Voices that don’t need to speak: augmentative communication apps that change lives

Introduction: Augmentative communication applications as tools for participation and self-determination

From speech compensation to building one’s own voice

Augmentative and alternative communication (AAC) applications shift the focus from compensating for speech limitations to expanding participation, self-determination, and personal expression. Individuals with complex communication needs face restrictions that affect their educational, occupational, and social goals, as well as their quality of life (Karal et al., 2025) . An application may offer symbols, text, stored phrases, speech synthesis, or personalized recognition, but its value depends on its ability to initiate interactions and communicate content chosen by the user. Reviews on self-determination show that AAC users have the right to participate in decisions about their rehabilitation, although self-assessment tools remain scarce in areas such as feelings, motives, and languages other than English (Savolainen & Pilesjö, 2025) .

Scope of evidence: autism, aphasia, apraxia, dysarthria, ALS and dementia

Recent evidence encompasses childhood autism, apraxia of speech, aphasia, dysarthria, amyotrophic lateral sclerosis (ALS), dementia, and acquired neurological disorders. However, the populations, designs, and outcomes differ: some studies examine specific skills in brief interventions, while others explore everyday use, accessibility, or quality of life. Therefore, an app that increases requests in an educational setting does not, by itself, demonstrate an improvement in friendship, self-advocacy, or spontaneous conversation. This distinction guides the review: word production is an initial outcome, while communicative participation requires evaluating diverse roles, interlocutors, and environments.

Beyond asking: communicative functions, autonomy and learning with mobile applications

Effects on functional communication and autonomy in natural environments

Experimental evidence supports the idea that mobile applications can enhance certain functional communicative behaviors, particularly when the intervention combines modeling, prompting, and reinforcement. In a single-case study with a multiple baseline design, an MPR- based intervention —modeling, prompting, and reinforcement—increased requests, refusals, and requests for help from an eight-year-old autistic child; the improvements were maintained two weeks later and generalized to a new professional (Kim & Noh, 2025) . The result is significant because it includes refusal and assistance, functions that allow for regulating an interaction and not just obtaining an object. However, the single-participant design limits generalizability to other ages, language profiles, and devices.

Generalization requires that the application be incorporated into ordinary activities, with different interlocutors and varying communicative demands. In adolescents with complex needs, two high-tech interventions used scripts organized by location or by category; both increased expressive accuracy in the community, but the location-based tool showed greater effects and generalization in that setting (Han et al., 2025) . The comparison suggests that the organization of the interface should correspond to the ongoing activity. A categorical screen can facilitate conceptual searching, while a location-based structure reduces decision-making when communication occurs in predictable spaces.

Autonomy does not equate to using the device without assistance. It includes choosing what to say, when to say it, how to correct one’s output, and whether to continue the conversation. In a review of self-assessment instruments for AAC users, most measures explored thoughts and beliefs, but gaps remained in feelings, motives, psychometric properties, and non-English languages (Savolainen & Pilesjö, 2025) . Evaluations should therefore ask the user whether the application reflects their preferences and facilitates everyday decisions, rather than simply recording response frequency.

The gap between requesting, commenting, regulating, and expressing emotions

The literature reveals a persistent asymmetry: prompting is typically taught earlier and measured more frequently than commenting, regulating interaction, or expressing emotional states. A systematic review of interventions for comments identified fourteen single-case studies focused on children with communication disabilities; strategies included prompt hierarchies, assisted modeling, and direct prompting, applied during shared reading, play, or intensive instruction (Spencer et al., 2025) . The most consistent evidence favored hierarchies of least to most assistance that incorporated assisted modeling during reading, although the small number of studies makes it impossible to recommend any one strategy with certainty.

Commenting fulfills functions that a request does not: sharing attention, transferring information, showing interest, and building social closeness. The forms studied included single-symbol utterances, combinations of two symbols, sentence frames, and grammatically correct productions (Spencer et al., 2025) . This variety indicates that progress should not be defined solely by long sentences. Proper observation records whether the person contributes new information, responds to another person’s perspective, or introduces their own topic.

Naturalistic usage data confirm the same limitation. After introducing AAC tools and caregiver training, minimally verbal autistic children showed greater communicative autonomy in natural contexts, but gaps in regulatory and emotional functions persisted (Jiang & Wang, 2025) . Training can increase communication opportunities, but it does not guarantee that the application contains emotional vocabulary, options for disagreeing, or resources for repairing misunderstandings. Autism-focused applications that incorporate emotional expression and progress tracking attempt to address this gap, although published results focus on acceptability by caregivers and teachers, not on independent changes in emotional communication (Reroma et al., 2026) .

Intervention design should progress from narrow functional objectives toward broader communicative repertoires. A graded prompt system can reduce initial errors, gradually withdraw support, and preserve the user’s initiative. The applied description of the least-prompt system proposes adjusting the support to the learning profiles, task demands, and communicative skills of each autistic individual (Ajemigbitse & Mantzoros, 2025) . Thus, the application ceases to be a menu of pre-set responses and becomes a means to learn how to initiate, comment, reject, ask questions, express emotions, and repair conversations.

Personalization and accessibility: adapting technology to diverse bodies, languages, and trajectories

Speech recognition, voice generation, and support for people with dysarthria or ALS

Accessibility depends on how information is received and how a response is generated. For people with dysarthria, automatic speech recognition may preserve some of the natural voice, but its conversational performance should not be taken for granted. In a follow-up study of up to twelve months with three people with ALS, all described lower-than-expected recognition accuracy and considered live captions useful only in certain contexts (Cave, 2024) . The participants appreciated the idea and hoped for improvements, but errors could affect comprehension and lead to unfair attributions of responsibility.

Personalization requires technical training and ongoing support. A participatory project on an application for dysarthric speech involved sixty-six participants, their families, and support teams for at least three months; user-driven testing led to accelerated voice training, simplified navigation, and the incorporation of switch access (Howarth et al., 2024) . This approach shows that accessibility is not simply about recognizing an acoustic signal. It also encompasses physical access, technological confidence, equipment configuration, and the ability to adapt uses to the individual’s chosen activities.

Systems that integrate speech recognition, conversational context, and personalized synthesis aim to produce more expressive messages, including emotional nuances and responses tailored to the interlocutor (Yiwen et al., 2025) . However, this approach stems from an exploratory feasibility study and focus groups with professionals, and therefore does not yet establish sustained benefits for users in real-life conversations. Furthermore, in ALS, the progression of dysarthria necessitates that applications allow for changes in input mode without imposing an abrupt transition to text.

Multilingual and specific design for apraxia, aphasia, autism and dementia

The interface language determines who can use an application independently. My Voice , developed for children with apraxia of speech and also described for aphasia, offers Arabic and English, text-to-speech conversion, and three construction modes: keyboard, phrases, and words (Sweidan et al., 2024; Sweidan et al., 2025) . Survey evaluations after one month of use were positive, but they do not provide comparative measures of functional communication or longitudinal data. The availability of Arabic addresses a documented gap, although it does not equate to complete adaptation to dialect, literacy, culture, or voice preferences.

In dementia, reviews of commercial apps have found a limited selection and insufficient personalization for daily care activities. One evaluation of 27 apps observed variation in engagement, functionality, aesthetics, and information; all were compatible with English speakers, several included other languages, and many had limitations in supporting tasks such as eating or dressing (Muñoz et al., 2024) . Another review of 11 apps found better ratings for ease of use and navigation than for visual appeal, and only one app had been created by a developer with speech-language pathology experience, while another communicated peer-reviewed results (Muñoz et al., 2026) . These findings suggest valuing authorship, evidence, and personalization alongside ease of navigation.

In autism, apraxia, aphasia, and dementia, the same interface cannot be assumed to be suitable. Symbol size, memory load, selection speed, vocabulary, touchscreen access, and text capabilities must be adjusted to individual abilities and preferences. Accessibility, therefore, is a property of the relationship between person, device, language, and activity, not a fixed application label.

From application to social connection: quality of life, literacy, friendship and community participation

Augmentative communication in education, healthcare, and everyday life

Social participation is a distinct outcome from selection accuracy. A systematic review of educational and healthcare interventions identified ten eligible studies from an initial 256 records and found a positive relationship between AAC strategies, communication skills, and quality of life, although the empirical base was limited (Gómez-Asencio et al., 2025) . Educational settings can offer learning opportunities, while healthcare environments require resources for communicating symptoms, preferences, consent, and questions. In both cases, the application must be available at the time of the interaction and understandable by interlocutors who may not know the user.

Literacy expands these possibilities by enabling the production and understanding of messages beyond pictographic repertoires. The literature shows that individuals who need AAC can acquire diverse reading and writing skills when they receive adapted instruction, effective tools, and trained professionals (Light et al., 2025) . Barriers persist when students are excluded from literacy instruction, when spoken responses are required, or when AAC that facilitates the conversion of graphic symbols to text is unavailable. Mobile design should include text-rich environments and materials that allow for participation without oral response.

The community requires a different approach to planning than that used in clinical settings. Location-based interventions increased expressive accuracy and were generalized across community settings among adolescents with complex needs (Han et al., 2025) . This finding supports interfaces linked to routes, shops, transportation, and real-world activities, but it does not, on its own, provide information about satisfaction, safety, or relationship quality. Assessment should combine behavioral observation, accessible self-assessment, and peer evaluation.

Friendship, self-defense, and the right to participate in one’s own decisions

Friendship requires reciprocal communication, choice, and continuity, not just correct answers. A review of 46 studies identified themes related to the formation and maintenance of friendships, peer support, barriers, the effects of AAC, and clinical recommendations (Finke et al., 2025) . Users seek relationships based on shared interests, experiences, and traits, so support should facilitate encounters and activities, not just interactions with professionals or family members. An app can contribute through personal vocabulary, spontaneous messages, and quick access to shared topics, but it cannot replace the social conditions that allow people to meet others.

Self -advocacy requires expressing disagreement, asking for clarification, reporting inappropriate treatment, and participating in decisions about health, education, and support. Self-assessment tools can facilitate this participation, although the corresponding review found 22 measures and gaps in feelings, motives, non-English languages, and psychometric properties (Savolainen & Pilesjö, 2025) . The quality of communicative life is also incompletely assessed: a review of adults with acquired neurological disorders found 10 studies with related instruments, but little information on adaptations and interlocutor support during assessment (Grigoriou et al., 2025) . The next question for technology design is not how many words a person produces, but whether they can express preferences and verify that these have been understood.

Critical synthesis: intervention strategies, controversies, and research gaps

The role of caregivers, modeling, training, and graded support systems

The most consistent results emerge when the app is integrated into an intervention, not when it is delivered as a standalone tool. Caregiver training was associated with greater natural communicative autonomy in minimally verbal autistic children, although it left regulatory and emotional functions unresolved (Jiang & Wang, 2025) . Assisted modeling and prompt hierarchies are supported for teaching feedback, especially in shared reading, but the review identified few studies and limitations in the number of points per phase, blinding, and the description of procedures (Spencer et al., 2025) . The app should therefore be accompanied by explicit decisions about when to model, when to wait, and when to withdraw support.

The system with the fewest prompts offers a sequence for adjusting support to the task and the autistic person’s competence (Ajemigbitse & Mantzoros, 2025) . Its clinical value depends on professionals recording independence, initiative, and transfer to different interlocutors. The MPR-based study showed maintenance and generalization with a new interventionist, but its single-participant design and three target behaviors do not allow us to conclude that the procedure produces a broad communicative repertoire (Kim & Noh, 2025) . The intervention must distinguish between acquisition, maintenance, generalization, and spontaneity.

Training is also necessary for speech recognition technologies. In ALS, users described lower-than-expected conversational accuracy and needed guidance on system customization and limitations (Cave, 2024) . In the development of an application for dysarthric speech, on-site and remote technical support enabled modifications to the interface, training, and access (Howarth et al., 2024) . These studies converge on a practical conclusion: accessibility includes support, troubleshooting, and protection against individual attribution of technological errors.

Between technological innovation and still fragmentary evidence

The literature combines single-case studies, scoping reviews, market assessments, participatory projects, and exploratory technology proposals. This heterogeneity makes it difficult to directly compare results or determine which application is superior for a given population. A review of mobile applications in Turkey identified 71 and analyzed 36 using a checklist of icons, content, and usability; it found some strengths and many weaknesses (Karal et al., 2025) . In dementia, the scarcity of applications supported by developers with clinical experience and by peer-reviewed studies poses a similar limitation (Muñoz et al., 2026) .

The gaps affect both outcomes and the people involved. There is a need to assess feedback, emotions, regulation, literacy, friendship, self-advocacy, and quality of life, along with requests. Accessible self-assessment tools and measures that incorporate the perspectives of interlocutors are also lacking (Grigoriou et al., 2025; Savolainen & Pilesjö, 2025) . Future research should include AAC users in the design process, compare long-term usage trajectories, and describe linguistic, physical, and cognitive access. Innovation will be clinically defensible when its functions are linked to defined communicative outcomes and decisions expressed by the user.

Conclusion: Design applications that expand expression, not just the emission of words.

Communication focused on the person and participation

AAC applications can increase requests, refusals, requests for help, and community expressions when combined with modeling, prompting, training, and real-world opportunities for use (Han et al., 2025; Kim & Noh, 2025) . Their impact diminishes when only word production is measured and feedback, emotions, conversational repair, literacy, and relationships are neglected. Personalization should include language, voice, motor access, vocabulary, context, and preferences, with support for individuals with dysarthria, ALS, apraxia, aphasia, autism, and dementia (Howarth et al., 2024; Muñoz et al., 2024; Sweidan et al., 2025) .

The ultimate criterion is not that an application produces more messages, but that it allows the individual to decide what to communicate, with whom, when, and how. To meet this criterion, development must incorporate AAC users into meaningful roles, improve self-assessment measures, and study sustained social outcomes (Light et al., 2025; Savolainen & Pilesjö, 2025) . Appropriate technology expands expression because it supports initiative, identity, and the right to participate in one’s own decisions.

References

• Ajemigbitse, AR, & Mantzoros, T. (2025). Applying system of least prompts to augmentative and alternative communication intervention. Intervention in School and Clinic , 61 (4), 172–180. https://doi.org/10.1177/10534512251408794
• Cave, R. (2024). How people living with amyotrophic lateral sclerosis use personalized automatic speech recognition technology to support communication. Journal of Speech Language and Hearing Research , 67 (11), 4186–4202. https://doi.org/10.1044/2024_jslhr-24-00097
• Finke, EH, Therrien, MCS, Azios, JH, & McElfresh, JM (2025). We all need at least one friend who understands what we do not say: A scoping review of friendship and augmentative and alternative communication. American Journal of Speech-Language Pathology , 34 (2), 931–958. https://doi.org/10.1044/2024_ajslp-24-00251
• Gómez-Asencio, D., Gago-Valiente, F.-J., Arana-Álvarez, C., Sierra, MIM, López-López, M.-J., & Sánchez, EM (2025). Pedagogical interventions in augmentative and alternative communication (AAC): therapeutic benefits in healthcare and quality of life. a systematic review. Frontiers in Communication , 10 . https://doi.org/10.3389/fcomm.2025.1584169
• Grigoriou, S., LaValley, M., Κολά, Ε., Koul, R., & Pampoulou, E. (2025). Communication-related quality of life in adults with acquired neurological disorders who use augmentative and alternative communication: a scoping review of existing tools. Augmentative and Alternative Communication , 42 (3), 299–310. https://doi.org/10.1080/07434618.2025.2516146
• Han, N.K., Baik, K., Hong, K.-H., Lee, H.J., & Kim, Y.T. (2025). Comparing location-based and category-based AAC interventions for improving communication skills in adolescents with complex communication needs. Journal of Special Education Technology , 41 (1), 43–56. https://doi.org/10.1177/01626434251321105
• Howarth, E., Vabulas, G., Connolly, S., Green, D., & Smolley, S. (2024). Developing accessible speech technology with users with dysarthric speech. Assistive Technology , 1–8. https://doi.org/10.1080/10400435.2024.2328082
• Jiang, Y., & Wang, T. (2025). Patterns of AAC use and communicative functions in minimally verbal autistic children following introduction of AAC tools and caregiver training: A corpus-based analysis. Research in Developmental Disabilities , 168 , 105188. https://doi.org/10.1016/j.ridd.2025.105188
• Karal, M.A., Özgüç, C.S., Hopcan, S., & Polat, E. (2025). Evaluation of mobile AAC applications that support social-communication of individuals with complex communication needs in Türkiye. Speech Language and Hearing , 28 (1). https://doi.org/10.1080/2050571x.2025.2497684
• Kim, J., & Noh, J. (2025). The effects of MPR-Based AAC app intervention on functional communication skills of children with autism spectrum disorder. Korean Association for Behavior Analysis , 12 (3), 21–40. https://doi.org/10.22874/kaba.2025.12.3.2
• Light, J., Holyfield, C., McNaughton, D., Nieder, D., & Preece, J. (2025). Overcoming barriers to literacy for individuals who need or use AAC: lived experiences, key research findings, and future directions. Augmentative and Alternative Communication , 41 (3), 283–294. https://doi.org/10.1080/07434618.2025.2502032
• Muñoz, MH, Okine, J., Brown, EL, & Ruggiano, N. (2024). The state of the art in mobile augmentative and alternative communication designed for people with dementia: Evaluation of features and functions. Alzheimer’s & Dementia , 20 (S10). https://doi.org/10.1002/alz.094244
• Muñoz, MH, Okine, J., Brown, EL, & Ruggiano, N. (2026). Mapping augmentative and alternative communication applications for dementia care. Journal of Applied Gerontology , 7334648261451134. https://doi.org/10.1177/07334648261451134
• Reroma, JD, Cuizon, JC, Bolivar, JA, & Ortiz, E. (2026). SpeakBright: A mobile application for augmentative and alternative communication (AAC) for nonverbal ASD learners (pp. 209–212). https://doi.org/10.1109/icct-pacific69083.2026.11519037
• Savolainen, I., & Pilesjö, M.S. (2025). Self‐Assessment instruments related to functional ability for people who use augmentative and alternative communication—A scoping review. International Journal of Language & Communication Disorders , 61 (1), e70171. https://doi.org/10.1111/1460-6984.70171
• Spencer, TD, Tönsing, KM, & Dada, S. (2025). Augmentative and alternative communication (AAC) interventions that promote commenting: a systematic review. Augmentative and Alternative Communication , 42 (3), 285–298. https://doi.org/10.1080/07434618.2025.2477694
• Sweidan, SZ, Abdel-Qader, AA, & Darabkh, KA (2024). Children with aphasia (inability to speak well): An android application for voice augmentation and alternative communication. In International Conference on Information Technology Systems and Innovation (pp. 463–470). https://doi.org/10.1109/icitsi65188.2024.10929394
• Sweidan, SZ, Abdel-Qader, AA, & Darabkh, KA (2025). My voice: An augmentative and alternative communication android app for children with apraxia in arab countries. International Journal of Interactive Mobile Technologies (iJIM) , 19 (04), 65–92. https://doi.org/10.3991/ijim.v19i04.52125
• Yiwen, X., Monideep, C., Tianyi, Z., Katelyn, E., Aanchan, M., & Mirjana, P. (2025). Your voice is your voice: Supporting Self-expression through Speech Generation and LLMs in Augmented and Alternative Communication. ArXiv.Org . https://arxiv.org/pdf/2503.17479