Journals / Journal of Computational Design / 2026 / Cilt: 7 Sayı: 2

Intent-Output Alignment in AI-Assisted Interface Design: A Professional Evaluation of Six Prompting Strategies

Yapay Zekâ ile Arayüz Tasarımında Niyet-Çıktı Uyumu: Altı İstem Stratejisinin Profesyonel Değerlendirmesi

Veysel Emekçi(IZMIR INSTITUTE OF TECHNOLOGY)0009-0006-7136-5340
Pages
237–265
Publication Date
2026-09-30
🏛️Bibliographic metadata and full text access for this article are provided via DergiPark.
DergiPark Record ↗
✨

AI Summary & Key Takeaways

AI Insight
⚡ TL;DR(Too Long; Didn't Read)

Given the single session, the five participants and the single output generated per strategy, the findings should be read as context-specific and exploratory rather than generalisable.

📌 3 Key Takeaways

🎯
Objective & Hypothesis:Generative AI tools such as ChatGPT and Claude are reshaping design practice, and AI now takes part in the early phases of the design process, including idea generation, concept development and decision-making.
📊
Key Findings & Data:Because the quality of an AI-generated output depends on how the designer communicates with the model, prompt design has become the primary communication channel between designer and AI.
💡
Conclusion & Impact:Given the single session, the five participants and the single output generated per strategy, the findings should be read as context-specific and exploratory rather than generalisable.
🔬Design:Modeling & Simulation
📍Scope:Comprehensive Sample
⚙️Variables:İstem Stratejileri · Kullanıcı Arayüzü Tasarımı · Niyet-Çıktı Uyumu

Abstract

Generative AI tools such as ChatGPT and Claude are reshaping design practice, and AI now takes part in the early phases of the design process, including idea generation, concept development and decision-making. Because the quality of an AI-generated output depends on how the designer communicates with the model, prompt design has become the primary communication channel between designer and AI. Existing prompt engineering scholarship examines how the structure of a prompt affects model performance — through chain-of-thought reasoning steps or zero-shot reasoning triggers; the question of how far the theoretical intent behind a prompt survives into a visual design output has received little systematic attention. This exploratory study examines six prompting strategies through a fictional task management application, TaskFlow. Each strategy rests on a distinct theoretical foundation: Basic Prompt (zero-shot prompting), Design Brief (chain-ofthought), Simplification Focus (cognitive load theory), Inspiration Based (abductive thinking), Problem Focused (human-centred design) and Emotion/Tone Based (emotional design). For each strategy Claude Opus 4.5 produced a written design definition, which was converted into a functional prototype interface through Figma Make AI (Default). Five professional UI/UX designers evaluated the seven outputs (one reference interface and six strategy outputs) in a single blind focus group session, using applicability, user-centredness and inspirational quality as criteria. None of the outputs was adopted as a whole by any participant. The two strategies whose theoretical intent became most directly visible in the output, Emotion/Tone Based and Problem Focused, attracted the most positive assessments across all three criteria, expressing their intent through concrete visual counterparts. By contrast, the Design Brief strategy, which supplied the most extensive context, did not control the direction of the output, which drifted from task management towards habit tracking; within this dataset the amount of information supplied does not appear to guarantee the realisation of intent. As the intent of a prompt narrowed, the user base addressed by the output tended to narrow with it, although the Emotion/Tone Based strategy stood outside this tendency, possibly because its emotional design framework does not depend on a specific persona. None of the five participants framed AI as a co-designer or creative partner; all positioned it as a tool operating under the designer's judgement, which sits in tension with a literature that often casts AI as an active actor. Intent-output alignment was not a framework tested in advance but a pattern that emerged during the analysis of the focus group data; the study proposes it, across the three dimensions of presence, direction and boundary, as an analytical axis that warrants examination in further research. Given the single session, the five participants and the single output generated per strategy, the findings should be read as context-specific and exploratory rather than generalisable.

Özet

Üretken yapay zekânın (YZ) tasarım pratiğine girişiyle birlikte istem (prompt) tasarımı, tasarımcı ile YZ arasındaki temel iletişim kanalı hâline gelmiştir. Mevcut istem mühendisliği yazını ağırlıklı olarak verilen bilginin miktarı ve biçimi üzerine yoğunlaşmakta; istemde ifade edilen teorik niyetin çıktıda ne ölçüde gerçekleştiği sorusu ise, özellikle görsel arayüz tasarımı bağlamında, sistematik biçimde ele alınmamaktadır. Bu keşfedici çalışma altı farklı istem stratejisinin tasarım çıktıları üzerindeki etkisini incelemektedir. Kurgusal bir görev yönetimi uygulaması (TaskFlow) bağlamında her strateji için Claude Opus 4.5 modeli yazılı bir tasarım tanımı üretmiş, bu tanımlar Figma Make AI (Default) modeline aktarılarak işlevsel prototip arayüzlere dönüştürülmüştür. Çıktılar, farklı sektörlerden beş profesyonel UI/UX (Kullanıcı Arayüzü/Kullanıcı Deneyimi) tasarımcısının katıldığı tek oturumluk bir odak grup çalışmasında, uygulanabilirlik, kullanıcı odaklılık ve ilham vericilik ölçütleri üzerinden kör değerlendirmeye tabi tutulmuştur. Katılımcıların hiçbiri çıktıların herhangi birini bütüncül olarak benimsememiştir. Teorik niyeti çıktıda en açık biçimde görünür hâle gelen iki strateji (Duygu/Ton Odaklı ve Sorun Odaklı) üç ölçütte de en olumlu değerlendirmeleri almıştır. Buna karşılık en kapsamlı bağlamı sağlayan Tasarım Özeti stratejisi çıktının yönünü kontrol edememiş, tasarım görev yönetiminden alışkanlık takibine kaymıştır; bu veri kümesinde sağlanan bilgi miktarı tek başına niyetin gerçekleşmesini güvence altına almış görünmemektedir. Yönlendirmenin niyeti daraldıkça çıktının hitap ettiği kullanıcı tabanı da daralma eğilimine girmiş, Duygu/Ton Odaklı strateji bu eğilimin dışında kalmıştır. Katılımcıların tamamı YZ'yı tasarımcının yargısı altında çalışan bir araç çerçevesinde değerlendirmiş, hiçbir katılımcı YZ'yı ortak tasarımcı ya da yaratıcı partner olarak konumlandırmamıştır. Niyet-çıktı uyumu, araştırma öncesinde sınanmak üzere kurulmuş bir çerçeve değil, odak grup verisinin çözümlenmesi sırasında beliren bir örüntüdür; çalışma bu örüntüyü varlık, yön ve sınır olmak üzere üç boyutta işleyen ve ileri araştırmalarda sınanmayı bekleyen bir analitik eksen olarak önermektedir. Tek oturum, beş katılımcı ve strateji başına tek çıktı koşulları göz önünde bulundurulduğunda bulgular genellenebilir değil, bağlama özgü ve keşfedici niteliktedir.

🔖

Cite This Article

Emekçi, V. (2026). Intent-Output Alignment in AI-Assisted Interface Design: A Professional Evaluation of Six Prompting Strategies. Journal of Computational Design, Cilt: 7 Sayı: 2, 237–265. https://doi.org/10.53710/jcode.1993120

📄 Full Text PDF Viewer

📄

PDF görüntüleyici hazırlanıyor...

References (18)

  1. Ahmetoglu, Y., Brumby, D. P., & Cox, A. L. (2024). Bridging the gap between time management research and task management app design: A study on the integration of planning fallacy mitigation strategies. In Proceedings of the 3rd Annual Meeting of the Symposium on Human-Computer Interaction for Work (pp. 1–14). Association for Computing Machinery. https://doi.org/10.1145/3663384.3663404
  2. Alain, G., Crick, J., Snead, E., Quatman-Yates, C. C., & Quatman, C. E. (2025). Evaluating user interactions and adoption patterns of generative AI in health care occupations using Claude: Cross-sectional study. Journal of Medical Internet Research, 27, Article e73918. https://doi.org/10.2196/73918
  3. Chandrasekera, T., Hosseini, Z., & Perera, U. (2025). Can artificial intelligence support creativity in early design processes? International Journal of Architectural Computing, 23(1), 122–136. https://doi.org/10.1177/14780771241254637
  4. Chompunuch, S., & Lubart, T. (2025). AI as a helper: Leveraging generative AI tools across common parts of the creative process. Journal of Intelligence, 13(5), Article 57. https://doi.org/10.3390/jintelligence13050057
  5. Choudhury, M. M., Eisenbart, B., & Kuys, B. (2025). Artificial intelligence (AI) in the design process: A review and analysis on generative AI perspectives. Proceedings of the Design Society, 5, 631–640. https://doi.org/10.1017/pds.2025.10077
  6. Dang, H., Mecke, L., Lehmann, F., Goller, S., & Buschek, D. (2022). How to prompt? Opportunities and challenges of zero- and few-shot learning for human-AI interaction in creative applications of generative models. arXiv. https://doi.org/10.48550/arXiv.2209.01390
  7. Haase, J., & Pokutta, S. (2026). Human-AI cocreativity: Exploring synergies across levels of creative collaboration. In M. J. Worwood & J. C. Kaufman (Eds.), Generative artificial intelligence and creativity (pp. 205–221). Academic Press. https://doi.org/10.1016/B978-0-44334073-4.00009-5
  8. Kojima, T., Gu, S. S., Reid, M., Matsuo, Y., & Iwasawa, Y. (2022). Large language models are zero-shot reasoners. Advances in Neural Information Processing Systems, 35, 22199–22213. https://proceedings.neurips.cc/paper_files/paper/2022/hash/8bb0d 291acd4acf06ef112099c16f326-Abstract-Conference.html
  9. Krueger, R. A., & Casey, M. A. (2015). Focus groups: A practical guide for applied research (5th ed.). SAGE Publications.
  10. Medeiros, K. E., Cropley, D. H., Marrone, R. L., & Reiter-Palmon, R. (2025). Human-AI co-creativity: Does ChatGPT make us more creative? The Journal of Creative Behavior, 59(2), Article e70022. https://doi.org/10.1002/jocb.70022
  11. Norman, D. A. (1988). The psychology of everyday things. Basic Books.
  12. Norman, D. A. (2004). Emotional design: Why we love (or hate) everyday things. Basic Books.
  13. Subramonyam, H., Pea, R., Pondoc, C., Agrawala, M., & Seifert, C. (2024). Bridging the gulf of envisioning: Cognitive challenges in prompt based interactions with LLMs. In Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems (pp. 1–19). Association for Computing Machinery. https://doi.org/10.1145/3613904.3642754
  14. Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12(2), 257–285. https://doi.org/10.1207/s15516709cog1202_4
  15. Tang, Y.-c., Huang, J.-j., Yao, M.-t., Wei, J., Li, W., He, Y.-x., & Li, Z.-j. (2019). A review of design intelligence: Progress, problems, and challenges. Frontiers of Information Technology & Electronic Engineering, 20(12), 1595–1617. https://doi.org/10.1631/FITEE.1900398
  16. Verganti, R., Vendraminelli, L., & Iansiti, M. (2020). Innovation and design in the age of artificial intelligence. Journal of Product Innovation Management, 37(3), 212–227. https://doi.org/10.1111/jpim.12523
  17. Wei, J., Wang, X., Schuurmans, D., Bosma, M., Ichter, B., Xia, F., Chi, E. H., Le, Q. V., & Zhou, D. (2022). Chain-of-thought prompting elicits reasoning in large language models. Advances in Neural Information Processing Systems, 35, 24824–24837. https://proceedings.neurips.cc/paper_files/paper/2022/hash/9d560 9613524ecf4f15af0f7b31abca4-Abstract-Conference.html
  18. Wilson, H. J., & Daugherty, P. R. (2018). Collaborative intelligence: Humans and AI are joining forces. Harvard Business Review, 96(4), 114–123. https://hbr.org/2018/07/collaborative-intelligence-humans-and-aiare-joining-forces