The Rise of the Agentic Hub

Plus: MIT’s warning on agent drift, GitHub’s control plane, real-world agents at scale, the move beyond single-LLM systems, and more...

In partnership with

Edition 162 | March 2, 2026

AI agents are good enough that we’re now debugging humans for authenticity.

Welcome back to Building AI Agents, your biweekly guide to everything new in the field of agentic AI!

In today’s issue…

  • OpenAI and Anthropic turn SaaS into an agent hub

  • MIT finds AI agents regularly break their rules

  • GitHub launches enterprise agent controls

  • Real-world agents: Walmart, Burger King, EvenUp

  • Why top agents won’t rely on one LLM

…and more

🔍 SPOTLIGHT

Nano Banana 2 | Building AI Agents

The Rise of the Agentic Hub

Over the last few weeks, both OpenAI and Anthropic have launched something that sounds mundane, even boring, but is quietly radical: platforms that string your existing SaaS tools together under one agentic interface. OpenAI calls theirs Frontier, a top-down enterprise play focused on agent governance and fleet management. Anthropic's Cowork comes at it from the opposite direction, starting as a productivity tool that just graduated into a full, enterprise-grade product. They're solving different problems, but the destination looks the same.

What makes Cowork's launch so telling is how ready the ecosystem already was on day one. Names like DocuSign, major CRMs, and popular productivity suites showed up as official connectors at launch, clearly the result of coordinated, behind-the-scenes partnerships rather than a spontaneous land grab. The platform was prepared, documentation and all, before the market fully grasped what it was for.

So is this the moment SaaS finally dies? Are the margins of every tool company about to collapse under the pressure of these new, agentic layers? That's the breathless take from certain corners of the internet, but the reality, as it tends to be, is more complex.

What these platforms do is sit on top of your existing tools, linking your Google Docs, your Gmail, your CRM, and your DocuSign into one surface that an agent can act on. You give it a multi-step plan rather than a single prompt, and it carries that plan out on its own, pulling from and pushing to every tool you've wired in. Think of it as how a sharp, competent operations manager works across a dozen tabs at once.

The business effects split in two directions. SaaS company multiples will likely compress, because when an agentic layer can replicate what a lightweight CRM, a basic project tracker, or a simple analytics dashboard does, the moat around those tools gets shallower. But the push toward integrations is also picking up speed. Every SaaS company with survival instincts is racing to provide API support, or MCP access, for these hubs before it gets left behind.

The opportunity for builders is concrete. OpenAI formed "Frontier Alliances" with Accenture, BCG, Capgemini, and McKinsey, dedicated practice groups built around helping enterprises deploy the platform. If the Big 4 consulting firms are investing in agentic platform integration as a service line, the same model scales down to agencies and consultants serving mid-market companies that need help configuring connectors and designing workflows.

For small and midsize businesses, this is where the real leverage sits. A 30-person company cannot afford to build custom agent systems from scratch, hire the engineers, stand up the data pipelines, and iterate for months. But that same company can plug its tools, its Gmail, its CRM, its DocuSign, into Cowork in days and walk away with an agentic layer handling document routing, admin work, and the internal tasks that used to eat entire roles.

These platforms do have clear limits, naturally. If you need a custom phone agent handling sensitive customer calls, or a tightly controlled tool built for regulatory needs, the generic agentic layer is not going to work for you today. The sweet spot is internal: the admin tasks, the cross-tool coordination, the information pulling that nobody ever had time to fix.

Think of these platforms not as the death of SaaS but as a new kind of hub for the agentic era, one that doesn't replace your tools but finally makes them work together.

The SaaS stack isn't dying. It's just getting a brain, finally.

As always, keep learning and building!

—AP

Subscribe to keep reading

This content is completely free, but you must be subscribed to Building AI Agents to continue reading.

I consent to receive newsletters via email. Terms of use and Privacy policy.

Already a subscriber?Sign in.Not now