Is AI sounding the end of RPA ?

Since its emergence, Robotic Process Automation (RPA) has been at the core of enterprise digital transformation, enabling the automation of repetitive tasks and freeing up time for higher value activities. However, with the rise of generative and agentic AI, this technology is increasingly being challenged. Is this the end of RPA, or simply the beginning of a new phase in its evolution ? Let us explore this question through recent technological developments and the reflections they raise.

RPA facing new challenges

In its traditional form, RPA relies on bots capable of replicating human actions within IT systems. While this technology has delivered significant productivity gains, it presents several intrinsic limitations :

• Rigidity : RPA scripts are often sensitive to changes in interfaces or processes, requiring frequent and costly updates
• Limited scope : traditional RPA bots are well suited to structured processes but struggle when dealing with unstructured data or tasks requiring decision making

At the same time, new technologies are challenging this approach.

Generative and agentic AI : an alternative or a complement ?

The promise of agentic technologies

Anthropic has attracted significant attention with the "computer use" capability of its Claude 3.5 Sonnet model, enabling AI to interact with desktop applications like a human user. Similarly, solutions such as Runner H from the French startup H embody a new generation of tools.

These agents, powered by Artificial Intelligence, are designed to act autonomously, making decisions and planning actions to achieve defined objectives. Unlike traditional RPA, they offer :

• Robustness to change : the ability to adapt to variations in interfaces and processes, reducing maintenance costs
• Decision making capabilities : they do not simply execute tasks but can analyse complex contexts and make informed decisions
• Flexibility : through the integration of natural language and multimodal models, they can interact seamlessly with both systems and users

‍

A coexistence in the making

While agentic AI technologies appear to offer significant advantages over traditional RPA, it would be premature to ঘোষণা its disappearance. Indeed, agentic AI is still in its early stages, and its large scale adoption faces several challenges.

First, its implementation remains technically complex. Designing such agents requires advanced skills and specialised expertise, which can slow down deployment, particularly in organisations with lower technological maturity.

Costs also represent a major barrier. Companies that have already made significant investments in RPA infrastructures are often reluctant to transition abruptly to new solutions, especially as the return on investment of agentic technologies is not yet fully proven.

Finally, these decision making AI systems raise regulatory and compliance challenges. In highly regulated industries, requirements related to traceability and auditability complicate adoption, as organisations must ensure that these technologies meet strict governance standards.

Towards technological convergence

Market players such as UiPath, Automation Anywhere and ServiceNow do not view these advancements as threats, but rather as opportunities. By integrating Artificial Intelligence into their solutions, they are creating hybrid tools that combine the stability of RPA with the flexibility and power of generative AI.

This convergence paves the way for more seamless automation, capable of handling both high volume processes and complex tasks requiring cognitive capabilities.

‍

The future of RPA : transformation rather than disappearance

RPA, as it once existed, may be declining, but it is evolving to meet the demands of modern environments. Agentic technologies will not replace RPA overnight, but will expand and enhance its scope. In the long term, we will likely see :

• Widespread hybridisation : organisations will combine RPA and AI to leverage the strengths of each technology
• Differentiated adoption : sectors where adaptability and decision making are critical, such as finance or logistics, will prioritise autonomous agents, while others will continue to rely on enhanced traditional RPA
• Intelligent and strategic automation : the combination of these technologies will optimise costs, performance and organisational agility

Conclusion : no, RPA is not dead

RPA is not disappearing, it is undergoing a transformation. The rise of generative and agentic AI does not signal the end of this technology, but rather the beginning of an era of augmented automation. Organisations will need time to fully harness this potential, but will ultimately benefit from more robust, flexible and high performing automation.

As highlighted by Morand Studer, Partner at eleven strategy, the future will not be defined by a competition between RPA and AI, but by their integration. This coexistence will provide organisations with tailored solutions, capable of addressing their specific needs while adapting to the demands of an ever evolving environment.

‍