ARISE: Agentic AI-RPA Implementation for SMEs Automating Marketing and Customer Relations in Resource-Constrained SMEs

Service SMEs, particularly independent hotels and camper rental companies, face a structural productivity challenge driven by persistent labour shortages. Owner-managers must combine operational responsibilities with customer-facing activities such as marketing and enquiry handling. This creates a constraint trap: limited capacity for communication and outreach restricts revenue growth and reinforces the inability to scale.
Traditional Robotic Process Automation (RPA) provides limited relief, as it depends on structured data and predefined rules, whereas customer interaction and content creation are inherently unstructured and judgement-intensive.
This project addresses this gap by configuring an integrated Agentic AI–RPA architecture that combines intelligent interpretation with automated execution. By embedding AI-driven decision-making into RPA workflows, routine cognitive tasks—such as interpreting enquiries, drafting responses, analysing social media performance, and generating content—can be partially automated. This shifts human involvement from continuous task execution to targeted oversight and exception handling, enabling staff to focus on higher-value activities such as customer experience design and strategic positioning.
The project configures two TRL 5/6 pilots using the existing OpenClaw Agentic AI framework: a Marketing Agent (Hotel De Plataan) that analyses social media metrics and generates and optimises content, including timing and posting strategies; and a Customer Relations Agent (Holland Campers) that classifies enquiries, retrieves relevant data, and drafts contextualised responses. Both workflows incorporate human-in-the-loop validation prior to publication or dispatch.
Using an Action Research approach combined with a quasi-experimental pretest–posttest design, the project evaluates how AI–augmented automation measurably reduces labour intensity while improving process performance and reallocating human effort towards higher-value activities.
The project delivers validated AI-RPA workflow prototypes, documented SME learning trajectories, and application-specific implementation playbooks. Disseminated via the MKB Digiwerkplaats Den Haag, these outputs provide resource-constrained service SMEs with practical, risk-reduced pathways to adopt Agentic AI as a cognitive labour multiplier.