 |
| (Image credit: Future) |
At its recent ONE conference, low-code leader OutSystems launched a new platform called Agent Workbench, designed to help enterprises build and manage AI agents more reliably. But in the same breath, CEO Woodson Martin cautioned that much of the AI hype we see today is exaggerated.
In an interview with TechRadar Pro, Martin remarked, “I have not met a single CIO who is saying, my goal is to replace all of my existing systems with something new built by an AI.” He emphasized that the real value is in layering AI over existing systems not tearing everything down.
Hype vs. Practicality
Martin argues that the idea of a complete AI revolution that all legacy systems will vanish overnight is unrealistic. Instead, he sees transformation happening in incremental layers, applying AI to enhance rather than replace. “The revolution is a little oversold right now,” he says, but he still sees meaningful change ahead.
He points out that under the surface, a lot of work still goes into infrastructure, data pipelines, APIs, integration, and governance. AI may appear magical on top, but the foundation must be solid.
OutSystems Bet on Agentic AI in Enterprise
While cautioning against overhyping AI, OutSystems is not stepping back from innovation. With Agent Workbench, the company aims to help organizations go beyond pilots and embed agentic intelligence into real workflows.
The platform includes features like an agent marketplace, integration with enterprise data systems, and support for multiple large language models. It’s built to fit inside existing software landscapes instead of forcing total rewrites.
The idea is to let organizations spin up intelligent agents to handle tasks like automating customer support, classifying data, or bridging disconnected systems without disrupting core operations.
Why Martin Pushes Back on AI Extremes
Martin sees three main risks in overhyping AI:
- Unrealistic expectations: Businesses might expect revolutionary change overnight, only to be disappointed by incremental results.
- Neglect of foundation: Without robust data, architecture, and governance, AI systems can underdeliver or introduce risk.
- AI as silver bullet myth: Thinking AI fixes everything can lead to misallocation of resources and missed priorities.
Instead, he argues for a measured, layered adoption of AI one that respects what already exists and builds responsibly on top of it.
Industry Response and Mixed Views
Not everyone agrees with Martin’s caution. Some industry leaders continue to envision more radical transformations driven by AI, especially in sectors like autonomous systems, life sciences, and robotics.
But even in those domains, many accept that breakthroughs often come from combining new AI capabilities with existing assets, rather than wholesale disruption.
Where Agentic AI Might Make a Real Difference
Here are some areas where a platform like Agent Workbench could deliver meaningful value:
- Business process automation: Automating repetitive tasks across systems.
- Data orchestration: Agents that connect siloed systems, pulling together insights.
- Customer support augmentation: Intelligent triage or routing assistants working alongside human teams.
- Workflow enhancement: Agents that suggest next steps or trigger processes in apps.
Because these use cases don’t require replacing entire systems, they align more closely with Martin’s vision of practical AI layering.
Balancing Optimism with Realism
Martin’s message is not anti-AI it’s pro-responsible AI. He believes that hype cycles distort expectations and delay adoption. But neither does he dismiss AI; rather, he positions it as part of the future, not the whole future.
In that sense, OutSystems’ launch of Agent Workbench offers a tangible path: enterprise-focused, governed, and designed to work with what’s already built.
As AI continues to evolve, the biggest winners may be those who take a sensible, incremental approach rather than chasing grandiose claims.
For more on how AI and enterprise strategy intersect, check my post on how organizations can adopt AI sustainably.