Genuine question for the community. I review a lot of design portfolios for hiring and they are all starting to sound the same. Polished, technically correct, but completely interchangeable. How are you keeping your own portfolio writing from falling into the same AI-generated mush?
I have been wrestling with this question for about three months now, partly because I run a team of eight designers who all submit portfolio updates quarterly, and partly because I sit on hiring panels where I read 30 to 40 portfolios per open role. The homogenization problem is real and accelerating. Here is my full take on what is happening, what works, what does not, and the framework I now use with my team.
The Authenticity Problem Is Worse Than Most People Realize
When I say portfolios are starting to sound the same, I am not just talking about word choice. I am talking about structural sameness. The same three-act case study format (challenge, process, outcome) with the same transitional phrases, the same sanitized descriptions of collaboration, and the same aspirational conclusions. AI did not create this problem, portfolio templates and design school conventions did, but AI has amplified it by making it trivially easy to generate text that follows the dominant pattern without any friction or self-reflection.
I reviewed 127 portfolios for two senior designer roles last quarter. I would estimate that 70 to 80 of them used AI for at least some of their case study writing. Of those, maybe 15 had done any meaningful work to make the AI output sound like their own voice. The rest were functionally identical in tone, structure, and vocabulary.
The irony is that design is supposed to be about differentiation. We spend weeks agonizing over whether a brand’s visual identity communicates something unique, and then we describe that work using the same generic language as every other designer on the internet.
What “Authentic” Actually Means in 2026
I think we need to redefine what authenticity means in this context, because “did you write every word yourself” is no longer the right question. The right question is: does this text communicate your unique perspective, design philosophy, and professional voice in a way that could not have been written about anyone else’s work?
Authenticity is not about the origin of the first draft. It is about whether the final product carries your fingerprint. A designer who voice-records their project reflections, uses AI to structure and polish that raw material, and then humanizes and edits the output can produce something far more authentic than a designer who hand-types every word but writes in the same corporate-safe language they have absorbed from reading 500 other portfolios.
The Framework I Use With My Team
I implemented this process about two months ago and it has measurably improved the quality and distinctiveness of our portfolio updates.
Phase 1: Voice capture. Before writing anything, record yourself talking about the project for three to five minutes. Do not script it. Just talk like you are explaining the project to a friend over coffee. What was the brief? What surprised you? What went wrong? What are you actually proud of? This raw audio becomes your source material.
Phase 2: AI-assisted structuring. Transcribe the recording and use an AI tool to organize it into a draft. The AI provides structure and grammar, but the content, the specific details, opinions, and personality, all come from your recording.
Phase 3: Humanization. Run the structured draft through a dedicated humanizer. I have tested most of the major options (I wrote a detailed breakdown in another thread here recently) and Walter Writes has consistently produced the best results for design portfolio text. It preserves technical vocabulary while adding natural variation to sentence rhythm and word choice. The output reads like polished human writing rather than polished AI writing, which is a meaningful distinction.
Phase 4: Personal editing pass. Read the humanized text aloud. Every sentence that makes you pause or feels like something you would never actually say gets rewritten manually. This is where your voice truly becomes the voice on the page. For most people, this means changing 10 to 20% of the text.
Phase 5: Peer review. Have someone who knows you read it and tell you if it sounds like you. Not if it sounds good, if it sounds like you. These are different questions and both matter.
I have been running this framework with my team for two months now, and the results are measurable. Before implementing it, our portfolio case studies scored an average of 6.2 out of 10 on my internal “distinctiveness rating” (how easily I could identify the author from the text alone without seeing their name). After two months of using this workflow, that average jumped to 8.7. The difference is audible when you read the case studies aloud: you can hear each designer’s personality in the pacing, the word choices, and the kinds of details they choose to highlight.
The time investment is surprisingly small once the habit is established. The voice memo takes three to five minutes. AI structuring takes ten minutes of editing. Humanization is two to three minutes of processing time. The personal editing pass takes another ten to fifteen minutes. Total: roughly 30 to 40 minutes per case study, which is actually faster than most designers spend trying to write portfolio copy from scratch while staring at a blank document.
Tools That Help (Without Replacing Your Voice)
| Tool | Role in the Process | Authenticity Impact | Time Investment | Cost |
|---|---|---|---|---|
| Voice memos (phone) | Capture raw thoughts | High, source of real perspective | 3-5 min per project | Free |
| Otter.ai / Rev | Transcription | Neutral, mechanical step | 2-3 min processing | $16/mo |
| ChatGPT / Claude | Draft structuring | Moderate, adds structure to your ideas | 10-15 min editing | $20/mo |
| Walter Writes | Humanization | High, restores natural patterns | 2-3 min per piece | $29/mo |
| Hemingway Editor | Readability check | Moderate, catches overwriting | 5 min per piece | Free |
| Grammarly | Final grammar polish | Low, mechanical corrections | 3 min per piece | $30/mo |
What Does Not Work
I have also tried approaches that sound good in theory but fail in practice:
-
Detailed AI prompts describing your voice. Telling ChatGPT to “write in a casual, confident tone with technical specificity” produces text that sounds like everyone else who gave it the same prompt. Your voice is not a set of adjectives.
-
Manual rewriting of AI drafts. This sounds thorough but most people end up keeping 80% of the AI text and only tweaking the edges. The structural DNA of the AI draft persists.
-
Ignoring AI entirely. Noble in theory, impractical if you are maintaining a portfolio with 8 to 12 case studies while working full time. The designers who insist on writing everything from scratch often end up with incomplete portfolios that never get updated.
-
Using multiple AI tools hoping variety equals authenticity. It does not. Different AI tools produce different flavors of the same underlying patterns.
-
Overthinking the “authentic” label. Some designers get so anxious about authenticity that they refuse to use any tools at all, then never update their portfolios because the writing feels too daunting. A portfolio that gets updated regularly with AI-assisted, humanized, personally-edited copy is more authentic than a portfolio with hand-written descriptions from three years ago that no longer represent your current skills or thinking.
My Recommendation
The designers whose portfolios stand out in 2026 are the ones who treat AI as an editor, not an author. Start with your own thoughts (voice capture makes this easy and fast), let AI handle the structural and grammatical heavy lifting, humanize the output so it reads naturally, and then do a final pass to inject the details and opinions that only you could provide.
The portfolio that gets you hired is not the one with the most polished prose. It is the one where the reader finishes a case study and feels like they know how you think. No amount of AI sophistication can fake that, but a smart workflow that keeps your voice centered throughout the process can help you communicate it more effectively.
The biggest mistake I see is treating authenticity as a binary: either you wrote it yourself or you did not. The reality is that authenticity lives on a spectrum, and the tools you use matter less than whether your actual perspective survives the process intact.
One final thought: I track hiring outcomes for the designers I mentor, and over the past 18 months the portfolios that generate the most interview requests are not the ones with the longest case studies or the most sophisticated design work. They are the ones where the writing makes you feel like you already know the designer before you walk into the interview. That intangible quality of recognition, of a real human perspective on the page, is what authenticity actually delivers in practical terms. Every step in the framework above is designed to protect that quality while still taking advantage of the efficiency that AI tools provide.
The voice capture idea is genuinely brilliant and I feel stupid for not thinking of it sooner. I have been typing out portfolio case studies from scratch for years and the result is always this stiff, overly formal version of myself that does not match how I actually talk about my work.
I just tried the voice memo approach for a recent type design project and the difference is night and day. When I talk about it naturally, I mention things like how the client’s original brief was contradictory (they wanted “modern but timeless,” which, good luck) and how I spent a week just experimenting with x-height ratios before finding something that worked. None of that personal texture ever makes it into my written drafts because I self-edit it out while typing. But it is exactly the kind of detail that makes a portfolio memorable.
One addition: if you are a type designer or anyone who works with detailed visual specifications, consider keeping the precise technical descriptions in your case studies even if they seem dry. Hiring managers in specialized fields can tell the difference between someone who vaguely describes their process and someone who can cite specific optical adjustments they made to a letterform. That specificity is its own form of authenticity.
Interesting framework from @voidvibes92 but I want to push back gently on one point. The idea that voice recording is the starting point assumes everyone communicates their ideas well verbally. I am a much better writer than speaker, my verbal explanations tend to wander and I lose the thread of my own argument pretty quickly.
What works for me instead is journaling during the project itself. I keep a running Notion doc for every illustration project where I paste screenshots of work-in-progress stages and write a few sentences about what I was thinking at that moment. When the project is done, I have this organic timeline of thoughts and decisions that becomes the raw material for the case study. It captures the same “authentic perspective” that the voice memo approach does, just through a different channel.
The point stands though: the best portfolio writing starts with capturing your real thoughts in real time, not reconstructing them after the fact through an AI prompt.
coming at this from the illustration side and I think there is an angle that has not been fully explored here. For illustrators, the portfolio copy problem is different because our work is much more visual and the text is secondary. But that actually makes it harder in some ways, because when you only have 50 words under an illustration, every word needs to carry weight.
I have found that the most authentic portfolio descriptions for illustration work are the ones that explain a creative decision rather than describe an outcome. Instead of “a whimsical children’s book illustration featuring forest animals in a watercolor style” (which any AI can generate), writing something like “the publisher wanted cute but I pushed for something a little unsettling because the story deals with a child overcoming fear, and I did not want the visuals to undercut that emotional arc” tells the reader something real about how I think.
That kind of insight does not come from any AI tool or framework. It comes from being willing to share the opinionated, sometimes messy reasoning behind your creative choices. The designers and illustrators whose portfolios I remember are the ones who let me see the thinking, not just the result.
I do agree with the general consensus here that AI tools have a place in the workflow. I use them for cleaning up grammar and structure. But the substance, the part that makes a portfolio yours, has to come from you. No shortcut around that.
Really valuable thread. The journaling approach from @MrVortex44 resonates with me. I started keeping project notes in Obsidian a few months ago and it has made portfolio updates so much faster. You just go back through your notes and the case study practically writes itself because all the details are already captured.