Firstly, concerning Phase 1 (Studies 1 and 2), which investigated Hungarian primary and secondary school EFL teachers’ DI and TEDI beliefs and practices, the analysis was based on self-reported data, which means that possible discrepancies between the participants’ actual and reported practices remain an issue. In the future, it would be important to triangulate data by observing the teachers’ lessons and examining their lesson plans, as well as by collecting data about the students’ experiences, which, taken together, could help provide a more in-depth understanding of the state of affairs of DI and TEDI implementation in Hungarian primary and secondary school EFL classes.

Another important consideration in the interview study (Study 1) of Phase 1 is the composition of the sample. While it included teachers with varying levels of experience, six out of the nine participants had no more than 12 years of teaching experience, which may have influenced their beliefs about DI and TEDI. Although the analysis of both the interview (Study 1) and questionnaire (Study 2) data did not reveal significant differences in these beliefs based on teaching experience, this does not rule out the possibility that more experienced teachers may hold different views on these concepts. Conducting future research exclusively with more experienced, senior teachers would provide data that could then be compared with the outcomes of Phase 1.

Another limitation of the questionnaire study in Phase 1 (Study 2) concerns the fact that the respondents were recruited through snowball and convenience sapling, which limits the generalisability of the results due to the “compromised nature of non-probability sampling” (Dörnyei, 2007, p. 99). Following Dörnyei’s guidelines (2007), I sought to mitigate this effect by describing in sufficient detail the limitations of the sample while also highlighting the characteristics that the sample shares with the defined target population, and, secondly, by being careful about the claims I made concerning the more general relevance of my findings. Still, it is important to acknowledge that the inherent biases of non-probability sampling could have influenced the results and that replicating this questionnaire study would be important in corroborating the findings of Study 2.

Besides, the relatively low R2 values of the regression models in Study 2 also need to be addressed. As was noted in the discussion chapter, these weaker values suggest that the current set of independent variables explains only a limited proportion of the variance in DI implementation. In the future, it would be important to explore additional dimensions discussed in the literature, such as classroom size, the diversity of the students, teacher workload, and national curricular demands (Chien, 2015; Gülsen, 2018; Tzanni, 2018) to improve the explanatory power of the models.

As for Phase 2, the question of the small sample size needs to be addressed, as this phase involved five focal participants in the interview study (Study 3) and the observation study (Study 4), and 39 students in the questionnaire study (Study 5). The aim of this phase was not to generalise but to conduct a close-up exploration of TEDI practices, in order to “understand the intricate tapestry of classroom events” (Dörnyei, 2007, p. 176), and, based on those outcomes, establish a transferable framework for those interested in harnessing the potential of technology for DI. Given this agenda, the study was inherently designed to be a small-scale, exploratory investigation. Nevertheless, it is important to acknowledge that Phase 2 of the research project focused on a narrow segment of Hungarian public education—schools that are exceptionally well-equipped with technology and predominantly serve students from higher socioeconomic backgrounds—which could limit the transferability of the findings. To address this, future research should include schools with a wider range of pedagogical profiles. Since limited technological facilities remain a challenge at both primary and secondary levels in Hungarian public education (Ringert, 2019; Tartsayné Németh et al., 2018), it would be important to examine how TEDI could be used in schools with fewer digital tools. Such studies could help identify practices suitable for minimally resourced teaching contexts and thus further refine and strengthen the applicability of the TEDI framework identified in this research project.

Furthermore, regarding the observation study (Study 5), it should be noted that although I had hoped to visit more than one class of each teacher, due to their busy schedules, I eventually had the opportunity to observe only one lesson per teacher, which may pose threats to the reliability and validity of the data. This limitation was partly mitigated by the informal discussions I had with the teachers after the observed lessons, which helped further understand their practices as they described the observed group, demonstrated the main features of some of the applications they used in the lesson, and shared their opinions on the benefits of these tools. Nevertheless, in the future it would be useful to observe further lessons of each teacher. Thankfully, several participants expressed interest in continuing our cooperation and kindly offered to open their classroom doors for further research.

Continuing the investigation at these—and other—research sites could help address two further limitations of Phase 2. Firstly, regarding the effectiveness of TEDI, Phase 2 focused on possible improvements in affective and cognitive variables such as motivation, language learning experience and self-efficacy, but developments in students’ language proficiency were not examined. Incorporating a (quasi-)experimental element in the investigation could help understand how TEDI practices impact students’ actual language skills, which is an important aspect in assessing the effectiveness of TEDI in TEFL.

Secondly, the rapidly developing nature of educational technology means that some of the ICT tools investigated in this book may soon become partially outdated or replaced by newer technologies. While the framework laid out in this project is hoped to prove useful in the long run and adaptable to new technologies, one aspect that is already lacking from the discussion and would be necessary to address in the future is artificial intelligence (AI). Although certain TEDI practices discussed in this book are based on AI through their adaptive technology (e.g., ELSA, BOOKR), AI technologies themselves (e.g., chatbots, adaptive tutoring systems, machine translation, automated writing evaluation) were not the focus of the research since they had not yet become widespread at the time of the data collection (2021–2023). As several recent studies have highlighted the potential of AI in making the learning process more personalised (Ayeni et al., 2024; Choukaier, 2024; Pratama et al., 2023; Rane et al., 2024), it will be important in the future to explore how the “brave new world of educational AI” (Hockly, 2023, p. 1) could enhance DI in TEFL.

A final note concerns the underlying presumption of this monograph: that using technology for DI holds benefits for both students and teachers. While this research project primarily focused on these benefits, it also sought to explore the concerns raised by the participants (e.g., internet connection issues, technology overuse, the challenge of keeping up with trends). As Larsen-Freeman and Anderson (2011) argue, in the ever-increasing presence of technology in education, it is essential that teachers “be knowledgeable about technology, and if they choose to use it … they do so in pedagogically sound ways” (p. 268). In the future, it will be important to further examine concerns related to educational technology, such as the digital wellbeing of learners and teachers, and issues around data collection, surveillance, and privacy (Hockly, 2023). The outcomes of such inquiries could guide informed decisions about technology use, while also helping to harness its potential to provide meaningful language learning opportunities for all.