Can we automate away the main challenges of end-to-end testing?

Agile methodologies enable companies to drastically increase software release pace and reduce time-to-market. In a rapidly changing environment, testing becomes a cornerstone of
the software development process, guarding the system code base from the insertion of faults. To cater for this, many companies are migrating manual end-to-end tests to automated ones. This
migration introduces several challenges to the practitioners. These challenges relate to difficulties in the creation of the automated tests, their maintenance and the evolution of the test code base. In this position paper, we discuss our preliminary results on such challenges and present two potential solutions to these problems, focusing on keyword-driven end-to-end tests. Our
solutions leverage existing software artifacts, namely the test suite and an automatically-created model of the system under test, to support the evolution of keyword-driven test suites.