Internet agents are at the heart of web search engines and support the user with flexibility on information search. While search engines are built on AI, they are tightly anchored to the principles of HCI and human-agent interaction. Search engines are usually made popular through social networks, but users are skeptical and rely on trust and competency of results before adopting a preferred engine. The effective use of any intelligent software requires evaluation practices to measure how the user performs in relation to the technology. Synthesizing on user performance, studies show that several attributes in the theory of action describe the sequence of steps behind a person interfacing with computers. The study presented here offers a balanced coverage of how users perform with Internet agents. Two agent types were tested, fixed agents that learn statically from the user queries and evolutionary agents that learn dynamically from user communities with similar inquiries. Four search engines were assessed to examine which factors are useful to search performance and to HAI usability. Likewise, the research design informs techniques for writing a thesis or a dissertation project.