Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

Generative AI grows up despite doubts

Learn why the hype might just be real.

[Photo: Getty Images]

Profile photo of Eric Barroca

CEO · Vertesia

ChatGPT’s burst onto the scene sparked one of the fastest technology adoption waves in history. Within months, enterprises experimented with generative AI for everything from marketing copy to coding. Investment surged, pilots multiplied, and the rhetoric was breathless. 

Now, just as quickly, headlines have shifted. Gartner’s most recent Hype Cycle places generative AI sliding toward the “trough of disillusionment.” Skeptics point to failed pilots, thin ROI, and fears of hallucinations. 

But let’s be clear: generative AI is still moving forward; just in a more advanced, deliberate way, and the real shift ahead isn’t about the models themselves—it’s about how enterprises move from scattered apps and one-off agents to truly agentic enterprises that are governed, orchestrated, and embedded into the business fabric. 

A familiar pattern 

Every major technology shift follows a similar arc. There’s a wave of excitement, a period of disillusionment, followed by the steady climb toward maturity, productivity, and value.

  • Cloud began with simple prototypes, then faced challenges in security, governance, and cost management—before becoming the backbone of modern IT. 
  • Mobile didn’t reshape industries until app ecosystems and cloud backends made it practical at scale.
  • Blockchain sparked massive hype, followed by disillusionment, and is now quietly being integrated where it makes sense. 

Generative AI is in that same moment: Not fading away, but entering the realism phase where deployments matter more than demos. The key difference this time? Enterprises that build systems of agents that are orchestrated, observable, and well governed—to move faster through the disillusionment dip and closer to true ROI. 

The real issue isn’t the models 

The gap between promise and payoff isn’t about the models themselves. The models are advancing rapidly, costs are dropping, and capabilities are undeniable. The real challenge lies in how enterprises deploy them. 

Too many are stuck in pilot purgatory—proofs of concept that showcase potential but never reach ROI because they don’t address the harder questions: 

  • Who governs access and permissions?
  • How do we ensure accuracy and avoid hallucinations?
  • How are AI solutions integrated with enterprise systems?
  • Where are the audit trails, observability, and security guardrails? 

Without these answers, projects stall as isolated apps and one-off agents rather than scaling into true enterprise infrastructure. 

Underhyped, not overhyped 

If anything, generative AI’s potential has been underhyped due to misplaced expectations. If you expect overnight disruption, you’ll be disappointed 

Generative AI’s long-term impact is on par with the internet—a technology requiring years of groundwork before transforming industries. If you understand that the next phase is about building orchestrated, governed, and embedded systems and becoming truly AI-first, then you see the real story is progressing. 

Agentic systems unlock ROI: Agents with memory to learn over time, access controls to operate within rules, connectivity to the right data and tools, and orchestration to coordinate across workflows. Until organizations make that shift, AI will remain a headline rather than a business benefit. The compounding value begins when AI agents stop working in isolation and start working together.

CTOs must clear the runway 

CTOs shouldn’t mistake this realism. It’s a moment to double down on the foundations that will determine who extracts lasting value from AI. That means: 

  • Structured content pipelines: Models get clean, usable inputs.
  • Enterprise orchestration: Manage multi-agent workflows and enforce permissions.
  • Governance and observability: Outputs can be trusted and audited.
  • Integration into core systems: CRM, ERP, and HR platform integrations ensure that AI augments existing workflows rather than creating new silos. 

These are the prerequisites for scale, and that trajectory mirrors the rise of cloud platforms, which unlocked enterprise adoption by providing governance, scalability, and trust. 

AI grows up  

It may feel like cooling hype, but the AI story is growing in maturity. Flashy early demos have given way to the methodical work of building reliable, governed systems capable of enterprise demands. 

For CTOs, this phase is an opportunity. Organizations that recognize AI’s growing-up moment and build on platforms designed for orchestration, governance, and integration will be positioned to capture lasting value. 

After all, generative AI is just learning how to walk steadily—before it runs. 
 
Eric Barroca is the cofounder and CEO of Vertesia.