Users utilize web applications to perform everyday tasks in order to achieve personal goals. Personalized Web-Tasking (PWT) is the automation of such web interactions while exploiting personal context to enrich users experience. However, web-tasking is affected by unpredictable context behaviour -- environment, user, and infrastructure -- and situational changes. Given that current web systems are challenged to respond effectively to such changes, we proposed to design PWT applications as self-adaptive software systems that exploit personal context to deliver user-centric functionalities. This paper presents our first approach implementing PWT applications using a grocery shopping web-tasking scenario. Our prototype PWT system transforms web-tasking knowledge information (i.e., user's web interactions) into RDF graphs (i.e., runtime models that contain the user's web-tasking). We conclude our paper with a discussion about our results and implementation challenges.