Karen Karapetyan · Senior Product Designer at AleNet Tecnologies
“A curious, quick learner with a solid process foundation and a real drive to grow. A pleasure to work with.”
Karen Karapetyan · Senior Product Designer at AleNet Tecnologies
Assessed
Oct 3, 2026
Verification
AI-Powered Product Design Workflow Review
Verdict: borderline
Arthur brings a good foundation in UX/UI design processes and is already using AI in his daily work, which puts him ahead of many designers at his stage.
2 demonstrated3 partial
What stood out most was how quickly he understood every point we discussed and how ready he is to grow. His next step is adding depth to how he works with AI: sharing more about the product, the business, and the users so the AI can give him truly useful results, and building simple ways to measure AI's value and risks across research, data analysis, and the wider UX process. Adding a PRD before design and building high-fidelity pages on a design system will also unlock AI with Figma through MCP for him. These are habits, not missing talent, and with his attitude and solid base, I am confident he will make these improvements quickly and see a real difference in his work.
Arthur works confidently in Figma. The next step for him is structuring his files around a design system, which will let him connect Figma to AI through MCP and work much faster.
Product Design
Demonstrated
Arthur showed a good understanding of the UX/UI stages and how they connect. Adding a PRD before design will make his strong process even stronger.
AI-Powered Design
Partial
Arthur already uses AI across his design work, which is a great starting point. His prompts were mostly short requests like "create this," and once we discussed adding product, user, and business context, he immediately saw how much better the results could be.
Design Systems
Partial
Arthur's screens are clean, but when he uses AI to generate UI, he asks it to create pages from scratch rather than building from his own components and tokens, so the results don't stay consistent. Once we discussed giving AI a design system to work from, especially through Figma MCP, he immediately saw how it would make AI output faster, consistent, and ready for development, and he was keen to apply it.
UX Strategy
Partial
When we discussed how to measure AI's impact and risks in research and data analysis, Arthur did not yet have a method in place, but he engaged thoughtfully and asked practical questions about how to start.
Evidence
What happened in the session
Arthur walked through his UX/UI process with a good understanding of each stage and how they connect. This is a strong base to build on.
He is already using AI in several stages of his work, which shows initiative and an open mindset toward new tools.
His prompts were usually short, like "create this" or "I want this." When we talked about adding context about the product, the users, and the business goals, he quickly understood how this would lead to much more useful results.
We discussed measuring AI's value and its risks in research and data analysis. He does not have a method for this yet, but he was very interested in building one.
He currently moves into screen design without a PRD and builds high-fidelity pages without a design system. He saw clearly how adding both would also open the door to using AI with Figma through MCP.
Throughout the session, Arthur was open, curious, and quick to learn. He took every piece of feedback positively and asked practical questions about how to apply it. His attitude and ability to improve are his biggest strengths, and I'm confident he will grow fast.
Action plan
What to do next
1
Give AI the full picture. Use a simple 5-part structure in every prompt: role, context (product, users, business goal), task, constraints, and output format. Save your best prompts in a personal library and improve them each week. You will see the quality of results improve quickly.
2
Start every project with a PRD. Before opening Figma, write a short PRD covering the problem, users, business goal, scope, requirements, and success metrics. AI can help you draft it. Share it with your AI agent at every stage so all outputs stay connected.
3
Measure AI's value and risks. For each stage where you use AI, set one quality check and one risk rule (for example, verify insights against real data, and never share confidential information). Track the results for 4 weeks to see where AI helps you most.
4
Build on a design system. Turn one existing project into a system-based file with components, variables, auto layout, and clear naming. From then on, create high-fidelity pages from your system.
5
Connect Figma to AI through MCP. Once your file is system-based, connect it to Claude through Figma MCP and rebuild one flow end to end. Come back for a follow-up review in 8 to 10 weeks. I look forward to seeing your progress.
