Stop Pitting RPA Against AI Agents—An Ethics of Hybrid Automation

Here's a contrarian take: the industry's obsession with replacing robotic process automation (RPA) with AI agents is not just technically misguided—it's ethically lazy. We keep hearing that agents are the future and RPA is legacy. But if you care about accountability, auditability, and human oversight, the smartest move is to layer agents on top of RPA, not swap one for the other. That's the architecture that keeps humans in control while reaping the benefits of both.

The False Choice Between Thinking and Doing

RPA excels at execution. It follows predefined, rule-based scripts to perform repetitive tasks like copying data between applications and filling forms (Robotic process automation, Wikipedia). It's deterministic but brittle. AI agents, on the other hand, use large language models to understand natural language, reason, plan tasks, and call tools (Intelligent agent, Wikipedia). They handle variability but introduce uncertainty. The common advice is to choose based on your use case. But that framing misses the point.

The dominant emerging architecture is hybrid: layer AI agents for reasoning over an RPA execution layer (Robotic process automation, Wikipedia). Agents don't replace robots; they drive them. UiPath describes this as Phase 3 of RPA's evolution—'Agentic Automation'—where RPA serves as the execution layer that turns the plans and reasoning of AI agents into tangible actions (UiPath, RPA). This isn't just technical jargon; it's the foundation for ethical automation.

Why Hybrid Is the Ethical Choice

Ethics in automation boils down to accountability and control. RPA enhances compliance by enforcing process consistency and providing audit trails (UiPath, RPA). That's a feature we should not discard. AI agents, left to their own devices, can make opaque decisions. But when agents reason and plan while RPA executes, you get a clear separation: the agent can suggest a course of action, but the robot performs the step in a traceable way. This hybrid approach also supports human-in-the-loop checkpoints, an effective automation pattern (Robotic process automation, Wikipedia). You can build in approval steps where a human reviews an agent's plan before the robot executes it.

Consider the example from IBM: a multi-agent legal research assistant routed queries through a low-cost classifier first, escalating only complex cases—cutting contract review time from 90 minutes to 45 minutes (IBM, AI agents). Now imagine that same system using RPA to actually file the documents or update the case management system. The agent handles the judgment; the robot handles the drudgery. You get speed and accountability.

Addressing the Counter-Argument

Skeptics will say: why add the complexity of a hybrid system when agents can do everything? They point to agentic AI's ability to create subtasks, plan, and self-correct without human intervention (IBM, AI agents). They argue that autonomous agents are the future, and RPA is just legacy baggage. But that argument ignores a critical reality: AI agents require goals and predefined rules defined by humans (IBM, AI agents). They are not fully autonomous in a moral sense. And organizational change and human oversight, not technology, are the biggest barriers to AI automation adoption (Robotic process automation, Wikipedia). A hybrid approach makes oversight easier, not harder. You can always see what the robot did, and you can always interrupt the agent's plan. Pure agentic systems can be a black box, which is ethically problematic.

The Market Is Already Moving This Way—And So Should You

The market data supports the hybrid model. The global RPA market was estimated at $4.68 billion in 2025 and is projected to reach $35.84 billion by 2033 (Grand View Research, RPA market). That's not a dying market; it's a growing execution layer. Software held the largest share of the RPA market at over 70% in 2025, with cloud deployment over 55% (Grand View Research, RPA market). Cloud-native robots and intelligent orchestration for complex agentic processes are already available (UiPath, RPA). The autonomous enterprise market is projected to reach $114.0 billion by 2029 (MarketsandMarkets, RPA market). This is the direction of travel.

So here's my recommendation: don't rip out your RPA bots. Instead, start layering AI agents on top of them. Use agents for the thinking—routing, triaging, planning—and use RPA for the doing—executing the steps in a deterministic, auditable way. Build human-in-the-loop checkpoints into the workflow. This gives you the best of both worlds: the adaptability of AI and the accountability of RPA. It's the ethical choice because it keeps humans in control.

Bottom Line

The single best move is to adopt a hybrid architecture that layers AI agents over RPA, using agents for reasoning and robots for execution, with human oversight built in. That's how you scale automation without losing accountability.

Sources

• Robotic process automation (Wikipedia) - https://en.wikipedia.org/wiki/Robotic_process_automation
• UiPath (RPA) - https://www.uipath.com/rpa/robotic-process-automation
• IBM (AI agents) - https://www.ibm.com/think/topics/ai-agents
• Grand View Research (RPA market) - https://www.grandviewresearch.com/industry-analysis/robotic-process-automation-rpa-market
• MarketsandMarkets (RPA market) - https://www.marketsandmarkets.com/Market-Reports/robotic-process-automation-market-104435980.html