AI Portfolio Website Prompts for Students: Show Your Work
Useful AI portfolio website prompts start with evidence of your work: what you made, why you made it, your contribution and what a visitor can inspect. Ask the AI to organize those facts into a small, readable website. Do not ask it to invent experience, clients, project results or testimonials to fill empty sections.
This guide is for college students and adult beginners building a first project portfolio. It provides original practice prompts, not extracts from Hamraah Foundation's paid toolkit. A portfolio can explain your work; it does not guarantee an internship, admission, job or assessment result.
Prepare a project evidence sheet before prompting
Choose one completed project or a clearly labelled work in progress. Write a short factual record: its purpose, intended user, your role, the tools used, a decision you made, a difficulty you encountered and what you would change next. Add only links and screenshots you have permission to publish.
If it was a group assignment, separate the group's result from your own contribution. If a teacher supplied the brief or starting code, say so where relevant. If AI helped draft text or code, describe that assistance honestly rather than presenting every generated component as your independent work.
For a fictional example, a student might have built a library catalogue prototype with sample records. They can show how searching and empty results work. They cannot claim that a real library adopted it or that it saved staff time unless that actually happened and can be supported.
Prompt 1: turn verified notes into a content plan
Plan a simple portfolio for a college student using only the project notes below. The main visitor should understand what I contributed and be able to open the project evidence. Suggest a home introduction, one project case study and a contact section. Mark missing facts as questions for me. Do not invent employers, qualifications, results, testimonials or additional projects. Return a content outline before writing website code.
Paste your evidence sheet after the prompt. Review the outline first. Delete sections that have no useful content instead of letting the generator add fictional experience. One carefully explained project can be easier to assess than several identical cards with vague descriptions.
Decide the visitor's main action. For a first portfolio, it might be opening the case study or viewing a working demonstration. Keep that action prominent and make secondary links easy to find. Avoid collecting visitor information when a simple public contact method is enough.
Prompt 2: build a case study that explains decisions
Write a draft case study from these approved facts. Use sections for the problem, my contribution, the approach, evidence, limitations and next steps. Distinguish a prototype from a deployed service. Keep team contributions separate from mine. Use plain language and leave unsupported outcomes out. For every claim about a result, identify the evidence supplied or flag it for my review.
Read the draft against your original notes line by line. A phrase such as “improved usability” needs a basis; if you only changed a navigation label, describe the change and why you made it. Explain an unresolved issue when it helps a reviewer understand your reasoning.
A useful caption tells the visitor what a screenshot demonstrates. For the fictional catalogue, a caption might explain the empty-search state and the next action it offers. A screenshot of an attractive screen alone does not establish accessibility, performance or successful use by real users.
Prompt 3: generate a clear, accessible layout
Build the approved content as a responsive portfolio. Use semantic page regions, a logical heading hierarchy and descriptive link text. Keep reading order clear on small screens. Make navigation and interactive controls usable with a keyboard. Use real buttons and links for their correct purposes. Do not add decorative animation, fake contact submissions or invented content. Explain which behaviours still need manual testing.
The W3C page structure tutorial explains how headings and page regions help people understand and navigate content. Use that guidance when reviewing the generated layout. Asking for accessibility is a starting instruction, not proof that the finished site meets an accessibility standard.
Keep project evidence near the claim it supports. Use readable text rather than putting essential explanations only inside images. For each image, decide whether it conveys information and write an appropriate text alternative. Do not add a heavy visual effect if it obscures the work you want someone to examine.
Prompt 4: review truthfulness, privacy and working links
Review this portfolio against my approved evidence sheet. List unsupported claims, ambiguous ownership statements, placeholder links and personal information that may not need to be public. Check that project labels distinguish practice work from real client work. Do not change facts automatically. Return a short issue list with the affected section and a suggested correction for my approval.
Follow that review with your own checks. Open each project and repository link in a signed-out window: a link that works only for you is not useful public evidence. Remove tokens, passwords, private documents and personal details from screenshots and downloadable files. Obtain permission before identifying collaborators or publishing their work.
Try navigation with the keyboard and read the page on a narrow screen. Check whether the contact method works as described, without sending unsolicited test messages. The W3C Easy Checks provide an introductory accessibility review; they are not a complete audit.
Worked example: turn project notes into a defensible case study
Try this exercise with a small project you can explain. The example is fictional practice work, not a Hamraah student result or an excerpt from a paid toolkit. Its purpose is to show what evidence you should supply before asking an AI tool to write a portfolio page.
Approved notes: I built a local book-catalogue prototype using sample records. I added a title search and an empty-results message. I tested a matching title, a title with no match and a cleared search. I have not tested it with library users, connected a real catalogue or measured time saved. I wrote the interface; a tutorial supplied the starting data structure.
Use only the approved notes above. Draft a short portfolio case study with the problem, my contribution, a decision I made, the checks I performed and limitations. Do not invent a client, users, performance measurements or employment outcomes. Label it as a practice prototype. After the draft, list any missing evidence that would be needed for stronger claims. Keep that list separate from the public case study.
A defensible draft could explain: “I built a practice catalogue to explore title search. I added an empty-results message so the interface explains when the sample data has no match. My checks covered a matching title, no match and clearing the search. The prototype has not been evaluated with library users.” This describes observable work without turning a personal exercise into a deployed service.
If the generated version says it “transformed library access” or “saved staff hours,” reject those phrases: the supplied notes do not support them. Also check ownership. Crediting a tutorial does not erase your contribution, and describing your contribution does not make the tutorial's work yours.
Now choose evidence for the page. A screenshot can demonstrate the empty-results message; a public demonstration can let someone try the sample search. Neither proves how real users would respond. Caption each item with what it actually shows, and remove private data from anything you publish.
Finally, ask the tool to organize the approved case study using the W3C guidance on headings and page structure. Check the page yourself. A reviewer should be able to find your contribution, evidence and limitations without watching an animation or guessing what a project card means. Repeat the process with your own notes before considering a website prompt toolkit for a larger project.
Choose your next learning step
If you are still learning how to give an AI tool a clear brief, start with our beginner website prompting guide. If you are considering a paid resource, use the prompt toolkit selection guide to assess whether it matches your project and workflow.
Hamraah AI School's website prompt toolkit is a separate resource to explore for website-building prompts. Check its current contents and terms before choosing it. This article does not describe a student course, mentoring service or employment program. Your next step is to build and review a portfolio you can explain accurately, using a tool or resource that fits that task.