Artflow - AI Portfolio Planner
An AI led web application that helps young artists to plan a portfolio.
Team Next Frame Collective
Members Siwen Pang, Zhengyang Xie, Deqi Kong, Zhaonan Xu, Zichong Xu
Target Learning Audience 10/14/2025
High school & College students who are making an art portfolio
New Introduction and Rational for AI Assistance updated 10/14/2025
For portfolio projects, users do not know which ones to choose in order to best showcase their personal ability, and do not understand the preferences and focus of the viewers of the portfolio, resulting in a lack of targeted production ideas. For those who have never created a portfolio, it is easy to affect professionalism due to disorderly layout and outdated aesthetics. The statement content does not know how to distinguish the priority and reading order of information through font size, color, image retention, etc., resulting in poor page reading experience and difficulty for viewers of the portfolio to quickly grasp the key points.
AI Portfolio Planner Chatbot, the main goal is to serve high school and college students who are creating portfolios, helping them plan and provide feedback on their portfolios, leaving them with more time to refine their artistic creations. Users can use AI to brainstorm and fill in the gaps in design ideas. Hence, AI can serve as a scaffold to assist students in organizing design ideas. At the same time, for inexperienced students who may feel lost, AI can act as a starting assistant.
Interaction Modality 10/6/2025
An AI led application that creates personalized art portfolio plans.
Learning Objective 10/6/2025
Improve planning skills
Learn about art categories
Discover personal strengths and specialties
Literature Review 10/14/2025
Account 1: The Enhanced Role of Generative AI in Artistic Creation
Our learning goal is to enable students to more conveniently and accurately master the planning and process of creating art portfolios through generative AI–assisted learning.
The reason for choosing this learning goal is that in today’s educational environment, even within the high school student population, there is significant variation in how students understand art, recognize their own strengths and interests, and approach portfolio creation. Many students feel confused and anxious about the academic journey ahead and are uncertain how to present their artistic capabilities in a way that gains recognition from schools. Traditional portfolio planning often requires students to explore unfamiliar processes, categories, and research paths on their own, which can be overwhelming and inefficient. Our product serves as an intelligent assistant that provides personalized guidance tailored to each student’s background, level of understanding, and creative goals. This individualized support not only reduces the inefficiency caused by trial-and-error in the early stages but also significantly accelerates the research and planning process, enabling students to focus more of their time and energy on refining their work and developing their artistic expression. In the early stages of product development, as creators, we also need evidence that the learning goals we set are meaningful and achievable. The following research offers a strong example that supports our approach. He, Y., & Zhang, S. (2025). Enhancing art creation through AI-based generative adversarial networks in educational auxiliary systems. Scientific Reports, August 2025. At the beginning of this paper, the authors note that “Creative art education increasingly requires interactive, personalized tools that support students in developing aesthetic expression and technical skills” (He & Zhang, 2025), which aligns closely with the core idea behind our product. This paper introduces an AI-enhanced educational assistance system based on Generative Adversarial Networks (GANs) with the following specific functions: 1. A modular AI structure that generates sketches or style transformations based on students’ descriptions. 2. Multi-mode user interaction capable of analyzing student-drawn images and providing feedback and suggestions for improvement. 3. A progressive preview of students’ next steps and design thinking. 4. A learning management system compatible with existing platforms (such as Canvas and Moodle) that supports collaborative learning. Compared with the product we aim to develop, the system described in the paper offers more features and broader functionality. It can iterate continuously based on model training, balance diverse artistic requirements, and deliver targeted feedback in real time. Furthermore, the authors emphasize that the system is suitable for automatic deployment in remote learning environments, supports interdisciplinary applications, and operates with extremely low training costs (He & Zhang, 2025). The system has already been validated through classroom experiments involving 60 undergraduate students. Compared with traditional digital art tools, students who used the AI-based system achieved an average score increase of 35.4%, and their participation levels rose by 42.7% (He & Zhang, 2025). Overall, generative AI can offer substantial support to students in the field of art. The success of such models highlights their strong potential in the educational sector and their ability to transform how students plan, create, and refine their artistic portfolios.Account 2
expand
In The Integration of AI in Design Thinking for Enhancing Student Creativity and Critical Thinking in Digital Media Learning (Hashim et al., 2025), for students creating portfolios. The study highlights that beginners face significant learning hurdles in conceptualizing designs from scratch, as well as coherently describing and structuring portfolio ideas using professional terminology. Researchers discovered that while many learners are familiar with portfolios and possess numerous creative ideas, they struggle to articulate how to plan and execute an effective portfolio. Artificial intelligence tools can assist learners by providing inspiration, helping them overcome the initial hurdles of starting and building a portfolio. The paper mentions that AI excels at refining ideas and translating them into clear and coherent concepts. The author concludes that AI scaffolding directly addresses beginners' cognitive bottlenecks, effectively translating abstract creative concepts. However, the article cautions that without reflection, students may over-reliance on AI-generated expressions, potentially obscuring their authentic ideas.Needfinding Reports 10/14/2025
Methodology
We plan to use surveys and affinity maps to collect data, which is a very intuitive and easy to study way of data collection and analysis. We will first create a survey and distribute it, mainly targeting high school and college students who are creating or preparing to create art portfolios. They may be confused about the different requirements of different schools, have some questions about their own art styles that they are not good at, or are halfway through and don’t know what to do next. These questions will be covered in our survey, and the answers of the respondents will become the basis for our data analysis. After successfully collecting a sufficient number of responses, we will start analyzing the frequency of various multiple-choice answers and the content of open-ended questions, and classify these answers. Affinity map is a good choice, which allows us to post each observed situation in the designated classification area like sticky notes. This makes data analysis more convenient and intuitive, and we can quickly identify the common needs and areas of disagreement of the respondents, and provide improvement strategies for our products based on this data.
Survey Result
Through this survey, we identified several significant issues: While most people create portfolios primarily for school applications or job searches, many remain unclear about how to tailor their content to different objectives. Many think about their portfolio descriptions as lacking logical coherence and depth, and they also lack awareness of industry standards.
When creating a portfolio, the most troublesome and time-consuming parts are writing reflective summaries, organizing work, and obtaining feedback from peers or experts. Many also mention struggling with layout design, unsure how to make their portfolio visually appealing while meeting professional standards. Regarding AI, most hope it can help structure their portfolio, check alignment with industry trends, and select layout templates. Ultimately, the open-ended responses reveal a common pain point: “having ideas but not knowing how to turn them into a visually appealing portfolio,” or “not knowing where to even begin.” This indicates that many lack clarity on portfolio direction, have insufficient depth in their content, struggle with design execution, and are hoping AI can provide more assistance.
Key Findings
Clear purpose, but unclear direction: Most people create portfolios for school applications or job hunting, yet they don’t know how to tailor content for different purposes (e.g., job applications vs. school applications). As a result, the focus of their portfolios often lacks specificity.
Lack of depth and illogical structure are common issues: Many people think about how their portfolio project descriptions are too superficial and lack clear logic, and they struggle to make the content appear more ‘professional,’ which makes the work seem unprofessional.
Writing reflections and getting feedback are the most time-consuming tasks: Everyone thinks about the hardest parts being writing project reflections (not knowing what to write) and obtaining external feedback (no one to help review), with these two steps taking the most time.
Layout and aesthetics are major pain points: Many people mention that “the real headache is the layout,” unsure how to make their portfolios both visually appealing and well-organized. Many have the content, but don’t know how to present it effectively.
Research Question 10/6/2025
In what ways does the AI planner help students identify their personal artistic strengths and specialties, and how accurate are these identifications from the learners’ perspectives?
Design Feature Map 10/21/2025
| Needs identified | Design features | Change in status quo | If AI is being used, what is the AI doing? | What data do you need to prototype this? | How are you prototyping this? | Ethical or societal implications | How would you measure effectiveness | Paint me a utopia | |
|---|---|---|---|---|---|---|---|---|---|
| 1 | Students don’t understand what a portfolio is or why it’s needed. | AI explains key concepts: what a portfolio is, why it’s important, what reviewers look for. | Users gain a clear understanding of purpose and expectations. | Natural language explanation and personalized Q&A. | User goals, background, and field. | Use LLM to generate explanations and examples. | Risk of biased or inaccurate information. | (1) Concept quiz results. (2) User feedback on clarity. | Users feel confident and clearly understand the purpose and structure of a portfolio. |
| 2 | Students don’t know what field or direction to pursue. | AI helps users explore interests and suggests directions. | Users can identify a suitable field (e.g., Fine Art, Digital Media). | Conversational recommendation and interest analysis. | Chat history, user preference data. | Simulate Q&A exploration flow. | Biased recommendations or over-influence from AI. | (1) Direction selection rate. (2) Self-reported confidence. | Users quickly discover the most suitable artistic direction for them. |
| 3 | Students can’t plan a portfolio structure. | AI guides users to build a personalized plan (e.g., personal vs. school application). | Users can generate a step-by-step plan tailored to their goal. | Plan generation and scheduling. | User goals and timeline data. | AI-generated outlines and step breakdowns. | Plans may assume conditions like time available, mental state, etc. | (1) Execution rate. (2) Satisfaction survey. | Users know exactly what to do at each stage, reducing confusion and procrastination. |
| 4 | Students don’t know if their work meets expectations. | AI analyzes uploaded work and provides feedback on quality. | Users receive detailed analysis and scoring on their work. | Image analysis and evaluation feedback. | Uploaded portfolio files. | AI feedback generation. | Privacy and copyright issues. | (1) Alignment with expert ratings. (2) Feedback usefulness. | Users receive feedback comparable to professional critique. |
| 5 | Students struggle with consistency and style coherence. | AI analyzes style and genre consistency across the portfolio. | Users can adjust and refine their work for stronger thematic focus. | Classification and semantic analysis. | Metadata, style tags, and uploaded works. | LLM explanations. | AI may reinforce stereotypes or misclassify genres. | (1) Improved consistency ratings. (2) User self-evaluation. | Users produce cohesive, professional-looking portfolios. |
| 6 | Students lack examples of successful portfolios. | The tool recommends relevant example works for inspiration. | Users can benchmark and draw inspiration from high-quality work. | Retrieval and recommendation systems. | Open-source portfolio datasets. | Search engine + recommendation pipeline. | Copyright and source transparency concerns. | Inspiration feedback. | Users quickly find references and improve their portfolio direction. |
| 7 | Students forget previous feedback or edits. | The tool automatically saves version history and conversations. | Users can revisit past iterations and continue editing. | Context memory and storage management. | Edit logs and conversation history. | Database + versioning system. | Data ownership and privacy concerns. | (1) Frequency of history use. (2) User satisfaction with tracking. | Users can seamlessly resume work and track progress over time. |
| 8 | Students fear AI might replace their creative process. | Clear ethical declaration: AI does not generate art, only provides assistance. | Users trust and engage with the tool confidently. | Transparency and usage restrictions. | User interactions and system logs. | Interface messaging and policy prompts. | Misunderstanding AI’s role and over-reliance risks. | (1) Trust survey. (2) Continued usage rate. | Users feel AI is a supportive assistant, not a replacement. |
| 9 | Students want immediate feedback on their work. | The tool generates instant portfolio analysis reports. | Users see feedback within minutes instead of days or weeks. | Automated analysis and real-time reporting. | Uploaded files and metadata. | Rapid evaluation pipeline + LLM output. | Risks of inaccurate instant feedback or misuse of sensitive data. | Feedback turnaround time. | Users receive near-instant feedback and can iterate immediately. |
Prototype
References
Dhillon, P. S., Molaei, S., Li, J., Golub, M., Zheng, S., & Robert, L. P. (2024, May). Shaping human-AI collaboration: Varied scaffolding levels in co-writing with language models. In Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems (pp. 1-18).
Khadka, T. (2012). Writing proficiency of higher secondary level students (Doctoral dissertation, Department of English Education).
Boynagryan, T., & Tshngryan, A. (2024). AI Writing Assistant: A Comprehensive Study. American University Of Armenia.
He, Y., & Zhang, S. (2025). Enhancing art creation through AI-based generative adversarial networks in educational auxiliary system. Scientific Reports, 15, Article 29202. https://doi.org/10.1038/s41598-025-14164-z
Hashim, M.E.A.H. (2025). The Integration of AI in Design Thinking for Enhancing Student Creativity and Critical Thinking.
Team Contribution
Siwen created the following: GitHub layout & updates, initial project ideas, XMind logic chart, and provided initial concepts. Zhaonan purchased Xmind Team Plan so we can collaborate on the logic chart of the app together. We brainstormed together and finalized the idenitified learning needs and rational for AI assistance. Deqi and Zichong drafted those sections and the references. Zhaonan, Zhengyang, and Siwen then revised the written sections and refined them.
We first had a group meeting to assign tasks to each member. Zhengyang and Zichong searched for relevant cases and literature, and wrote two articles respectively. Deqi created a survey questionnaire based on the target users and surveyed them. Zhengyang and Siwen organized user data and summarized the results. Zhaonan reviewed all the data to be submitted soon. Then, through team meetings, the project was optimized and finally uploaded to GitHub. Zhaonan, Zichong, and Siwen jointly produced the presentation.
History
Idenitified Learning Need 10/6/2025
For those who are making a portfolio for the first time, they are not familiar with the content, production ideas, and layout of the portfolio design statement. The biggest challenge for users is not knowing how to start and not being sure what the target audience wants to see the most. It is also difficult for users to describe the project background, goals, and solutions in professional and insightful language within the industry. The content is often too colloquial and difficult to reflect depth of thinking.
According to research, some people lacked the organization skill even if they had the good ideas upon the problem, they were unable to organize their ideas (Khadka, T, 2012). For portfolio projects, users do not know which ones to choose in order to best showcase their personal ability, and do not understand the preferences and focus of the viewers of the portfolio, resulting in a lack of targeted production ideas. For those who have never created a portfolio, it is easy to affect professionalism due to disorderly layout and outdated aesthetics. The statement content does not know how to distinguish the priority and reading order of information through font size, color, image retention, etc., resulting in poor page reading experience and difficulty for viewers of the portfolio to quickly grasp the key points.
Heading 2
Heading 3
Bold text Italicized text
Blockquote
- First item
- Second item
- Third item
- First item
- Second item
- Third item
inline code