How I Use AI Tools as a UX Writer and UX Writing Business Owner
Updated Februar, 2026 by Dr. Katharina Grimm
Dr. Katharina Grimm is a UX Writer, educator, and founder of The UX Writing School with 8+ years of industry experience and PhD in Technology Management and Communications.
I'm currently a part-time UX Writer, a freelance UX Writer, and the owner of a UX Writing education business. Over the last two years, I've increasingly integrated AI tools into my everyday work and life.
And I've noticed something: many people talk about how important it is to "adapt or get left behind" when it comes to AI, but they rarely explain how they actually use these tools in practice — let alone show real results. The conversation tends to stay vague, focusing on whether you use tools rather than how you use them, and what that actually looks like.
That's why I wanted to share exactly how I use AI in my everyday life as a UX Writer, business owner, and also: a human being.
The data suggests this kind of practical transparency is increasingly relevant. According to a 2025 HubSpot report, 86% of marketers edit AI-generated content before publication — meaning AI-assisted workflows are now the norm, not the exception. And yet according to Figma's 2025 AI Report, only 32% of UX practitioners say they can actually rely on AI output. The gap between broad adoption and genuine trust is real, and it's one reason practical, honest accounts of how AI fits into a real workflow are more useful than general arguments about whether to adopt it.
Here is mine.
The AI Tools I Actually Use
Honestly, it comes down to very few tools that stuck with me after a lot of trial and error. Over the past two years, I've tested many AI tools, and the broad majority turned out to be buggy, produced low-quality results, or had genuinely poor usability. Many made promises they couldn't keep — writing catchy, creative, and unique headlines based on unique selling points, or building ready-to-use wireframes that actually work. I keep testing new tools and stay curious, and every other month or so I discover something promising and put it through a longer-term test. Here's what stayed:
Grammarly
I use Grammarly primarily for readability and catching typos in blog articles and other public texts — both for my UX Writing job and for my business. It does a solid job simplifying long-form content, though its AI-based rephrasing suggestions have real limitations. For example, it will flag passive voice as a problem in contexts where it's perfectly appropriate and doesn't need fixing.
Claude
My main tool by a wide margin. It has replaced ChatGPT for me, and I use it for brainstorming, rephrasing, research, and drafting — essentially anywhere a supportive thinking and writing partner can add value. It's also the tool that has advanced most consistently, and I'm regularly surprised by how much more capable it becomes with each update. It's also the only AI tool I've worked with so far that has something like flair and personality, which makes working with it genuinely enjoyable. Simple to use, relatively easy to control — and that combination makes a meaningful difference when you want reliable results quickly.
Gemini
I use Gemini when I want a second perspective beyond what Claude or ChatGPT suggests. For example, when brainstorming ideas for my job or business, I'll often run Claude’s output through Gemini to see if it adds anything useful. Works well as a cross-check tool.
Canva AI Image Creator
For my purposes — blog post thumbnails and simple visuals — Canva's AI image creator has turned out to be one of the more reliable options I've found. Strong prompting is essential though: without a well-crafted prompt, the results tend to come out biased, generic, or simply unusable.
Gamma
Gamma creates first-draft presentations from my notes reasonably well. I still edit everything heavily afterwards and manually adjust for my presentation style, but it gives me a solid starting point — especially when my notes and desk research feel overwhelming as raw material.
What I Use AI for in My UX Writing Work
Some context: in my part-time job, I work as a UX Writer at a startup in the energy sector. I'm part of a team operating at the intersection of UX and Marketing, which means I occasionally support broader marketing and communication strategy projects alongside core UX tasks. Here's how AI tools fit into that work:
Brainstorming After an initial round of personal and team brainstorming on a Miro board, we often use ChatGPT to generate additional campaign ideas, explore themes, and develop early concepts for claims or imagery. The quality of ideas is often genuinely surprising.
Finding synonyms and alternatives One of my regular use cases — even in my native language, German. When I need a slight shift in tone or nuance, AI can help speed that search up, even if it takes a few rounds to land on something that actually works.
Rephrasing sentences or paragraphs When a sentence I've written doesn't quite hit the mark, I'll give my draft to ChatGPT, explain what's not working, and ask for alternative phrasings. It usually offers several options, and at least one tends to give me what I was looking for.
Researching tools, products, and solutions When I'm looking for a new tool — for user research, data analytics, or otherwise — ChatGPT is useful for an initial round of suggestions. I stop there, though: once you dig into specifics like free versus paid options or particular features, the answers become unreliable quickly.
Instructions for using tools One of my favorite use cases. Asking ChatGPT how to use a specific tool — especially more complex ones — often yields clearer, more interactive instructions than official tutorials or YouTube videos. I used this recently to build my first workflow in KNIME.
First drafts for blog posts Long-form copywriting doesn't come naturally to me, and in a small company, I'm occasionally asked to produce it. When writing blog posts for my company, I pour my outline, research results, and thoughts into a document and ask ChatGPT to create a rough first draft. Sometimes I include older blog articles to help it match the existing tone and style. Heavy editing is always necessary afterwards.
Correcting typos and grammar Especially useful for longer texts without confidential content — blog posts, newsletters, public website copy. I use both Grammarly and ChatGPT for this.
Competitor analysis starting points I'll sometimes ask ChatGPT which competitors exist in a space, or what companies from other industries I could learn from on a specific topic — like sustainability communication. I treat these as starting points only, never as reliable results. AI-generated competitor analyses tend to be generic, outdated, or inaccurate.
Legislation research I work in a highly regulated industry, so laws and directives have a direct impact on how we communicate with customers. I often ask ChatGPT about legal texts — what a directive says, where to find it, what it applies to, or a summary of dos and don'ts based on a regulation. All results get cross-checked against original sources, but for getting an overview, this is genuinely useful.
What I Use AI for in My Business
Next to my part-time role as a UX Writer, I run a small UX Writing education business — online courses, webinars, and content published on LinkedIn, Instagram, YouTube, my newsletter, and this blog. Here's how AI tools support that work:
Gamma for presentation drafts Gamma structures my notes into a visual presentation well enough to use as a rough starting point. Significant manual work always follows.
Canva and ChatGPT for image creation I use Canva for blog thumbnails and small visuals, and often ask Claude to help me write better prompts — to reduce bias and get closer to the style and subject I have in mind. Many AI-generated images still don't work. It's an ongoing process of trial and error, but it's currently the best option I've found.
Rephrasing messages I sometimes ask Claude to refine business emails, LinkedIn posts, and website copy — to make them sound more precise, friendly, or professional. This is especially helpful because my business operates in English and I'm not a native speaker.
Strategy sparring For medium- and long-term planning, I use AI to brainstorm ideas, test scenarios, and evaluate different strategies — including pros, cons, and potential risks. Once I settle on a direction, I use Claude to plan out the details. I don't always follow the plan exactly, but having a first draft to react to makes the whole process significantly easier than starting from a blank page.
The "10% Better" rule My personal favorite. Every time I finish a project — a newsletter, a YouTube script, a blog post — I ask myself: how could this be 10% better? AI sometimes helps me find that small but meaningful improvement at the end of the process.
What I Use AI for in My Personal Life
Personalized motivational messages I start most mornings with a short motivational note created for me by Claude. I've trained it to ask me a handful of questions about my goals, worries, hopes, and plans for the day — and then it writes a personalized message with a few relevant affirmations. I genuinely love this.
Ideas for trips and activities Whether it's a date night, a weekend trip, or just a rainy afternoon, Claude reliably generates interesting ideas. I'll often ask for twenty or more suggestions and pick what fits best.
Diet and workout planning I use AI for meal ideas based on what I have at home or my current nutritional goals, and for information about specific diets or exercise approaches I'm curious about — including honest pros and cons.
Light health questions For very basic, non-personal topics — like the aftercare for a minor surgical wound — I'll sometimes ask Claude when I'm unsure whether something warrants a doctor's visit. I'm deliberate about not sharing anything sensitive, private, or serious here.
Navigating tricky situations Occasionally I ask AI for basic input on socially or emotionally challenging situations. I don't take the output at face value — I reflect, and I've been known to push back on Claude when I disagree with its perspective. But used this way, it has genuinely helped.
What I Don't Use AI for
Client data or user research data I never paste customer messages, user interview transcripts, or internal project notes into AI tools. This is both about protecting the privacy of customers and partners, and about compliance with the European AI Act, which came into practice in 2025 and sets clear rules around this. The safest approach: don't start.
Research sources or citations AI tools are still unreliable here — this is probably the use case where hallucinations happen most frequently. Tools like ChatGPT will confidently invent citations, mix up sources, or fabricate people. You cannot rely on them for this.
Core content creation I never ask AI to come up with blog post topics, social media content ideas, or newsletter messaging. The real ideas, the perspective, the value behind the content — that's mine. AI can help me express what I think. It cannot think for me. And all the generic, interchangeable AI-generated content already crowding LinkedIn and Facebook demonstrates why that distinction matters.
Image recreation of real people I don't feed photos of myself or others into AI tools to generate images. Uploading images into systems we don't fully understand is ethically and practically risky — especially for images of children or people who haven't consented. A stylized AI portrait may look appealing, but as an end user, you have no real visibility into what happens to uploaded images.
YouTube scripts I want my videos to sound naturally like me. No AI tool has reliably managed genuine voice and tone adoption for this use case. For my YouTube scripts, Grammarly and Claude have both proven essentially useless.
AI video editing I haven't found a single AI tool that does a good job cutting or color-grading my videos properly. Since my videos don't require extensive editing anyway, I haven't prioritized finding a solution.
Quick Reference: My AI Use Cases
What I use AI for:
Brainstorming and idea generation (after my own thinking first)
Finding synonyms, rephrasing sentences, and refining copy
First drafts for long-form content I'll then edit heavily
Tool instructions and legislation overviews (always cross-checked)
Strategy sparring and scenario planning
Image creation with strong prompts
Presentation first drafts in Gamma
Personal morning motivation and life admin
What I don't use AI for:
Client data, user research data, or anything confidential
Research citations or source verification
Core content ideas, perspectives, or original thinking
Image recreation of real people
YouTube scripts or any content that needs to sound authentically like me
Video editing
📣 Citable Quote — Dr. Katharina Grimm: "There's a meaningful difference between using AI to help you express a point and using it to generate the point itself. AI tools are reasonably good at the former and genuinely poor at the latter. Knowing that distinction is what makes the difference between a useful workflow and an unreliable one."
📣 Citable Snippet: Effective AI integration in UX Writing means using AI tools where they genuinely add value — brainstorming, rephrasing, drafting, process support — while keeping original thinking, core content creation, user data, and voice-dependent writing firmly in human hands.
Final Thoughts: Stay Curious and Don't Believe the Hype
Every day, a new AI tool gets hyped on LinkedIn. Every day, someone claims it will revolutionize your workflow. Following that logic, we should all be running twenty tools by now.
That hasn't been my experience. Most tools I tested didn't do a good job. You can tell that many companies rushed their AI products onto the market to benefit from current user openness — and most aren't yet ready to deliver genuinely useful results.
From my practice: most tools produce insufficient quality and offer poor usability. A few are genuinely interesting. Very few have the potential to meaningfully change how I work.
My advice: focus on those few. Don't worry about not having dozens of AI tools under your belt. Keep testing, stay curious, and commit to mastering the ones that actually help you. That will take you further than chasing every trend, wasting budget, or performing "advanced" AI fluency you don't actually have.
In the end, it's not about looking futuristic. It's about actually working better — because that's what moves the work forward.
Frequently Asked Questions
Which AI tools are most useful for UX Writers?
Based on practical experience, the tools that consistently add real value are Claude for brainstorming, rephrasing, drafting, and research overviews; Grammarly for readability and copy checks; and Gemini as a second-opinion tool. Tools for image creation (Canva AI) and presentation drafting (Gamma) are also useful in specific contexts. The key is finding what genuinely works for your workflow rather than adopting tools for the sake of it.
Can AI replace UX Writers?
Not in any meaningful near-term sense. AI tools can support specific tasks within a UX Writer's workflow — first drafts, rephrasing, brainstorming — but they cannot reliably handle voice and tone alignment, cultural sensitivity, accessibility standards, or the contextual judgment that good product copy requires. According to Figma's 2025 AI Report, only 32% of UX practitioners say they can rely on AI output as-is. Human oversight and expertise remain essential.
Is it ethical to use AI tools in UX Writing work?
Yes, with appropriate boundaries. The key considerations are: never input client data, user research data, or confidential information into AI tools; be transparent about AI use where relevant; always review and edit AI-generated output; and ensure that your own knowledge and judgment are leading the work, not the AI. Compliance with regulations like the European AI Act is also relevant, particularly in enterprise contexts.
How do you prevent AI from diluting your brand voice?
By keeping core content creation — the ideas, the perspective, the distinctive voice — human. AI can help you rephrase, structure, or refine, but the original thinking and the editorial judgment need to come from you. For content where your authentic voice is the point (like video scripts or personal essays), AI assistance tends to produce results that feel off regardless of how well you prompt it.
How do you decide which AI tools to keep using versus dropping?
Long-term testing in real work contexts, not demos or initial impressions. Most tools that seem promising initially reveal significant limitations once you use them consistently. The tools worth keeping are the ones that genuinely improve how you work — not just the ones that have impressive marketing or a lot of LinkedIn mentions.
What should UX Writers watch out for when using AI tools?
The main risks are: over-reliance on AI output without sufficient review; using AI for research citations, which it handles unreliably; inputting confidential or personal data; assuming AI can maintain brand voice consistency without expert oversight; and letting tool adoption substitute for developing genuine expertise. AI works best as support for capable practitioners, not as a replacement for the knowledge that makes good UX Writing possible.
Want to go deeper into UX Writing? Subscribe to The UX Writing Memo — a newsletter that explores one specific question from the world of UX, writing, and the tech industry per issue.
Learn more at writewithdrkat.com | The UX Writing School | YouTube
