<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Orezi Mena on Medium]]></title>
        <description><![CDATA[Stories by Orezi Mena on Medium]]></description>
        <link>https://medium.com/@thecreativemena?source=rss-d9fa9cbc8869------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*iWUAjJb8i2R1ZQ4QK08TFg.jpeg</url>
            <title>Stories by Orezi Mena on Medium</title>
            <link>https://medium.com/@thecreativemena?source=rss-d9fa9cbc8869------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 13:16:05 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@thecreativemena/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[How to get a downloadable link for your CV using Github Releases]]></title>
            <link>https://thecreativemena.medium.com/hosting-your-cv-on-github-using-releases-01402612a5be?source=rss-d9fa9cbc8869------2</link>
            <guid isPermaLink="false">https://medium.com/p/01402612a5be</guid>
            <category><![CDATA[resume]]></category>
            <category><![CDATA[job-hunting]]></category>
            <category><![CDATA[cv]]></category>
            <dc:creator><![CDATA[Orezi Mena]]></dc:creator>
            <pubDate>Fri, 02 Jan 2026 10:44:22 GMT</pubDate>
            <atom:updated>2026-01-02T10:45:39.960Z</atom:updated>
            <content:encoded><![CDATA[<p>I have tried so many methods and honestly i got frustrated, especially because it takes multiple clicks, and even the free ads method which is google drive would require the user to make 2–3 clicks. That frustration pushed me to GitHub. Not for code — for my CV.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aUVYJ6DoIT5Fz5DZ7i6HKg.png" /></figure><p><strong>Why GitHub Releases Actually Works</strong></p><p>GitHub has this feature called Releases. It’s designed for distributing files, finished assets that need to live somewhere permanent and accessible.</p><p>The link doesn’t expire. The file downloads cleanly. You can version updates over time without losing previous versions. And honestly? It just looks intentional.</p><p>Instead of your CV floating in some random folder, it becomes a published artifact.</p><p><strong>What You Need</strong></p><p>A GitHub account and your CV saved as a PDF. That’s it. No coding knowledge required.</p><p><strong>Setting Up the Repository</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rzfjj46k_27BQ1d22-bYnQ.png" /></figure><p>Log into GitHub and create a new repository. Name it something direct like cv or resume works perfectly. Make it public. You don’t need to add any files yet. Think of this repository as a container whose only job is to host your CV.</p><p><strong>Getting to Releases (Important)</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ruGskkuRThq4cQAYxkHKfA.png" /></figure><p>Here’s where things get slightly quirky: sometimes the Releases section doesn’t appear on a brand new, empty repository.</p><p>If you don’t see it in the sidebar, just upload any file first ,even a simple README. Once there’s at least one file in the repository, the Releases option will appear.</p><p><strong>Publishing Through Releases</strong></p><p>Navigate to the Releases section inside your repository. If there aren’t any releases yet, GitHub will prompt you to create one. Click to create a new release. Give it a version number, something simple like v1.0. The title can just be “CV” or “Resume.” The description is optional.</p><p>When you reach the section for attaching binaries, upload your CV PDF. Once it finishes uploading, publish the release.</p><p>Your CV is now officially live.</p><p><strong>Getting the Shareable Link</strong></p><p>After publishing, GitHub displays an Assets section. Your CV appears there as a downloadable file.</p><p>Copy that link. This is what you share.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_WN0QiLD-1d1rF-2FqsUVw.png" /></figure><p>When someone opens it, the CV either downloads immediately or opens directly in their browser, depending on their settings. No friction. No access requests. No broken links.</p><p><strong>Updating Over Time</strong></p><p>This is where Releases really shines, you can have a version history for your CV.</p><p>When you update your CV, create a new release with a new version number and upload the updated PDF. Each version remains accessible, which can be useful if you ever need to reference an older version.</p><p>If you prefer consistency, keep the file name the same across releases. The version number handles the tracking.</p><p><strong>Where This Link Works</strong></p><p>LinkedIn. Portfolio websites. Email signatures. Job applications. Anywhere you need a reliable CV link.</p><p>It removes friction and communicates something subtle but important: you care about clarity, reliability, and presentation.</p><p><strong>A Small Shift That Matters</strong></p><p>This is a minor change in workflow. But it treats your CV like what it is a published asset, not just a document floating in the cloud.</p><p>GitHub Releases makes that shift real.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=01402612a5be" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why I Built a Sub Agent and Documentation Workflow to Stop AI Hallucinations]]></title>
            <link>https://medium.com/design-bootcamp/why-i-built-a-sub-agent-and-documentation-workflow-to-stop-ai-hallucinations-28657eb17783?source=rss-d9fa9cbc8869------2</link>
            <guid isPermaLink="false">https://medium.com/p/28657eb17783</guid>
            <category><![CDATA[sub-agents]]></category>
            <category><![CDATA[workflow]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[claude-code]]></category>
            <dc:creator><![CDATA[Orezi Mena]]></dc:creator>
            <pubDate>Wed, 03 Dec 2025 09:56:03 GMT</pubDate>
            <atom:updated>2025-12-04T08:30:45.382Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fyc6pXLUQv2cdWNQvSlZpA.jpeg" /></figure><p>You know that moment when your AI assistant confidently suggests a function name or an import or a module, and you paste it in only to see red error messages and broken builds. That is what I call the hallucination cliff.</p><p>When you build serious systems, whether it is financial platforms, creative tooling, or multi-repo design and engineering projects, that cliff is not just annoying. It becomes a reliability risk. After falling off it enough times, I built a safeguard. A simple, human-controlled sub agent process that sits on top of my coding flow. Every time the AI touches code, it must reference canonical documentation. Every update gets logged. The result is far fewer hallucinations and far more confidence in what ships.</p><p>In this piece, I share why hallucinations happen and exactly how I engineer around them. If you use AI tools for real work and not small toy projects, this can save you hours of debugging and a lot of frustration.</p><h3>Why AI Assistants Hallucinate in the First Place</h3><p>Large language models do not understand your codebase. They predict the next token based on patterns rather than fully grasping your project’s actual state. That works for small snippets and simple demos. Although once you enter real-world complexity with custom logic, evolving APIs, multiple files, and context that changes every week, they lose track.</p><p>When context becomes unclear, the model fills the gaps with guesses. Those guesses look correct but are not. That is hallucination.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0YYw1Ms3E8owKTuqyYL7QA.jpeg" /></figure><p>It shows up as nonexistent functions, wrong modules, invented class methods, fake imports, or entire flows that do not exist anywhere in your repository. For developers building production-grade systems, hallucinations are not a funny quirk; they are a liability.</p><p>There is also a dangerous version called package hallucination, where the AI suggests dependencies that look legitimate but do not exist. If you install them blindly, you expose yourself to bugs and potential security risks.</p><p>So hallucination is not an LLM malfunction. It is a predictable side effect of using probabilistic models on live, evolving, private codebases without grounding or guardrails.</p><h3>What Most People Try and Why It Often Fails</h3><p>Developers try different mitigation techniques, although many stop working once the project grows.</p><p><strong>Prompt engineering</strong> helps a little. You can say things like “use only existing APIs” or “avoid inventing functions,” although this is not enough once your project becomes complex.</p><p><strong>Public documentation</strong> also helps slightly. But public docs do not represent your internal logic, recent refactors, or project-specific architectural decisions.</p><p><strong>Retrieval augmented generation</strong> is one of the best known approaches. You fetch external text and feed it to the model so that the output becomes grounded in some context. Although if the retrieved information is outdated or incomplete, the AI still guesses.</p><p>In short, without precise and up-to-date grounding, hallucination returns as soon as the project evolves.</p><h3>My Workflow</h3><h4>Sub Agent with Docs and Git Logs That Keep the AI Grounded</h4><p>I work on projects that change constantly. I needed something stable. Here is the exact formula that works for me.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*iAGgAV-mSJg5ihXl6UOT5Q.jpeg" /></figure><p><strong>1. Canonical documentation in markdown</strong></p><p>Every project has a set of markdown files that act as the official truth. Architecture. API layout. Module breakdown. Design patterns. Nothing lives only in my head.</p><p><strong>2. A sub agent that reads the docs before generating code</strong></p><p>Before I let the AI write or modify anything, I run a sub agent that loads the relevant markdown files. This forces the AI to operate with real context instead of its loose memory.</p><p><strong>3. An update log that tracks every change</strong></p><p>Whenever I push new code or edit documentation, I record high-level changes in an update log file. It becomes a journal of everything evolving in the repo. The AI reads this too, so it never works with stale context.</p><p><strong>4. Version control for both code and documentation</strong></p><p>Because everything is tracked in Git, I always have full traceability. I can revert, compare diffs, and audit decisions whenever I need.</p><p><strong>5. Guardrails in prompts</strong></p><p>Every time I call the AI, I include strict instructions. Use only what is documented. Avoid creating new dependencies without confirmation. Ask if uncertain.</p><p><strong>6. Human review as the final step</strong></p><p>The AI is a junior engineer. I review every change. Nothing gets merged blindly.</p><p>This combination of documentation, update history, guardrails, and human review keeps hallucination extremely low, and it scales well for complex projects.</p><h3>How This Aligns With What Experts Recommend</h3><p>My workflow mirrors what many researchers and practitioners highlight as best practice.</p><p>Studies on RAG show that retrieval helps only when the retrieved context is well structured and accurate. Poorly managed retrieval still produces hallucinations.</p><p>Other research shows that grounding generated code in verified API references improves correctness by a significant margin.</p><p>Industry guides also push for a mix of retrieval, explicit constraints, schema checks, and human verification when deploying AI assisted code.</p><p>In other words, the world is learning the same lesson. Treat AI like a junior developer who needs specs, history, and supervision. Not like a magic autocomplete.</p><h3>Why This Matters for Anyone Doing Real Work</h3><p>If you are building anything beyond tiny demo projects, you need reliability. This workflow gives you that.</p><p><strong>You get stability</strong>, because the AI cannot drift far from the truth.</p><p><strong>You get maintainability</strong>, because clean docs and a live update log make onboarding easy for collaborators and for your future self.</p><p><strong>You get security</strong>, because the AI cannot randomly invent dependencies that might introduce vulnerabilities.</p><p><strong>You get speed</strong>, because the AI becomes fast and predictable instead of wild and unreliable.</p><p>For creatives, founders, designers, and developers who juggle multiple projects, this workflow provides a balanced mix of speed and structure.</p><h3>Quick Checklist for Your Own Workflow</h3><ul><li>Keep documentation in markdown inside your repo.</li><li>Maintain a live update log that captures recent changes.</li><li>Always load documentation before generating or modifying code.</li><li>Give the AI clear constraints and rules.</li><li>Review everything before merging.</li><li>Use retrieval or external memory when your project becomes large.</li><li>Keep human judgment at the center of the system.</li></ul><p>AI coding assistants are powerful tools, although they lose accuracy when context becomes fuzzy. When you combine documentation, version control, retrieval, and human supervision, you eliminate most of the hallucination risk and turn the AI into a dependable collaborator.</p><p>If you adopt even a simplified version of this workflow in your next project, you will ship faster and with more confidence.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=28657eb17783" width="1" height="1" alt=""><hr><p><a href="https://medium.com/design-bootcamp/why-i-built-a-sub-agent-and-documentation-workflow-to-stop-ai-hallucinations-28657eb17783">Why I Built a Sub Agent and Documentation Workflow to Stop AI Hallucinations</a> was originally published in <a href="https://medium.com/design-bootcamp">Bootcamp</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why You Should Stop Using One AI Agent for Everything (And What to Do Instead)]]></title>
            <link>https://medium.com/design-bootcamp/why-you-should-stop-using-one-ai-agent-for-everything-and-what-to-do-instead-58b7eeb3b21d?source=rss-d9fa9cbc8869------2</link>
            <guid isPermaLink="false">https://medium.com/p/58b7eeb3b21d</guid>
            <category><![CDATA[vibe-coding]]></category>
            <category><![CDATA[claude-ai]]></category>
            <category><![CDATA[design-engineer]]></category>
            <dc:creator><![CDATA[Orezi Mena]]></dc:creator>
            <pubDate>Tue, 02 Dec 2025 10:18:15 GMT</pubDate>
            <atom:updated>2025-12-02T12:20:50.055Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lBQZjjhNQn_yb-cOHerOVw.png" /></figure><p>There’s a certain excitement that comes with starting a new project with AI. You open your editor, fire up Claude or Cursor, and before you even think twice, you’re already prompting:</p><p>“Generate my backend.”<br> “Build the UI components.”<br> “Set up the architecture.”<br> “Write tests.”</p><p>It feels productive in the moment, like you’re finally moving at the speed of thought.</p><p>Until everything falls apart. A week later, your codebase is tangled. Your logic contradicts itself. And the only realistic path forward i starting all over, Control + A, then delete… phew.</p><p>We don’t talk about this enough. But the issue isn’t the AI. It’s that we’re asking one agent to play too many roles. And no great product, human or machine, thrives under that kind of pressure.</p><h3>Designers Already Know This: Systems Beat Chaos</h3><p>If you’ve ever built a design system, you know this truth intimately. Creating color tokens, spacing scales, typography rules, component variants… It’s tedious at first. It feels like sweeping the studio before painting the masterpiece.</p><p>But you do it because you know the downstream effect: faster execution, cleaner decisions, better consistency, and a product that actually scales.</p><p>Working with AI is no different.</p><p>Most people prompt AI like it knows everything. But a good designer knows better: <strong>structure always beats magic. </strong>And structure, in the world of AI engineering, looks like sub-agents.</p><h3>What Exactly Are Sub-Agents?</h3><p>Think of sub-agents as digital teammates, each one designed for a specific responsibility.</p><p>Instead of one model handling everything from backend logic to UI pixel-nudging, you break the work into specialist roles:</p><p><strong>Backend Agent:</strong> Creates API schemas, models, routes, integrations.<br> <strong>Frontend Agent:</strong> Builds clean components, maintains consistency with your design tokens.<br> <strong>QA Agent:</strong> Writes tests, spots edge-case failures before you do.<br> <strong>Design Agent:</strong> Guards your design language across the codebase.<br> <strong>Architecture Agent:</strong> Ensures your long-term structure won’t betray you later.</p><p>This isn’t about having “more AI.” It’s about working the way good engineering teams already work.</p><p>You wouldn’t ask your product designer to configure your servers. You wouldn’t ask your QA engineer to choose your typography scale. You wouldn’t ask your backend dev to craft your iconography set.</p><p>So why do we ask a single AI to do it all?</p><h3>The Childlike Wisdom Hidden Here</h3><p>This idea didn’t come to me from reading research papers. It came from remembering how children ask for help.</p><p>In childhood, every adult has a job:</p><p>Your mum helps with your homework. Your uncle opens the stubborn jar.<br> Your older cousin fixes the game console. A neighbor shows you how to tie your shoelace.</p><p>As kids, we intuitively knew what adults forget: <strong>you ask the right person for the right help. </strong>Growing up complicates that. We start believing that “independence” means “do it all yourself,” even if you break in the process.</p><p>But like the stories from <em>Don’t Grow Up (My Book)</em>, there’s a lesson we often forget: a village gets you farther than a lone genius.</p><h4><strong>Sub-agents bring back that village.</strong></h4><h3>Why Sub-Agents Make Your Work Better</h3><p>Here’s what happens when one AI tries to do everything:</p><p>Code becomes inconsistent. Logic contradicts itself. Architecture drifts. You spend more time debugging than building. You reset the project more often than you move it forward.</p><p>But with sub-agents? Responsibilities are clear. Outputs are clean. Decisions are traceable. You can iterate without rewriting the world. Your project grows like a tree, not like a wildfire.</p><p>Multiple agents don’t confuse the work. They clarify it.</p><h3>This Isn’t Over-Engineering, It’s Good Leadership</h3><p>When you use sub-agents, you’re no longer “prompting ChatGPT.” You’re managing a team. You create briefs. You assign roles. You give each agent the boundaries it needs to excel. You orchestrate collaboration.</p><p>Suddenly the project becomes less chaotic, more intentional.</p><p>And you, the design engineer, become less of a “prompt typist” and more of a creative director of intelligent systems.</p><p><strong>That is the future of engineering.</strong></p><h3>The Practical Bit: How This Actually Works</h3><p>The setup is simpler than you think. You define your sub-agents upfront, then summon them when needed.</p><p><strong>Start by creating your team.</strong> Write out clear role definitions for each agent you’ll need:</p><pre><strong>Backend Agent: </strong>You specialize in building clean, scalable APIs. <br>You handle database schemas, route logic, and integrations.</pre><pre><strong>Frontend Agent: </strong>You build UI components that strictly follow <br>our design system. You prioritize accessibility and performance.</pre><pre><strong>QA Agent:</strong> You write comprehensive tests and catch edge cases. <br>You think about what could break before it does.</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*KfhwRoc0hKdcqCer.png" /></figure><p><strong>Then add them to your project context.</strong> In tools like Cursor or Claude Projects, you can store these definitions as custom instructions or project knowledge. The system learns who does what.</p><p><strong>When you need something built, just call on the right specialist.</strong> Say “Backend Agent: create the user authentication flow” or “Frontend Agent: build the dashboard cards.” The system already knows the role and responsibilities.</p><p>The beauty? You’re not juggling multiple chat windows. You’re orchestrating specialists within one workspace, and the AI knows which hat to wear based on who you’re calling.</p><p>It’s like having a team on standby, ready when you need them, quiet when you don’t.</p><h3>I will leave you with this…</h3><p>If you’re building anything meaningful with AI, stop relying on a single agent to do everything.</p><p>Build your team.</p><p>Break the work down. Assign roles. Let your sub-agents collaborate the way real engineering teams do.</p><p>Because the moment you stop expecting one AI to be the genius but rather become a better team player, you start building products that actually scale.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=58b7eeb3b21d" width="1" height="1" alt=""><hr><p><a href="https://medium.com/design-bootcamp/why-you-should-stop-using-one-ai-agent-for-everything-and-what-to-do-instead-58b7eeb3b21d">Why You Should Stop Using One AI Agent for Everything (And What to Do Instead)</a> was originally published in <a href="https://medium.com/design-bootcamp">Bootcamp</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Design Is the Elden Ring of Tech’s Next Cycle]]></title>
            <link>https://medium.com/design-bootcamp/design-is-the-elden-ring-of-techs-next-cycle-87203f793691?source=rss-d9fa9cbc8869------2</link>
            <guid isPermaLink="false">https://medium.com/p/87203f793691</guid>
            <category><![CDATA[future-of-work]]></category>
            <category><![CDATA[ai-design]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[innovation]]></category>
            <category><![CDATA[ux-design]]></category>
            <dc:creator><![CDATA[Orezi Mena]]></dc:creator>
            <pubDate>Mon, 17 Nov 2025 11:24:20 GMT</pubDate>
            <atom:updated>2025-12-02T12:20:52.454Z</atom:updated>
            <content:encoded><![CDATA[<p>AI can now generate full features, refactor entire codebases, and translate plain language into working prototypes. Tools like Cursor, Claude Code, and GitHub Copilot have collapsed the gap between idea and implementation. What once required a team of engineers can now be produced with a few prompts.</p><p>But even as AI accelerates the “building” side of technology, another truth has become undeniable. AI can generate function, but it cannot create intention. It cannot understand human nuance, emotional clarity, or the delicate balance between usefulness and delight.</p><p>This is why design is becoming the Elden Ring of tech’s next cycle. Not an accessory. Not an afterthought. The rare, powerful discipline that will separate the companies that move forward from the ones that stay stuck.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uSurrTRJd6uFlDhGX1eyKA.png" /></figure><h3>AI Can Code, But AI Can’t Design With Intuition</h3><p>Engineering has been transformed by AI. Boilerplate disappears. Debugging becomes collaborative. Frontend components are generated from Figma in seconds. Yet AI still struggles with one thing: understanding people.</p><p>AI can assemble a UI. It cannot sense friction. AI can recreate a pattern. It cannot imagine a feeling. AI can follow an instruction. It cannot decide what should be built.</p><p>Human designers remain the interpreters of emotion, context, frustration, and aspiration. They translate the unsaid into products that feel intuitive. They define meaning. They decide what deserves to exist.</p><p>As execution becomes cheaper, discernment becomes priceless.</p><h3>The Shift: Designers Are Not Just Designing, They Are Building</h3><p>A quiet revolution is happening inside product teams. Designers are no longer limited to mockups and handoffs. They are building the tools themselves. AI has unlocked a new level of autonomy.</p><p>This is not hypothetical. It’s happening inside real teams today.</p><h4>Atlassian</h4><p>Design director Joel Unger has shared publicly how designers use Cursor to prototype full interactions in actual code. Instead of static Figma flows, designers hand over interactive prototypes that engineers can run, inspect, and refine.</p><h4>Dive Club</h4><p>In their 2025 essay What’s Next for AI Design, Dive Club notes that tools like Cursor have made it possible for designers and researchers to build working internal apps without engineering support. They highlight dashboards, automation tools, and AI workflows built entirely by non-engineers.</p><h4>Whitespectre</h4><p>In a published case study, a team built a functional internal prototype in about four hours using Cursor and ChatGPT. Before AI, the same task required weeks of engineering time. Designers and PMs iterated live, validated assumptions, and only pulled in engineering when ready for scale.</p><p>The pattern is clear. Designers who understand workflows deeply can now build the tools that improve them.</p><h3>Why This Matters: Internal Tools Are the Hidden Growth Engine</h3><p>Internal tools rarely get attention, but they shape the speed and quality of every external product a company ships. When these tools suffer, the whole company slows down.</p><p>AI enables designers to transform internal systems in ways engineering orgs struggled to prioritize.</p><h4>Faster validation</h4><p>A designer can build the first working version of a tool in hours. The team tests something real, not theoretical.</p><h4>Better usability</h4><p>Internal tools now get the same care and consideration as user-facing products.</p><p>Spotify’s design team shared how one internal system collapsed a multi-week workflow spanning seven tools into a single interface. The result was better execution, fewer errors, and smoother launches for global campaigns like “Wrapped.”</p><h4>Reduced engineering bottlenecks</h4><p>Designers handle the early version. Engineering focuses on performance, reliability, and scale. Everyone moves faster.</p><h4>Higher adoption</h4><p>People actually use tools that feel intuitive.</p><p>Designers make that happen.</p><p>This shift is not about replacing engineers. It is about giving designers enough power to remove friction, create leverage, and accelerate everything around them.</p><h3>The Rise of the Hybrid Designer</h3><p>The most valuable emerging role in tech is the hybrid designer.</p><p>Part designer. Part systems thinker. Part workflow architect. Part builder.</p><p>This person understands:</p><ul><li>how teams work</li><li>where friction really lives</li><li>how to turn research into systems</li><li>how to prototype with AI</li><li>how to design with empathy and precision</li></ul><p>AI handles the complexity.</p><p>The designer defines the experience.</p><p>These hybrid designers are becoming the glue inside modern product teams. They create clarity between engineering, PM, operations, and leadership. They prototype fast, test fast, and reveal what truly matters before engineering commits.</p><p>They are not replacing developers.</p><p>They are multiplying them.</p><h3>Design as Strategy in the AI Era</h3><p>AI reduces the cost of building.</p><p>Design increases the value of what is built.</p><p>In this new landscape, the question is no longer “Can we build this?” AI says yes to almost anything. The real question becomes “Does this matter?” and “Will it feel right to the person using it?”</p><p>Design answers those questions.</p><p>Hybrid designers execute on them.</p><p>The companies that elevate design now — and empower designers to build tools with AI — will ship faster, adapt faster, and innovate faster than their competitors.</p><h3>Design Is the Elden Ring of This Tech Cycle</h3><p>Elden Ring rewards adaptability, mastery, and the ability to read the terrain better than your opponent. That is exactly what modern design represents.</p><p>AI provides the strength.</p><p>Design provides the judgment.</p><p>AI generates the lines.</p><p>Design composes the story.</p><p>The designer who can think, interpret, and build with AI becomes the most powerful player in the room. The multiplier. The clarity-maker. The one who dissolves friction and unlocks velocity.</p><p>This is why design has become the Elden Ring of the next tech cycle.</p><p>And why the teams that understand this will lead the future of product</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=87203f793691" width="1" height="1" alt=""><hr><p><a href="https://medium.com/design-bootcamp/design-is-the-elden-ring-of-techs-next-cycle-87203f793691">Design Is the Elden Ring of Tech’s Next Cycle</a> was originally published in <a href="https://medium.com/design-bootcamp">Bootcamp</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Catching Fraud, Keeping Secrets]]></title>
            <link>https://medium.com/design-bootcamp/catching-fraud-keeping-secrets-ffa1f7f3c2e6?source=rss-d9fa9cbc8869------2</link>
            <guid isPermaLink="false">https://medium.com/p/ffa1f7f3c2e6</guid>
            <dc:creator><![CDATA[Orezi Mena]]></dc:creator>
            <pubDate>Fri, 30 May 2025 19:16:47 GMT</pubDate>
            <atom:updated>2025-06-01T15:53:18.891Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>A Plain-Language Walk-through of the Mobile Flow I Designed</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gRWrqArzuDHZYUIr7pULcw.png" /></figure><p>Banks and regulators wrestle with a classic tug-of-war:</p><ul><li><strong>Banks</strong> need to hide the tricks their fraud-catching AIs use, or criminals will learn to dodge them.</li><li><strong>Regulators</strong> need proof the bank is playing fair and following the rules.</li><li><strong>Customers</strong> just want their money safe and their personal data private.</li></ul><h3>The new trick: Zero-Knowledge (ZK) proofs</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*gaSx0NuXA530pbSP.png" /></figure><p>Think of a ZK proof like a <em>sealed envelope</em> that says, “Trust me, the answer inside is correct,” and then signs that statement in permanent ink. Everyone can check the signature, but nobody can open the envelope.</p><h3>Five Quick Screens, Zero Stress</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*w3VFHMkmPVeOcWyEwJVpRQ.png" /></figure><ol><li><strong>Enter or Scan: </strong>Type the transaction ID <em>or</em> scan a QR code.</li></ol><p>Why it matters: Two ways is good for desk workers and field teams.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3suhS792GScWKFvrcnbuIg.png" /></figure><p><strong>2. We Check the Math: </strong>A spinner and a short status line (“Validating proof”).</p><p>Why it matters: Lets you know something is happening, no tech jargon needed.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gc6khQDlK9jg2EO7P4pEOg.png" /></figure><p><strong>3. We Ask the Blockchain: </strong>Card shows “Ethereum Mainnet,” block number, and a progress bar.</p><p>Why it matters: Proves the check is recorded on a public ledger no one can edit.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*07qRFX2IAEo1LMcg3EgllQ.png" /></figure><p><strong>4. Result &amp; Details</strong>Big “Verification Successful” badge, amount, risk level, and short list of <em>why</em> it’s risky (five plain-English rules).</p><p>Why it matters: Auditors see everything they need. Crooks still can’t see how the rules work.</p><p><strong>5. Export: </strong>One-tap PDF for email, print, or archive.</p><p>Why it matters: Compliance teams love tidy paperwork.</p><h3>Why this design feels easy</h3><ul><li><strong>Four small steps:</strong>users always know where they are.</li><li><strong>Cards with friendly icons: </strong>each stage looks different, so progress feels real.</li><li><strong>Plain labels: “</strong>High Risk” instead of “Velocity Rule Trigger”</li><li><strong>Color for meaning, not decoration: </strong>blue for “in progress,” green for “safe,” red for “needs attention.”</li><li><strong>One export button: </strong>no hunting through menus.</li></ul><h3>What stays private</h3><ul><li>The secret recipe your AI uses to spot fraud.</li><li>Any personal data about the customer.</li><li>The exact numbers behind each rule.</li></ul><p>All the app shows is <em>which</em> rules fired, not <em>how</em> they work.</p><h3>Why everyday users should care</h3><ul><li><strong>Faster answers:</strong> no waiting days for a bank letter.</li><li><strong>Stronger privacy: </strong>your data never leaves the vault.</li><li><strong>Fair oversight: </strong>regulators can double-check instantly, so trust is built in.</li></ul><h3>Where this idea helps next</h3><ul><li><strong>Healthcare: </strong>confirm a diagnosis meets guidelines without exposing medical records.</li><li><strong>Credit scores: </strong>prove decisions are fair without sharing your entire history.</li><li><strong>Insurance: </strong>show risk grades without handing over raw driving data.</li></ul><p>Design can turn heavy cryptography into a 30-second tap-through. If you’ve got ideas, or you’re tackling a privacy problem of your own, let’s chat. Great security should feel simple, even if the math behind it isn’t.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ffa1f7f3c2e6" width="1" height="1" alt=""><hr><p><a href="https://medium.com/design-bootcamp/catching-fraud-keeping-secrets-ffa1f7f3c2e6">Catching Fraud, Keeping Secrets</a> was originally published in <a href="https://medium.com/design-bootcamp">Bootcamp</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Revolutionizing Local Business Discovery: Enhancing Meta’s AI Search Capabilities]]></title>
            <link>https://thecreativemena.medium.com/revolutionizing-local-business-discovery-enhancing-metas-ai-search-capabilities-fdede874027d?source=rss-d9fa9cbc8869------2</link>
            <guid isPermaLink="false">https://medium.com/p/fdede874027d</guid>
            <category><![CDATA[hashtag]]></category>
            <category><![CDATA[web3]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Orezi Mena]]></dc:creator>
            <pubDate>Mon, 21 Oct 2024 20:21:38 GMT</pubDate>
            <atom:updated>2024-10-21T20:21:38.730Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*5aH4YPqJ3r1yNo-J.png" /></figure><p>Large Language Models feels like a miracle. I mean, there’s a new technology we can query about anything and it will never tell us it doesn’t know. These tech answers all our questions, even some that are outside the guideline of use if phrased properly, but trust me, I won’t be teaching you that.</p><p>We have quite a number of LLMs today, and while they compete with each other for the best response to queries, it’s not my plan to weigh in on any of these models or state my thoughts on which is the best from my own experience.</p><p>Rather, I’ve noticed that while the likes of OpenAI’s ChatGPT and Anthropic’s Claude are very powerful tools, they don’t have enough utility as Facebook’s Llama 3 or Google’s Gemini, and here’s why:</p><p>Claude &amp; ChatGPT are built by companies that don’t have products with mass adoption like the Google and Meta ecosystem. Instagram has 2 billion active users, while Facebook has 3.07 billion active users. The Meta ecosystem is wild.</p><p>The rollout of Llama into their already existing product has been impressive so far, especially with the use of AI in search, where you can do a local query or ask AI from search. But I encountered a gap in this implementation.</p><p>Hashtags have been the best way for businesses to be discoverable across platforms like Google and for Instagram ranking. But for users finding vendors and businesses to patronize, if I don’t know the hashtags to use, it gets hard to find these businesses.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Hec-QyvBNPQ5UWp2FN_F4g.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MsV4ym6GODj_x_Nmrlr4vw.jpeg" /></figure><p>I was looking for a flower vendor that could easily make delivery to my address, so I asked Meta, and Llama didn’t only fail to find me a vendor close to my area, it also failed to even find me a vendor that offers the service.</p><p>I needed a balloon vendor and I’m getting suggestions of a furniture vendor, and when I queried further, it was teaching me how to use hashtags. I was disappointed.</p><p>This creates a frustrating user experience and misses opportunities to connect businesses with potential customers. Users searching for specific local services often receive irrelevant results or general advice rather than direct connections to suitable businesses.</p><p>So, how do we solve this user experience? I want to believe it’s quite easy. I mean, I’ve never built an LLM, but I know that indexing can solve this. Here’s my proposed solution for enhanced business indexing and AI-powered discovery:</p><ol><li><strong>Comprehensive Business Profiles:</strong> We need to ask businesses to provide detailed information that would help them become more discoverable. This includes primary and secondary business categories, specific services offered, service area, delivery options, keywords related to their business, operating hours, and special features or unique selling points.</li><li><strong>AI-Powered Indexing:</strong> Develop an AI system that can analyze business profiles, posts, and customer interactions to understand the full scope of services offered. It should generate relevant tags and categories automatically and update business information based on recent activity and customer feedback.</li><li><strong>Natural Language Understanding for User Queries:</strong> Enhance the AI’s ability to interpret user queries by implementing advanced natural language processing to understand context, intent, and location. It should recognize colloquial terms and local slang for services and infer unstated needs (e.g., “flower delivery” implies local service).</li><li><strong>Dynamic Search Results:</strong> Create a more intelligent search algorithm that prioritizes local results based on the user’s location and query context, offers filters for refining search results, and provides a mix of exact and related matches to cover all potential needs.</li><li><strong>User Feedback Loop: </strong>Implement a system for users to provide feedback on search results, which can be used to improve future search accuracy, identify gaps in business listings, and prompt businesses to update their information.</li><li><strong>Business Onboarding and Education: </strong>Develop resources to help businesses optimize their presence for AI-powered discovery, including a guided profile setup process, tips for keeping information up-to-date, and analytics on how users find and interact with their business through AI search.</li><li><strong>Integration with Existing Meta Features: </strong>Ensure the enhanced AI search integrates seamlessly with Facebook Marketplace, Instagram Shopping, and Meta Business Suite.</li></ol><p>For implementation, I’d suggest starting with a pilot program in a specific category like local services in select markets. Then we can gradually roll it out to more categories and regions based on feedback. We should continuously improve by regularly updating the AI model with new data and user feedback. And for tech-savvy businesses, we could create APIs to easily update their information programmatically.</p><p>If we pull this off, we should see improved user satisfaction with search results, increased engagement between users and local businesses, higher conversion rates for businesses discovered through AI search, and more comprehensive and up-to-date business information across Meta platforms.</p><p>By implementing this enhanced indexing and AI-powered discovery system, Meta can significantly improve the user experience for finding local businesses and services. It’s a win-win: users get connected more efficiently with the services they need, and businesses increase their discoverability and potential customer base.</p><p>It’s not rocket science, but it does require Meta to focus on the user experience and really think about how people search for local services. With some smart indexing and a bit of common sense in the AI’s programming, we could turn this into a truly useful feature. I shouldn’t ask Meta for bread and get a stone.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fdede874027d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Enhancing Crypto Wallets with Abstraction: Simplifying User Experience and Expanding Accessibility]]></title>
            <link>https://medium.com/design-bootcamp/enhancing-crypto-wallets-with-abstraction-simplifying-user-experience-and-expanding-accessibility-c3e997584ae6?source=rss-d9fa9cbc8869------2</link>
            <guid isPermaLink="false">https://medium.com/p/c3e997584ae6</guid>
            <category><![CDATA[design]]></category>
            <category><![CDATA[cryptocurrency]]></category>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Orezi Mena]]></dc:creator>
            <pubDate>Sat, 27 Jul 2024 13:55:23 GMT</pubDate>
            <atom:updated>2024-07-27T13:55:23.486Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/370/1*qYh79rjU60OjFzi-7rPB8A@2x.jpeg" /></figure><p>I need to receive a payment, but I only know how to use MetaMask on the Ethereum mainnet. I have no idea how to switch to other chains, and the sender wants to use Tron’s network. While I could solve this by using my Binance app to select the token and network to receive the payment, this isn’t a sustainable long-term solution. This is where abstraction comes in. Abstraction removes unnecessary information and decision-making, allowing the process to be seamless and efficient. We often treat crypto wallets merely as storage for our assets, but they are actually gateways into the blockchain ecosystem, much like login credentials for web2 platforms like Facebook and Twitter or your banking app. By abstracting away the complexities, wallets can function more like these familiar platforms, simplifying interactions and enhancing the user experience.</p><p>Abstraction can greatly enhance the wallet experience by simplifying complex processes and creating a more intuitive interface. By hiding the intricacies of blockchain transactions, cryptographic keys, and security features, users can interact with their wallets effortlessly, focusing on essential functions without being bogged down by technical details. This streamlines transactions, enhances security by reducing user error, and improves onboarding for new users by eliminating the need to learn which wallet works for which chain — one abstracted wallet can manage them all.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/370/1*J0cgblTNa11UymnA5fRZeA@2x.jpeg" /></figure><p>Abstraction also allows for seamless integration with various services and consistent user experiences across multiple cryptocurrencies, making digital asset management more accessible and user-friendly. While bridging allows users to transfer assets across different blockchain networks, it involves complex steps that can be daunting for the average user. Bridging typically requires understanding multiple networks, managing different wallet addresses, and navigating various technical processes, which can be time-consuming and prone to error.</p><p>Abstraction, on the other hand, simplifies the entire process by handling the technical details behind the scenes, allowing users to perform transactions seamlessly without worrying about which blockchain they are using. This offers a smoother and more intuitive experience, making it easier for users to interact with multiple networks effortlessly and reducing the risk of mistakes.</p><p>Traditionally, users need to hold a native cryptocurrency like ETH to pay for gas fees, which can be confusing and inconvenient. With account abstraction, these fees can be managed in the background, potentially allowing payment in any token the user holds or bundling multiple transactions into one. This makes the process more user-friendly and removes the need for users to maintain a balance of the native token solely for fees. Additionally, it allows for more seamless smart contract interactions, making decentralized applications (dApps) more accessible and driving broader adoption of blockchain technology.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/382/1*xVddxH_viCHGH-LeIE-uGQ@2x.jpeg" /></figure><p>As the crypto ecosystem continues to grow, abstraction will play a crucial role in making digital asset management more user-friendly, promoting broader adoption, and integrating decentralized applications into everyday use. Embracing abstraction ensures that crypto wallets evolve from mere storage facilities into powerful gateways to the blockchain world, much like our familiar web2 platforms and banking apps.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c3e997584ae6" width="1" height="1" alt=""><hr><p><a href="https://medium.com/design-bootcamp/enhancing-crypto-wallets-with-abstraction-simplifying-user-experience-and-expanding-accessibility-c3e997584ae6">Enhancing Crypto Wallets with Abstraction: Simplifying User Experience and Expanding Accessibility</a> was originally published in <a href="https://medium.com/design-bootcamp">Bootcamp</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Gated Communities: The Layer of Security]]></title>
            <link>https://medium.com/design-bootcamp/gated-communities-the-layer-of-security-ab5505080da9?source=rss-d9fa9cbc8869------2</link>
            <guid isPermaLink="false">https://medium.com/p/ab5505080da9</guid>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[communication]]></category>
            <category><![CDATA[security]]></category>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[community]]></category>
            <dc:creator><![CDATA[Orezi Mena]]></dc:creator>
            <pubDate>Fri, 24 May 2024 16:32:36 GMT</pubDate>
            <atom:updated>2024-05-24T18:55:25.267Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uWScjNUZFDY1fKUYac5peg.png" /></figure><p>We are all members of a community in some sense. It could be your neighborhood, your church, or even as small as your family. All these can be classified as communities. Trespassing would be rampant if not for the rules and boundaries set in place in neighborhoods. Similarly, I believe we need such systems in the larger community called the internet.</p><p>The blockchain ecosystem has amplified the value of groups, channels, and servers (all of which are communities) by exposing participants’ information to the general public. A layer of security should be implemented so that people I don’t know personally cannot access me outside of the channel or group.</p><p>Gated chat refers to a communication system where access is restricted or controlled. This means that not everyone can freely join or participate in the chat. Various methods of gating include:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gAvwQqgkQAS64763TkTdKg.png" /></figure><p>1.	<strong>Invitation-Only</strong>: Users can only join the chat if they receive an invitation from an existing member.</p><p>2.	<strong>Subscription-Based</strong>: Users may need to pay a subscription fee to access the chat.</p><p>3.	<strong>Approval Process</strong>: Joining the chat requires approval from a moderator or administrator.</p><p>4.	<strong>Membership Criteria</strong>: Users must meet certain criteria or qualifications to participate.</p><p>5.<strong>	Tokenization</strong>: Some blockchain gated chat systems might use tokens as a means of access or currency within the chat. Users may need to hold and use tokens to participate or access certain features.</p><p>6.	<strong>Wallet Address</strong>: Can only join if the address has been listed and approved by admin.</p><p>7.<strong> Network:</strong> Users can only join with a certain network, like Ethereum mainnet or Solana.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6wtwmt8KlKHxBgcI8BgJxA.png" /></figure><p>Gated chats are often used in private or exclusive communities, professional networks, or situations where privacy and control over participation are essential. It simply means controlled access.</p><p>Also, groups are usually public. Even when you painstakingly add participants manually to a group or a channel, the link is available to all participants, and they can share these links with others. For the most part, I think this poses a risk to privacy and online safety.</p><p><strong>How Do We Combat These Issues? By Gating.</strong></p><p>If you are not 18, you can’t go to the pub. Pornhub asks you to confirm you are 18 before you can access its content (I think they can gate it better). No visa, no entry. “Money for hand, back for ground” (a Nigerian proverb meaning no payment, no service). These are examples of how things have been gated.</p><p><strong>Why is Gating Needed?</strong></p><p>For security, period.</p><p>Okay, I will explain further. You know how people can message you privately from channels or groups you share on Telegram, Discord, or WhatsApp? It has its upsides; I have personally benefited from this, but it exposes users’ data, especially on WhatsApp. I believe Discord has a layer where the user sees the message as a friend request they can choose to ignore, but my curiosity has been sparked, so why not check?</p><p><strong>How Do We Solve This Security Concern?</strong></p><p>1. <strong>User Control:</strong> Allow me to decide how private or public I want to be.</p><p>2. <strong>Permission: </strong>Confirm if I want my username or details shared with other members.</p><p>3. <strong>Direct Messaging:</strong> Confirm if I want others to DM me through the group.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zrDtyP9qxT1wAmYudJdN3w.png" /></figure><p><strong>Benefits for Admins or Creators</strong></p><p>1. <strong>Controlled Membership:</strong> Gating ensures only the desired members are in the group.</p><p>2. <strong>Monetization:</strong> It can be monetized.</p><p>3. <strong>Efficiency:</strong> Imagine using a wallet to confirm only token holders can access a channel.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*L_cVDDJXv9NLguOgPOG4bg.png" /><figcaption><em>Zk-rollups aggregate transactions into batches, executing them off-chain and validating their correctness on-chain through a validity proof.</em></figcaption></figure><p><a href="https://www.immunebytes.com/blog/zksync-a-secret-sauce-leveraging-blockchain-technology/">ZkSync</a> and <a href="https://chain.link/education-hub/zero-knowledge-rollup">Rollups</a> have recognized the need for privacy in the blockchain. With <a href="https://trustwallet.com/blog/ethereum-name-service-beginners-guide">ENS</a>, which is public, I can use your ENS name to know your transactions and volumes on the blockchain. This doesn’t have to be public, and Rollups are solving that. Selective disclosure allows only those who need to know to have access to this information. We can leverage this intelligence for communities, making private information available on a need-to-know basis, or at least give users the power to grant access.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aihSpD9U0rl5ZvRALZCdzA.png" /></figure><p>For admins or channel holders, vetting processes are easier because gating can be a form of wallet verification, code, or geographical location. This allows them to determine user access without manually performing security checks. It’s a win for everyone.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ab5505080da9" width="1" height="1" alt=""><hr><p><a href="https://medium.com/design-bootcamp/gated-communities-the-layer-of-security-ab5505080da9">Gated Communities: The Layer of Security</a> was originally published in <a href="https://medium.com/design-bootcamp">Bootcamp</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Earnipay: A Journey of Discovery, Redesign, and Growth]]></title>
            <link>https://medium.com/design-bootcamp/earnipay-a-journey-of-discovery-redesign-and-growth-d15a87a39bb0?source=rss-d9fa9cbc8869------2</link>
            <guid isPermaLink="false">https://medium.com/p/d15a87a39bb0</guid>
            <category><![CDATA[user-research]]></category>
            <category><![CDATA[finance]]></category>
            <category><![CDATA[user-centered-design]]></category>
            <dc:creator><![CDATA[Orezi Mena]]></dc:creator>
            <pubDate>Thu, 01 Jun 2023 12:55:08 GMT</pubDate>
            <atom:updated>2023-06-02T00:20:23.502Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*J9YpNgpOSF_s0vJpgFBT4w.png" /><figcaption>Earnipay’s website in 2022</figcaption></figure><p>As the product designer at Earnipay, I had the opportunity to play a pivotal role in transforming the platform and addressing the challenges it faced. After securing a $4M seed round in January 2022, we struggled with product adoption and growth, despite employing various marketing techniques and making product tweaks. The target audience for Earnipay included both employers and employees, but without the employers on board, employees couldn’t benefit from the early salary access we offered.</p><p>As a point of interest, I would like to mention that our research, design, and implementation process took eight weeks to complete. Having cleared that up, let’s get down to the details.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jgW3sr1wrFikbiyaIZMSbg.png" /></figure><p>In order to uncover the root causes of our low adoption rates, I conducted comprehensive research. Among the companies we were not able to close, I interviewed eight of them and 25 of their employees. Insights gained from these interviews exceeded our expectations and shed light on critical process issues.</p><p>During the interviews, we uncovered some key challenges:</p><ol><li>Employees often discovered Earnipay before their employers, meaning they had to recommend the platform personally. Unfortunately, we lacked a streamlined process for this crucial step.</li><li>Approximately 45% of employees expressed a desire to share in the revenue generated from transactions made through Earnipay.</li><li>Around 60% of employers expressed concerns about integrating Earnipay into their existing systems, fearing it would complicate their payroll and book balancing processes.</li><li>Notably, 60% of the participants felt that our branding positioned us as a loan company rather than emphasizing the core value we aimed to provide.</li><li>72% of the participants complained about always having to sign in whenever they open the app</li><li>65% complained that the inclusion of two input types in the withdrawal process is confusing.</li></ol><p>Armed with these valuable insights, I led the charge to find creative solutions that would address these challenges head-on. Here’s how we approached it:</p><p><strong>Change of Brand Messaging</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VRokpTna4-MPlnfBXtIDeQ.png" /><figcaption>Website Explorations</figcaption></figure><p>Recognizing the need to refine our messaging, I spearheaded efforts to explore different approaches to communicate the true value of Earnipay. We developed various alternative brand messages and experimented to find a more compelling and accurate way of conveying our value proposition.</p><p><strong>Product Segmentation</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lQzrZKDXyFtg-D0lXh_hhg.png" /><figcaption>Showing the sectioning of the website</figcaption></figure><p>Based on the research findings, we realized that solely targeting employees was insufficient. To address this, I led the team in a strategic shift to simultaneously target employers and employees. We revamped our website, creating unique sections with tailored messaging for each audience, effectively addressing their specific needs and concerns.</p><p><strong>Introduction of the Payroll System</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*w1zZx1CuMdMX0D1_WN4FeA.png" /><figcaption>Earnipay Payroll System</figcaption></figure><p>Through careful exploration of integration options, we concluded that building our own payroll management system would be the most effective solution. As the product designer, I guided the development of this system as a feature within Earnipay, streamlining the onboarding process for our users and enhancing their overall experience. It was crucial to ensure that this feature complemented existing payroll providers rather than competing with them.</p><p><strong>Improved Mobile App Experience</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fqE-vcofScLDskgN35AHkA.png" /><figcaption>Earnipay Mobile App</figcaption></figure><p>We implemented effective solutions to tackle these issues we identified from the research and improve the mobile experience. Firstly, we introduced a persistent login feature that securely stored user credentials. This enhancement eliminated the need for repetitive sign-ins, allowing users to seamlessly access the app without the hassle of entering their login details every time. By providing a convenient and time-saving solution, we significantly improved user satisfaction.</p><p>In response to the confusion surrounding the withdrawal process, we made important changes to simplify and streamline it. Through extensive user research and usability testing, we identified the pain points and redesigned the process accordingly. By consolidating the two input types into a single, intuitive interface, we eliminated confusion and created a more user-friendly experience. Users now navigate a clear and unified flow, ensuring a smooth and efficient withdrawal process.</p><p>With these solutions in place, we successfully addressed the users’ concerns and significantly enhanced the mobile experience on the Earnipay app. The persistent login feature and simplified withdrawal process have made using the app more convenient, efficient, and user-centric.</p><p><strong>Simplification of the System</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cflLUnv2f3QhGoQD9zjmew.png" /><figcaption>Simplification of the admin system for multi-organization in modals</figcaption></figure><p>During the research process, I identified a significant friction point that had gone unnoticed. Our existing system required each company, including sub-companies under a parent company, to register individually with separate login details. This complexity created challenges for administrators. To simplify the process, I led the team in designing a streamlined registration system that allowed sub-companies to be easily included under the parent company’s account. We communicated this update through a newsletter, resulting in a surge of new business registrations.</p><p><strong>And then…</strong></p><p>The realization dawned upon us that even seemingly minor changes in a product can have a profound impact on its success. Guided by user data and feedback, I made strategic decisions and collaborated with the team to implement the necessary changes within a tight timeframe of three months. The efforts paid off, as we witnessed a remarkable surge in both product adoption and revenue generation in just 92 days.</p><p><strong>Its important to mention that..</strong></p><p>As the product designer, I am proud to have played a crucial role in this transformation, leveraging user insights to identify pain points, redesign processes, and introduce new features that resonated with both employers and employees. The success of this endeavour serves as a testament to the power of data-driven decision-making and a user-centric approach.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*J5_YndsysHkDnnWG" /><figcaption>Photo by <a href="https://unsplash.com/@peet818?utm_source=medium&amp;utm_medium=referral">Pete Pedroza</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Thanks so much for taking the time to read this. Please feel free to contact me via <a href="https://twitter.com/Thecreativemena">Twitter</a>, <a href="http://thecreativemena@gmail.com/">email</a>, or <a href="https://dribbble.com/thecreativemena">dribbble</a> if you are interested in what I have to say.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d15a87a39bb0" width="1" height="1" alt=""><hr><p><a href="https://medium.com/design-bootcamp/earnipay-a-journey-of-discovery-redesign-and-growth-d15a87a39bb0">Earnipay: A Journey of Discovery, Redesign, and Growth</a> was originally published in <a href="https://medium.com/design-bootcamp">Bootcamp</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[ARC — The Case Study]]></title>
            <link>https://medium.com/design-bootcamp/arc-the-case-study-1ede38b005f3?source=rss-d9fa9cbc8869------2</link>
            <guid isPermaLink="false">https://medium.com/p/1ede38b005f3</guid>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[defi]]></category>
            <category><![CDATA[cryptocurrency]]></category>
            <category><![CDATA[nft-marketplace]]></category>
            <dc:creator><![CDATA[Orezi Mena]]></dc:creator>
            <pubDate>Wed, 17 Aug 2022 09:09:47 GMT</pubDate>
            <atom:updated>2023-09-19T22:30:32.082Z</atom:updated>
            <content:encoded><![CDATA[<h3>ARC — The Case Study</h3><p>ARC makes the entire crypto market available from the palm of your hand. We&#39;re redefining the status quo by refining all of Defi.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SgfhcgE76j0oJBuPWPDmlw.png" /></figure><p><strong>ARC is a web three solution that makes you manage your entire crypto market from the palm of your hand by connecting the world of DEX, CEX and NFT on multiple blockchains effortlessly</strong></p><p><strong>Role — Product Designer</strong></p><p><strong>Client — Arc.market</strong></p><p><strong>Year — 2021</strong></p><h3><strong>Background</strong></h3><p>The crypto world has existed since 2009 but didn&#39;t get much limelight until the Defi space&#39;s introduction in 2015. The Defi space got a bigger boom in 2017 with the rise of Centralized Exchanges like Binance and Decentralized Exchanges like Uniswap.</p><p>With the rise of crypto, new tokens get listed daily. Many of these tokens are found more on DEXs, with Pancakeswap, Uniswap, Zapper, etc., a hub to list and interact with connected to your wallet, e.g. Trustwallet. This made people have investments scattered across multiple platforms, making it difficult to manage their investments properly and estimate their worth.</p><p>Arc was devised to be the solution to unifying the crypt space, including all CEX, DEX and even NFT investments, to make the lives of crypto investors easier.</p><h3>My Role</h3><p>I came into Arc as the only designer on the project to handle the research, brainstorming with the CEO, who was the major visionary of the project. I took those ideas, improved on them, and created high-fidelity mockups that would be tested and implemented.</p><p>I was shortly joined by another designer who would work directly under me to coordinate the two initial projects of Arc, including the web3 aggregator and the NFT marketplace.</p><h3>The briefings began</h3><p>When I joined the team, there was an old design file that the previous designer had designed. Hence, my first act was to ask for feedback on the product, and why the dashboard was to be redesigned, so I started with meetings with the CEO as it was his idea to redesign the project. Some calls were scheduled in which his opinions on the current project were shared and how he felt the project didn&#39;t fully reflect his vision.</p><p>But as a professional designer, my goal is to align business goals and users&#39; needs together, so I went on ahead to test the existing product with a few members of the discord community to get the pain points from the user&#39;s perspectives as well, and this was used to draft out the new designs.</p><h3>More research</h3><p>The next step in the research phase was to look at major protocols across several aspects that were doing well in their niche. The available product has covered many niche areas, including Defi to DEX, NFT, and analysis tools.</p><p>So these were the competitors researched below<br>DEX: 1inch, Zapper, Uniswap, Sushiswap<br>NFT: Opensea, Rarible, Foundation.app<br>Analysis: Dextools, Coinmarketcap</p><p>These helped in getting an overview tho analogous; it helped to structure the UX of the product as the base shell so we can improve the design, UX, and overall value proposition of the protocol.</p><h3>Creating Personas</h3><p>The personas created were aimed at having a mental picture of the type of users we would have to be able to empathize with the users properly.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kxZD6trDctiWSLDdvY9Ucg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2OcnUeQ6m3kgQryJY0fMuA.png" /></figure><h3>Brainstorming with HMW (How might we).</h3><p>Taking further steps to refine the problems with how we might help to see the solutions from the user&#39;s pain points and find the best ways to offer value and solutions to the users. More simple and flexible flows</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GGgTRcmPlZGcG1vQu1W1QA.png" /></figure><h3>Design Decisions</h3><p>After carefully analyzing the data and working with an MVP and time constraints, the following solutions were designed below.</p><p><strong>Dashboard</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DgN8mHoXUZc2YvsZiWvdcw.png" /></figure><p>The dashboard is structured so that you both get an overview of your investment holdings. At the same time, you can toggle to view the individual holdings, view your earnings, and interact with your other exchanges. Here you can see your CEX holdings and investments and interact with them likewise your DEX investments and also interact with them like stake and unstake, also bridge and other forms of trading.</p><p><strong>Bridge</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YtI_dhAQ5lYr2cfnOsjYVw.png" /></figure><p>The bridge Arc gives was designed to keep a record of all the activities and, simultaneously, give you a clean and easy way to move the same asset from different blockchains. We also take it forward by swapping different tokens between chains.</p><p><strong>Swap with multiple routes</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VgLv2nE1IhXt4NlItSFEJw.png" /></figure><p>The swap interface was designed to allow the conversion of one token to another on the same blockchain. At the same time, we give users the ability to import tokens from any blockchain by pulling data from sources like DEX tools, Coingecko, Coinmarketcap, etc. you can swap any token in the world for any token on Arc. We give you the best routes, aggregate all protocols with such tokens, and select the best for you.</p><p><strong>Trade</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rdWaf2FtzUYRWWb1eC2mzQ.png" /></figure><p>Trading s one feature that has been very difficult to DEX due to its fast-paced nature and need for high liquidity; with Arc as an aggregator, we connect to the pools of multiple DEXs and access their treasury pools, and this gives more than enough to handle instant trading for both spot and futures trading. This gives the CEX exchange in a decentralized space 🚀</p><p><strong>Fiat On Ramp</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DHRvLKVart2G3jbaG-kryw.jpeg" /></figure><p>One of the biggest needs for most users of DEX is how to buy tokens; yes, we can manipulate the tokens from one form to another and invest when we have them, but getting them initially is the issue. SO Arc integrates Wert, a tool that allows the purchase of several tokens from payments with your card or bank account, and this allows you to purchase crypto that can be traded or swapped at will.</p><p><strong>Pools &amp; Farms</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vb-IPGTg_ozzGm84czeVEw.png" /></figure><p>Pools and Farms are investment opportunities given for non-trading investors to stake their tokens and earn passive income. By so doing, they would be contributing to the protocols treasure, which helps the speed of transactions in general. All pools and farms generate rewards that can be claimed at any time.</p><p><strong>Aggregated NFT Marketplace</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6X9JacsRA4KXTc4Cci04Bw.png" /></figure><p>The world of NFTs has become increasingly popular among investors. However, interacting with your NFTs can be a cumbersome task, especially when you have investments in multiple marketplaces. That&#39;s where Arc comes in. It simplifies the process by allowing you to connect all your marketplaces in one place. With Arc, you can easily manage your assets, accept bid offers, and even sell your NFTs without having to navigate through multiple marketplaces.</p><p><strong>Lend NFT</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*z5WX5EI4n7RDPR-0UxbW2g.png" /></figure><p>The usage of non-fungible tokens (NFTs), particularly in the GameFi industry, has become increasingly popular. To excel in some games, it is essential to possess good gear and weapons which are usually NFTs. While guilds typically provide support to their members, Arc takes it a step further by allowing users to borrow NFTs for a specific period and return them within a given time frame.</p><p><strong>Arbitrage</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pBbpGPC1GTAlHTi_Qd89qw.png" /></figure><p>Arbitrage trading is a profitable opportunity for many traders. It involves buying assets from the cheapest protocol and selling them at the highest price. Although it sounds simple, it&#39;s actually quite challenging to find such opportunities. Our platform is designed to assist you in your Arbitrage trading endeavors by providing you with all the necessary tools. You&#39;ll have access to important news, Twitter updates (since that&#39;s where most protocols announce), and a list of Arbitrage opportunities that are refreshed every 10 minutes based on the prices of assets we aggregate from various protocols. Our platform will help you identify the protocols that offer potential arbitrage opportunities.</p><p><strong>Mobile Views</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wLQeEtYTn19UFxF--3Exyg.png" /></figure><p>The majority of Arc’s users were mobile users from the research, so high priority had to be placed on the mobile designs to optimize the experience</p><h3><strong>Key Takeaways</strong></h3><p><strong>Learnings</strong></p><p>This is by far one of the most interesting web3 projects I have worked on. It allowed me to understand more in-depth the complexities of the Defi and NFT space, prioritize user needs, and simplify their burdensome activities to give maximum value.</p><p>The web3 space holds a lot of treasures and ways to keep changing lives, and I am glad to have impacted thousands for now on this one😁.</p><p><strong>Conclusion</strong></p><p>Arc is a solution that is set to bring about financial freedom to millions and affect a lot of lives, so the ease of the processes and the clarity of the tasks make it easier to drive adoption and, by so doing, invite more people that can benefit from the amazing product.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1ede38b005f3" width="1" height="1" alt=""><hr><p><a href="https://medium.com/design-bootcamp/arc-the-case-study-1ede38b005f3">ARC — The Case Study</a> was originally published in <a href="https://medium.com/design-bootcamp">Bootcamp</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>