← CHI 2026 map

CHI '26 · Honorable mention · full-paper review · confidence medium-high

Where Will They Click Next? A Social Foraging Model for Collaborating Teams

Shahnewaz Leon , Sandeep Kaur Kuttal

DOI PDF Program page Page JSON

PFIS-T is a credible and well-scoped CHI contribution: it moves next-step prediction from an individual-centric framing to a team-aware one, and the evaluation directly supports that shift in a realistic collaborative debugging setting. The main caveat is that the evidence is still bounded to a controlled lab task, so the broader systems vision remains prospective.

Axes Lens

Rare contribution shape, typical evidence profile. The point here is not a score. It is to show what kind of claim the paper makes, and whether the evidence pattern is unusual or baseline in this 268 -review set.

Review Summary

PFIS-T’s value is that it makes a fairly clean conceptual move: instead of treating next-step prediction as a solo user modeling problem, it explicitly models social information foraging by folding in teammates’ recent navigation and synchronous communication. That is a meaningful departure from the common baseline in this area, where team context is often ignored or reduced to static collaboration metadata. The paper’s novelty is therefore not just a new predictor, but a reframing of the predictive target itself: the model treats social cues as moment-to-moment information scent. The validation is also appropriately aligned with the claim. The authors do not overreach into deployment; they evaluate in a controlled virtual-lab study with ten three-person debugging teams and compare against the strongest individual baseline, PFIS3. That makes the reported accuracy gain relevant evidence for the specific task class. At the same time, the limitations are clear and important: the model relies on recent navigations and contemporaneous speech/text, and the evidence comes from a single collaborative debugging context rather than diverse team settings. So the paper is strongest as a technical and methodological contribution to team-aware prediction, with a plausible but still unproven path toward collaborative IDEs and adaptive awareness tools. In CHI terms, this is a solid honorable-mention style paper because the idea is crisp, the evaluation is credible, and the scope is honest, even if the broader generalization to all collaborative work remains to be shown.

What Changed

Canon before

Prior work on information foraging and next-step prediction is largely individual-centric, so team support has typically been indirect or absent rather than modeled as a social, moment-to-moment predictive process.

Departure from common sense

Existing computational models of information foraging remain individual-centric, leaving teams without support for social foraging —leveraging partners’ actions and communication to navigate complex projects.

Actual novelty

We introduce PFIS-T, the first predictive computational model of social information foraging in synchronous teamwork. PFIS-T treats teammates’ recent navigational activity (implicit social cues) and their conversational data (explicit social cues) as moment-to-moment sources of information scent—traces signaling where valuable information may be found.

Evidence

The paper’s core contribution is a predictive model, PFIS-T, that extends the PFIS family by incorporating both implicit teammate navigation and explicit conversational cues to predict a programmer’s next action. The evaluation is a controlled virtual-lab study with ten three-person teams doing collaborative debugging, and the reported performance gain over PFIS3 supports the claim that social cues improve next-step prediction in this setting.

Limits

Method limits

The evidence is strongest for the specific collaborative debugging setting studied. The model is trained and evaluated on navigation and conversation traces from ten three-person teams, so the methodological scope is limited to this task structure and data modality.

Deployment limits

The paper’s deployment implications are suggestive rather than demonstrated: it points to collaborative IDEs and adaptive surfacing of social trails, but does not show in-the-wild integration, long-term use, or effects on team outcomes beyond prediction accuracy.

Boundary conditions

The approach depends on access to recent teammate navigation and synchronous communication. Its validity is therefore bounded by settings where those signals are available, timely, and meaningful for the task; the paper’s own framing centers on collaborative debugging rather than all forms of teamwork.

Manual check: No manual check required from the provided evidence packet; the main uncertainty is generalization beyond the evaluated debugging setting.

Position in field

This sits at the intersection of information foraging, team-aware prediction, and collaborative software engineering tools. Its field position is as a model contribution that reframes next-step prediction from an individual behavior problem to a social, team-level inference problem.

Abstract

Modern knowledge work is increasingly collaborative, especially in information-intensive domains such as crisis response, scientific discovery, and software engineering. Software engineering epitomizes these trends through practices like pair programming and collaborative debugging. Yet existing computational models of information foraging remain individual-centric, leaving teams without support for social foraging-leveraging partners’ actions and communication to navigate complex projects. We introduce PFIS-T, a predictive computational model of social information foraging. Building on the PFIS model family, it integrates implicit cues from teammates’ recent navigation and explicit cues from synchronous communication to predict a programmer’s next action. We evaluated PFIS-T with ten three-person debugging teams, finding that it substantially outperforms the strongest individual baseline, PFIS3, predicting 81.5% of navigations and improving accuracy by 16.7%. These results show how predictive models can operationalize social foraging and point to opportunities for collaborative IDEs and interactive systems that adaptively surface social trails to improve coordination and awareness.