Abstract
: Autonomous Artificial Intelligence (AI) agents are finding enormous interest from both industrial and academic developers. One such vital area of application of AI agents is in automated web and computer use. These automations were traditionally done by the Robotic Process Automation (RPA), which are rule based and static in nature. AI agents offer advantages over RPA by performing Web and Computer use automations using natural language instructions as well as they can dynamically adapt to changing interfaces using strong reasoning intelligence embedded in underlying AI models. In this paper, we present a study which can de-mocratize the Autonomous Web agent within the corporate environment. This will help various departments to improve efficiency by independently fulfilling repetitive, monotonous and complex web navigation operations using basic Python proficiency. We have initially developed an autonomous web agent and then built user interaction layer (RPA Agent) on top of it for its general use. Together, the agent receives a high-level instruction and autonomously carries out the task, effectively imitating human-like actions. Various building blocks of this RPA agent are discussed in detail. To demonstrate its usability, RPA agent is used to solve the problem statement by the marketing department. Compared to conventional RPA development time, the agent has provided required market insights in significantly less timespan.