The final speaker in this session at the SEASON 2025 conference is Elsa Lichtenegger, whose projects is exploring the question of search intent. Why do people search, what drives them, what is their motivation, what are their goals? Search intent describes the user’s underlying goal or purposes, which is expressed through the initial search query and may be refined through further search iterations.
A definition of search intent by Broder (2002) tends to be dominant in the available literature on the topic; this has considerable impact on extant research. It distinguishes three types of user intent: informational intent (seeking information); navigational intent (seeking to reach a specific Website); and transactional intent (seeking to complete an action). This is simple and intuitive, is immediately applicable to search engine queries, and fits what search engines looked like when they first became key portals to information – however, how applicable is this still to the multipurpose platforms that search engines have evolved into over the past two decades or more?
Perhaps, then, there is a need to reconsider our understanding of search intent: perhaps there are other types of search intent that we might need to consider. How might we unveil the intent underlying users’ engagement with search engines, though? This should not start with the search query itself, but with the overall search session; users might be asked to reflect on their immediate goals, broader goals, and initiating context for their search sessions, for instance. This could be operationalised by asking respondents in a research project to look at their recent search history, for instance.
To do this with conventional search engines is no longer enough, though; it can and should also be extended to considering AI chatbot sessions, and perhaps also to other similar platforms.
Applying Broder’s taxonomy, informational intent tends to be highly prominent; this is true for both search and chatbot sessions. Intent can also be ambiguous and not easily determined from the query formulation alone; this highlights the value of engaging with users directly to recall and explain their intent. Further, Broder’s navigational intent tends not to be relevant for chatbot sessions: few users use them to find Websites to visit. Conversely, some chatbot sessions are highly conversational, in a way that search sessions will never be.
A refined taxonomy emerging from these observations across search and chatbot sessions, then, shows three broad patterns: one is knowledge-seeking, which is about understanding, verifying, and explaining facts and concepts; a second is guidance-seeking, where users seek advice or instructions for solving specific problems and making decisions; a third is output-seeking, where they seek particular products, tools, or resources that meet their needs.
This represents a less resource-centred approach that Broder’s taxonomy and focusses instead on goals, and also moves away from a query-level analysis to consider the full session; through this reconsideration, informational seeking splits into knowledge and guidance seeking, while Broder’s other types also redistribute across the new types of search intent. Queries that did not fit the original framework now find a home in the revised taxonomy, too, accounting for new intents that have become possible in the new world of information-seeking.
