Lost by Over-Management: Adaptive Notification Model for Handling Weakly Planned Activities ^†

Abstract

The study explores the scenarios and approach to the design of the software for managing notifications about the fuzzily planned activities. Though many such scenarios can be solved by using traditional time and activity management tools such as organizers, diaries, planners, or schedulers, practical situations often arise when people tend to avoid overmanagement for real-life situations, when the plans might be flexible, and the planned activities might depend on location, contextual, and time information, which may not necessarily be well known or configured in advance. In this contribution, we describe examples of such situations and define the concept of soft planning. Following the principles of the human-driven design paradigm, we conducted a small-scale survey to gather insights into user preferences and identify drawbacks of existing digitalized activity planning and decision-making tools, often based on configurable notification management software. The findings reveal that while notifications are useful, users often encounter issues such as information overload, lack of contextual awareness, and disruptions caused by the notifications arriving at inconvenient or even inappropriate times.

1. Introduction

Planning regular activities has been a behavioral model of human life organization since ancient times. In present-day terms, the technology of planning involves numerous formal and informal methods to organize, manage, schedule, and optimize human activities in daily life. These methods have been evolving along with the growth of society, while being continuously developed to match the new demands of each era. In ancient civilizations, the first calendars were invented to help in organizing agricultural activities, religious practices, and other social events, such as in ancient Egypt, for example (see Figure 1), where the lunar calendar was used for religious functions mostly, while the civil (solar) calendar was more commonly used for everyday life organization [1].

Calendars, however, act as prescriptive generalized models rather than personalized solutions. The evolution of planning systems spans over hundreds of years, beginning with notes written on stones, all the way to the use of paper for managing records, time, and tasks. As explored in The Accidental Diarist: A History of the Daily Planner in America [3], these tools have transitioned over time from simple organizational aids to personal artifacts that reflect the daily routines of their users. Interestingly, access to such artifacts not only delivers knowledge about the characteristic societal processes of the past but often brings valuable insights into the lives, affairs, and endeavors of important individuals. For example, the few examples of authentic handwritten signed documents, such as house purchase certificates and mortgage papers, created part of the initial grounds for the emergence of the so-called Shakespeare authorship problem, which is full of mysteries and highly disputable [4].

Along with the rise of industrialization, the need for proper time management also increased. Daily planners became truly essential for coordinating work schedules, tracking productivity, and balancing the responsibilities of the involved stakeholders.

Major scenarios suggested by such “tools” have been digitized with the emergence of the computer-based solutions for scheduling the regular activities, time and project management, meeting planning and organization, issuing and keeping reminders, etc. These solutions are typically deterministic since they naturally require the precisely defined conditions such as a known time frame, location, resources, and connected activities. For example, to organize an efficient research meeting, one needs to decide on the following matters:

• Define the scope and the major topics to be discussed;
• Decide on the list of participants;
• Select a time slot that is convenient for the participants (tools like Doodle [5], When2Meet [6], or Calendly [7] can be helpful);
• Define a location (or a link for an online activity);
• Decide on the required resources (handouts, equipment, stationery items, distributable materials, software, etc.);

Despite the efficacy of such traditional planning solutions within the well-structured and well-organized environments, they often fall short in situations in which planning parameters are partially unknown or just do not need to be perfectly known or defined in advance. This contribution tries to address such weakly (not to be confused with “weekly”) planned scenarios. Activity planning in such cases suggests the need for a more adaptive approach that would accommodate a number of possible ever-changing or unfixed situations and deferred or loosely organized activities. To address these special though relatively infrequent situations, a model was sketched in [8], which introduced a concept of soft planning that would complement the traditional time- and location-based organizers by incorporating the features to describe and support the uncertain and evolving conditions happening in the process of task and reminder management.

The remaining text is organized as follows. In Section 2, we exemplify the weakly planned scenarios through a number of practical situations. Section 3 explores the aspects of ensuring a soft planning model through the principles of human-centric design (HCD), focusing on in-advance elicitation of user expectations and preferences acquired from a small-scale survey presented in Section 4. Section 5 introduces an adaptive notification model and presents an approach to designing a prototype of the system through the conceptualization of how soft planning and adaptive notifications can be integrated into a practical system.

2. Example Scenarios of Soft Planning

In this section, we sketch a number of practical situations that can be considered good illustrations of soft planning scenarios.

2.1. Proximity-Based Shopping Reminder

One notices that they have run out of a product they have at home. They may need to purchase some more, but the problem does not require any immediate solution as sketched in Figure 2. However, it would be nice to be reminded while visiting a nearby food store.

2.2. Context-Aware Task Management

A homemaker needs to change the water filter as soon as possible (not necessarily at a strictly defined moment in time). This task can only be accomplished at home, so it would be helpful if the reminder appears only while at home and at an appropriate moment of time, for example, not during the nighttime or too early in the morning. Figure 3 depicts the case.

2.3. Seasonal Travel Recommendation

A traveler discovers an interesting but not-so-famous sightseeing spot that would be great to visit at a different moment in time (for instance, during a better season) as diagrammed in Figure 4. The exact time when this traveler can come back to the area cannot be determined yet, but it would be helpful to receive a reminder about this place if it happens that one day the traveler is around the area again.

2.4. Leisure Activity Suggestion

A movie enthusiast loves to go to the cinema (or theater) to watch a new much-awaited but not-yet-released production. Soon, a movie is going to be released: the reminder might be helpful while having some free time in a place with a good theater nearby. Figure 5 illustrates this scenario.

2.5. Location-Based Task Reminder

A student requests a document from the research supervisor that needs to be submitted to the student affairs office. The supervisor left the document on his office door so the student could take it anytime. There is no urgency, but in order not to forget about the duty prior to the due date, a reminder might be helpful when the student goes to attend class at the nearest location on campus. Figure 6 draws on this use case.

2.6. Location-Driven Immersive Learning

An AI-driven immersive learning scenario suggested in [9] can be reconsidered as a softly planned activity (as sketched in Figure 7). A learner is encouraged to study the language vocabulary in suitable contexts. For example, while waiting for a train at the platform, the learner could receive a reminder to take a lesson focusing on the words and phrases commonly used at the station (announcement, ticket counter conversations, asking platform directions, etc.). Similarly, while visiting a hospital and waiting for an appointment, relevant exercises could be practiced, aiming at acquiring the necessary typical vocabulary and phrasal patterns to describe symptoms or request medication.

The examples presented in this section introduce the situations that are not critical and do not necessarily require the precision of a traditional, robust time management or scheduling system. Instead, they might need a more flexible and adaptable approach. Although making them fit into a traditional time management system might be possible, it might take away their uncertain nature, and some might consider them a kind of over-management.

3. Related Work

Human-centric design is acknowledged in this paper as a broad paradigm establishing the principles of designing products and services (naturally including digital products and services such as software) according to the careful address of the personalization-focused user demand, cultural and social contexts, as well as deployment and usage constraints. These constraints and contexts may include adaptability to different age groups, social groups, technology backgrounds, discrepancies in the region- and culture-associated communication and information presentation styles, and many other factors [10,11]. The model of adaptive notifications handling uncertainty and personalizing user interactions within the soft planning scenarios represents an applied specific concept to be examined through the prism of the more common human-centric design principles.

3.1. Human-Centric Design Principles

Human-centric design, also known as human-centered design, is a methodology that places the users at the center of the design process. According to the International Standards Organization (ISO), the definition of human-centered design is as follows:

“Human-centred design is an approach to interactive systems development that aims to make systems usable and useful by focusing on the users, their needs and requirements, and by applying human factors/ergonomics, and usability knowledge and techniques. This approach enhances effectiveness and efficiency, improves human well-being, user satisfaction, accessibility and sustainability; and counteracts possible adverse effects of use on human health, safety and performance.”

—ISO 9241-210:2019(E) [12]

Though the ISO definition has established the concept frame, there is much space remaining for fine-tuning this definition to the practicality of building human-centric systems prioritizing the needs, experiences, and feedback of users throughout the design process. For example, the general principle of sustainability can be further interpreted as the requirement for a system to be intuitive and functional while providing neither over-complicated nor over-simplified user interfaces. In other words, according to Norman [13], human-centered design is a process of ensuring that people’s needs are met and that the resulting product or service is usable but also understandable.

Adaptability to user behavior, preferences, background, and context is a key parameter of the soft planning concept, which is why it seems natural to observe and apply the principles of the broader human-centric paradigm. Additionally, adapting human-centered design principles can promote inclusivity by ensuring that the designed tools are accessible and usable for a diverse range of users with varying levels of technological proficiency, or even for users with disabilities. Sargiotis, in [14], supports this idea by stating that human-centric design can be used to ensure that cutting-edge technologies such as AI are developed and deployed responsibly across different domains and are aligned with evolving and often reconsidered ethical standards.

3.2. A Concept of Adaptive Notification System

Adaptability and responsiveness are among the key attributes of the human-centered design paradigm [15]. As we mentioned earlier, in the scope of managing weakly planned activities and decision making, we work with uncertain things, which means there is a need to consider the actual context and user preferences. One of the major issues that needs to be resolved is how to manage the notifications in a way that they are prompt, meaningful, and relevant. Mehrotra and Musolesi, in [16], discussed several concerns, including the possibility of notifications arriving at inappropriate times, thus potentially disrupting the user instead of helping them. This needs to be avoided, as it defeats our original purpose of avoiding annoying notifications.

There have been several interesting studies on how notification time and content can affect how users perceive them. These studies examine the impact of tailoring mobile notifications based on different personality traits [17], predicting the user’s preferences [18], and improving user interaction by using AI to personalize the notification content [19]. Particularly, in [17], it was found that tailored notifications might affect some groups of people while having no visible impact on other groups. By contrast, [19] discusses how personalization can increase user satisfaction and foster a deeper emotional connection between users and apps, leading to higher retention rates.

Thus, notification systems require adaptability, especially in contexts in which users’ preferences, priorities, and environments are constantly changing. For soft planning scenarios, this adaptability becomes even more important, as it involves dynamic and uncertain plans (some of which may not be accomplished), which can only be effectively managed with a system capable of adjusting to the actual purpose, time, location, and resource context. While human-centric design principles guide the system’s usability and inclusivity, adaptive notifications provide the technical mechanisms to address the inherent uncertainties of soft planning. Together, they not only respond to user needs but also evolve alongside their changing contexts, ensuring that the system remains both useful and non-intrusive.

4. User Anticipations and Preferences: The Survey

4.1. Overview

The purpose of this survey was to gather insights into user experiences and preferences while utilizing digital planning tools in daily life, with a focus on identifying the drawbacks (and, in part, over-management) of existing “traditional” systems for handling reminders through notifications. We also examine the extent to which the users are interested in adaptive features, such as context-aware and AI-driven notifications.

The small-scale survey presented in this contribution is a preliminary step in ensuring that our idea aligns with what users actually want and anticipate to address their needs effectively and empathically, as assumed by human-centric design principles. Understanding these user perspectives will create the necessary meaningful and arguable grounds for the design process.

4.2. Key Findings

To understand the importance of the digital planning tools, we asked the participants about their usage frequency. The results show that 60% use these tools frequently, either daily or almost every day, while 20% use them weekly, and another 20% rarely use them. This indicates that planning tools are widely utilized, although with varying frequency.

Focusing more on the notifications, we asked participants about their usefulness and timing. Figure 8 and Figure 9 depict some results and suggest that notifications are a valuable part of daily planning. However, half of the participants reported that notifications can be disruptive or distracting at times.

We also explored issues with the current notification systems. Half of the participants stated information overload as a major concern, while the other half mentioned a lack of contextual awareness. Other issues also included notifications arriving at inconvenient times and missing important notifications when devices are in sleep mode.

Regarding adaptive notifications, participants were asked how important they are for planning tools in connection to factors such as location, time, or activity (see Figure 10). The responses were mixed; therefore, more detailed questionnaires might be required.

We also asked the participants why they would or would not use a system that sends reminders based on their actual situation. Many participants found it potentially helpful for context-specific tasks, such as receiving reminders upon arriving at school or recommendations about eating at a nearby restaurant. Some would also appreciate features like location- and transportation-based notifications for recurring events, which would enhance convenience and accuracy. In some answers, the value of postponing notifications and automatic muting during the calendar events was highlighted, as this would prevent disruptions. Adaptive reminders were viewed as particularly beneficial for non-essential activities. However, some participants expressed hesitation, concerned with the effort required to set up such a system.

Lastly, we also asked the participants to share their concerns about using a system that could access their real-time data (such as location) or grant AI assistants access to these data. Figure 11 shows the results.

While 60% of participants expressed privacy concerns, 20% were not concerned, and 10% were neutral. Despite these concerns, 90% of participants were comfortable or somewhat comfortable with AI handling their data, provided that there is a clear privacy policy. Only 10% objected, emphasizing the need for transparency and data protection.

4.3. Summary

Overall, the survey results highlight the widespread use of digital planning tools and the importance of notifications in daily planning, although issues such as information overload, lack of contextual awareness, and timing disruptions remain significant concerns. There is an interest in adaptive notification models that would support features such as context-aware reminders to reduce disruptions, the latter being aligned with our focus on enhancing the user experience. However, privacy concerns must be addressed through transparent data practices.

5. Toward Adaptive Notifications: The System Structure

The findings from our survey, presented in Section 4, show the key challenges of the traditional notification systems. Potential users are interested in adaptive features that would adjust notifications dynamically based on user preferences, schedules, and the real-time context.

Additionally, we believe that it is beneficial to integrate AI into the soft planning system to personalize the content of notifications. As supported by the findings of [17,19], we can see that users can perceive or interact with notifications differently depending on the content of the notification or how it is written. For example, some users may prefer concise, action-oriented messages, while others might have a better understanding with more detailed or empathetic tones. Additionally, the system can also adjust the content of the notification to match the urgency of the task.

For weakly planned activities, such adaptive notifications can enhance flexibility and responsiveness and help users manage dynamic plans. By adjusting the flow of information to align with user preferences and situational demands, we believe that irrelevant notifications can be reduced.

By combining adaptive notifications and AI for soft planning, we aim to construct a system that would enhance the user experience, increase planning efficiency, and provide a seamless, intelligent, and non-intrusive approach to managing weakly planned or optional activities.

Figure 12 presents a simplified system structure illustrating how soft planning, adaptive notifications, and AI agents can be integrated into a functional system. The modules are aimed at facilitating adaptive notifications based on user input and actual data.

The user input primarily suggests scheduled events, where the activity details can be defined along with specific conditions, such as time constraints and location preferences, in order to determine when the notifications should be triggered. Since the actual data, such as user location and current activity (e.g., walking, driving, or in a meeting), are taken from the user’s mobile device, these data are first processed in the context of the module, which checks whether the conditions taken from the user input are fulfilled. The decision-making module determines whether a notification should be sent by taking into consideration factors such as priority levels for different types of notifications, preferred notification issuance times, and “do-not-disturb” situations set by the user.

Optionally, before notifying the user, the system can process the constructed notification through the AI module, where its content is analyzed and refined based on user preferences and past interactions. Once ready, notifications are sent to the user through the notification delivery module. Additionally, users can also provide their feedback by dismissing them, delaying them, or marking them as irrelevant. This feedback would help to refine the system’s “understanding” of the user’s preferences over time and therefore enable adaptation to the user’s needs and ensure the relevance of the system’s response.

The diagram in Figure 12 outlines the key components and their interactions and illustrates the main processes, contextual factors, and notification delivery workflow.

6. Discussion

The paper explores an approach to constructing an adaptive notification system addressing the example scenarios of soft planning. By considering users’ context-related factors, such as transportation mode, actual schedule, location, travel time, and user availability, the system aims to improve timely and personalized reminders for weakly planned activities.

The major results of the current study are as follows:

• We described examples of practical scenarios illustrating soft planning contexts, in which traditional approaches to time, resource, and location-based scheduling are applicable but might lead to a sort of over-management.
• We positioned the project on constructing an adaptive notification system within the scope of the human-centric design paradigm and examined the factors that need to be assessed while working with the prototype implementation.
• Based on the small-scale survey, we defined the important challenges and risks that must be addressed in order to create an effective, user-friendly, and non-disruptive implementation.

Avoiding notification fatigue is a major concern in managing notification workflow [20]. Timely notifications must arrive at the right moment; however, too many notifications might lead to information overload. Additionally, there are important privacy concerns, as efficient personalization inevitably requires accessing user data, such as location or activity patterns. Therefore, transparent policies are required to guarantee the user’s control over which data are shared.

To address these concerns, we plan to take a human-centered design (HCD) approach during development. This involves user testing, feedback collection, and iterative refinement of the adaptive notification system. By involving users directly throughout the development process, we hope to find a balance between timely notifications and minimizing disruption while also exploring how to enhance personalization and privacy control. This iterative approach will hopefully create a notification system that aligns with diverse user needs and supports flexible real-world scenarios.

Finally, we believe that there are psychological factors that impact how the users interact with the system. Some individuals may still prefer using solutions that favor precise and fixed time management, while others may avoid over-management and prefer having more flexibility in their planning of non-critical or optional activities and tasks. An adaptable and soft approach to scheduling might appear to be helpful for people who tend to procrastinate and are seeking a way to add more regularity to their task management.