Lab is currently developing Project NANDA, an ambitious initiative to build the foundational infrastructure for a true ‘Internet of AI Agents’. Just as the early web required DNS and HTTP protocols to scale, NANDA is creating the essential building blocks to support trillions of autonomous AI agents that can discover, verify, and collaborate in real-time across decentralised networks.
The Agentic Web is Here
Emails. Phone numbers. Wallets. Browsers. Decentralised registries. Stitch these pieces together and you start to see the early scaffolding for an agent-native internet. Leading researchers call it the Agentic Web, and most people have not noticed it is under construction.
For leaders in the knowledge economy, the implications are profound. We are moving from a world where we use AI as a tool, to a world where we manage AI as a workforce.
If your organisation is still treating AI as a glorified search engine or a writing assistant, you are preparing for the wrong future. The real competitive advantage will belong to the organisations that understand how to integrate, govern, and deploy autonomous agents at scale.
The infrastructure is being built right now. The question is whether your leadership team is READY to use it.
This piece first appeared in Talent Pools Nexus, our weekly read for leaders in research, insights and consulting. Subscribe on LinkedIn.
