Blog 006: Building with AI
This is a brand new project.
I’ve been wanting to stretch my design muscles in a real way. Not just prompts or exercises, but something I can actually build, test, and see evolve. And now with AI changing the game, it feels like the perfect time.
So this is me, trying stuff.
As a product designer, I want to learn by doing — building something real from the ground up. This journal is where I’ll track the whole thing. No polish, no fluff. Just notes from the road.
Who knows, maybe it turns into a proper portfolio piece later. Or maybe it catches the right eyes along the way. Either way, I’m showing the work. Let’s see where it goes.
Here is the product as it stands now: PAIRSONNA
-
Everything.
Seriously — I’m doing all of it right now.UX and product design
System and feature architecture
AI matchmaking logic
Prompt writing and testing
Brand voice and tone
Troubleshooting platform limitations
Journaling the process 👋
Eventually, I’ll hand parts of this off. But for now, it’s just me — trying to figure it out one step at a time.
-
Pairsonna is a digital platform I’m building to connect creators and makers into real partnerships — not just transactions. The idea is simple: instead of hiring someone or buying from them, you build something with them. Think co-owned product lines, shared storytelling, and collaborative launch tools. Less Etsy, more partnership studio.
-
The goal is to create a platform that facilitates meaningful, equitable partnerships — with built-in tools that support not just product development, but also trust, ownership, and storytelling.
It’s not about churn-and-burn dropshipping. It’s about slow commerce, shared vision, and making cool stuff with the right people.
-
UX Complexity: Two user types (makers and creators), multiple roles, overlapping needs — and it all has to stay simple and clean.
Shifting Scope: Started small, then kept uncovering new layers: storefronts, AI matching, project containers, messaging logic, and more.
No-code Constraints: Building everything in Lovable (and previously Webflow) means bumping into platform limitations that force creative problem-solving.
Balancing Vision + MVP: Trying to stay focused on the core experience without losing the bigger picture.
-
I’ve spent years floating between design disciplines — from architecture to furniture, from graphics to UX — and one thing I’ve always felt missing was the space between. That fertile, messy, creative middle where people come together to make something bigger than either could alone.
I wanted to build a tool that makes those kinds of connections easier to start, and stronger to sustain. Especially now, with AI and e-commerce changing everything, it felt like the right time to reimagine how we collaborate.
ENTRY 001 - WHERE IT ALL STARTED
Funny enough, this whole thing kicked off while I was interviewing at Shopify. I didn’t end up getting the job, but in the middle of the process I went down this rabbit hole about e-commerce trends, creator pains, all that.
I stumbled on an article from Dovetale, and one part stuck like glue: it talked about how Shopify could help smaller creators by pairing them with makers — real partnerships, not just merch drops or influencer deals. Like, imagine a YouTube chef teaming up with a blacksmith to sell knives. That hit me hard.
I couldn’t stop thinking about it. Creators have audiences. Makers have skills. Why isn’t there a space that brings them together for something real — like co-ownership, shared revenue, actual products with stories behind them? That little spark is what lit the fuse for what would eventually become this project.
ENTRY 002 - RESEARCH MODE
Today I sat down and looked at what’s already out there. Etsy, Pinkoi, Maker’s Row, Artisans Cooperative, DirectCreate… all of them kind of circle around this space but none of them actually solve the thing I’m obsessed with. They all focus on transactions. Buy, sell, done.
What’s missing is the full partnership piece. Helping a creator and a maker go from idea to product to sales, with the support they’d need along the way. Things like legal templates, revenue splits, a shared dashboard to manage it all. None of the current platforms even come close to that.
The research gave me more confidence. There’s definitely a gap. I just need to figure out if I can shape this idea into something people will instantly get.
ENTRY 003 - WRITING IT OUT
I decided to put the idea into a full business proposal. Mostly to see if it still made sense once I got it out of my head and onto paper. I pulled together all the research, added market data, even laid out a roadmap and competitor breakdown.
It read kind of like a startup pitch. Big market numbers, phases, features, all that. Honestly, it felt a little overbuilt, but that was the point. I wanted to stress test the idea. Would it still hold up if I shared it with friends? Would they get it right away or poke holes in it?
What I learned is that people do get it. The gap is clear. The challenge is making sure it doesn’t sound too heavy or complicated. I don’t want this to feel like a corporate play. It has to feel personal and human, because that’s what the whole idea is about.
ENTRY 004 - FINDING THE NAME
“CreatorConnect” worked as a placeholder, but it was way too generic. Felt like a tech tool, not something alive. I kept tossing names around until one finally clicked: Pairsonna.
Pair + persona. Pair for the connection. Persona for the identity, style, personality. Together it means pairing people based on who they really are, not just numbers or skills.
The second I said it out loud, it felt different. Suddenly the project had a name with story and weight. Something you could imagine on a landing page, with a clean logo and a tagline like “Pair with purpose.”
That’s when it stopped being an idea floating in my head and started feeling real.
ENTRY 005 - COMPETITIVE RESEARCH & FEATURE MAPPING
I went deeper into research today, this time mapping out features side by side with what’s already on the market. Etsy, Pinkoi, Maker’s Row, Artisans Cooperative, DirectCreate… I lined them all up and started asking, “What do they actually give people, and where are the gaps?”
The answer was pretty clear. They all stop at selling. None of them help creators and makers actually partner up, share ownership, or manage a project from start to finish. That gap is huge, and it’s exactly where Pairsonna fits.
So I sketched out what would make Pairsonna different:
End-to-end support with legal templates, contracts, and revenue splits.
A dashboard that becomes the shared workspace.
AI-powered matchmaking so people don’t waste time scrolling endlessly for the right fit.
Storytelling as a core feature, not an afterthought.
Looking at it all together, it feels like the idea has a real backbone. Not just a concept anymore, but a structure I can actually build on.
Longer version (before the super long version):
Snapshot: Competitors
Pairsonna – Collaboration-first. AI matchmaking, shared dashboard, legal + revenue split tools, storytelling pages, integrates with Shopify/Etsy. Not a marketplace; focuses on co-creation.
Etsy – Marketplace giant. Huge audience, easy to list/sell. No collaboration tools, no revenue sharing, minimal vetting. Purely transactional.
Pinkoi – Curated design marketplace. Quality-vetted Asian designers, stylish catalog, cross-border sales. No collaboration or co-ownership. Strong curation, weak creation support.
Maker’s Row – US manufacturing directory. Connects entrepreneurs with factories, supports scaling. No storytelling, no revenue split, limited project tools. Great for sourcing, not partnerships.
Artisans Co-op – Co-op marketplace. Seller-owned, fair fees, strict handmade criteria, community ethos. No matchmaking, no project management. Fairer Etsy, but still just sales.
Direct Create – Craft collab network. Connects Indian artisans with designers, some project collaboration, low fees. Less tech/AI, limited to heritage crafts. Closest to Pairsonna, but regional and less structured.
Note on Research
I’ve been using ChatGPT as a sidekick for research. First, to map the competitive landscape — who’s out there, where they overlap with Pairsonna, and what makes us stand apart. Second, I set up an agentic AI to scrape the web for real pains and needs from creators and makers, pulling in direct quotes with sources. That part’s still ongoing.
Everything it surfaces, I read through and verify myself. It’s been a surprisingly fast way to spot gaps and patterns, but it still needs my eyes on every word, as it should.
ENTRY 006 - HOW TO BUILD IT
Today I hit the big question: do I build everything from scratch, or plug into tools people already use?
On one side, building it all myself would make Pairsonna feel super integrated and unique. On the other, it would take forever and cost way too much. Meanwhile, creators and makers already live on Slack, Trello, Google Drive, Shopify… why not meet them where they are?
So I landed on a hybrid approach. The core stuff — matchmaking, revenue splits, legal templates, storytelling — has to be native. That’s what makes Pairsonna different. But things like messaging, scheduling, file storage? Better to integrate existing tools through APIs. Slack for chat. Trello for tasks. DocuSign for contracts. Shopify for sales data.
It keeps the early version lean, faster to build, and still powerful. Down the road, if users rely heavily on one feature, we can replace it with something custom. For now, the goal is speed and adoption without losing what makes Pairsonna special.
Right!? Ok, let’s see where that goes…
ENTRY 007 - VISUALIZING THE DASHBOARD
While pushing the boundaries of how much AI can tackle in this process, I tried tossing a prompt into Uizard just to see what would happen. Honestly, it was more of a fun experiment than anything else. The UX was kind of a mess, the UI was meh, but nothing I couldn’t whip up faster in Figma if I already knew the flow.
That said, it did spark some ideas for MVP. Seeing co-creating tools, calendars, even a few analytics widgets pop up on screen was interesting — like a little glimpse into where these kinds of tools could be heading. Even if everything felt out of place, it was still worth the try.
It’s still rough sketches at this stage, but seeing it mapped out made the whole thing feel tangible. Not just an idea — something I could actually picture people clicking around in.
Here is a user flow / dashboard architecture of what came to mind ↓
ENTRY 008 - STORYTELLING AT ITS CORE
The more I think about it, the more I realize storytelling has to be front and center. It’s not just about selling a product. It’s about why the product exists, who made it, and what it represents.
Most platforms completely skip this. They show you a photo and a price. Done. But people don’t just buy things, they buy into stories. The artisan’s craft, the creator’s vision, the journey from raw material to finished piece — that’s where the real value lives.
So I imagined a “Storytelling Showcase” right inside the dashboard. A place where both the creator and the maker can upload photos, videos, sketches, even little anecdotes. They’d build a product story together, in real time, as the project develops. At the end, you’d have a full “story page” ready for the shop, social media, or marketing campaigns.
This feels like a real differentiator. It gives products emotional weight and makes the collaboration visible to the world. That’s the kind of thing that could set Pairsonna apart.
ENTRY 009 - PULLING IT ALL TOGETHER
At this point I stepped back and looked at the whole thing. Matchmaking, legal templates, dashboards, integrations, storytelling, revenue splits… it’s a lot, but together it paints a pretty clear picture of what Pairsonna is meant to be.
The big vision is simple: a platform that makes it easy for creators and makers to team up and actually build something together. Not one-off merch. Not quick influencer deals. Real partnerships that have structure and support from start to finish.
It feels ambitious, maybe even a little crazy, but the pieces make sense. If it works, it could change how creators and artisans connect, scale, and tell their stories.
Now it’s about figuring out how to get from this big picture down to an MVP that proves the idea.
ENTRY 010 - WEBFLOW EXIT / LOVABLE ENTRY
So... I hit the wall with Webflow.
It was working, kinda — but the minute things needed actual logic, the no-code ceiling came crashing down. I was duct-taping interactions just to fake real functionality. It stopped feeling like progress and more like wrestling a CMS with pretty buttons.
The tipping point was realizing I needed a real dashboard flow. Something dynamic. Something that responds to users, not just static clicks. So I paused, poked around for alternatives... and stumbled into Lovable.
Didn’t know what to expect. Turns out, it’s exactly what I needed.
ENTRY 011 - LOVABLE, DAY 1
Alright, jumping in.
Lovable drops me into this VS Code-style builder with chat. Pretty barebones, kinda intimidating at first — but also low-key thrilling. I’m not coding per se, but I’m closer to code than ever. I tell it what I want, it scaffolds the thing. Still need to think like a builder though, which I like.
Got my basic app shell set up. Home page. Routing. Environment variables dropped in. Supabase wired. Starting to feel real.
The "auth" setup makes me sweat a bit, but we work through it together — Lovable tells me where to go, I click and follow along. Wild how fast stuff compiles. This might work.
One more thing.
I’m not building this completely alone. Alongside Lovable, I’ve been using ChatGPT pretty much as my design sidekick. It’s been helping me think through ideas, shape clear prompts, and troubleshoot stuff when I hit a wall inside Lovable.
It’s kind of like having a second brain around — especially when I’m navigating tools I’ve never used before.
ENTRY 012 - FIRST REAL OUTPUT > DASHBOARD
The first major "aha" moment — I’m looking at an actual user dashboard, and it’s mine.
Built the MVP layout with cards for each action: discover matches, storefront preview, messages (placeholder for now), submit feedback. Super basic, but already way more dynamic than anything I could fake in Webflow.
Feels like the first time this project looks like... a product. Still rough, but real.
ENTRY 013 - SUPABASE PUNCH TO THE FACE
Oh man. The Supabase part hits hard.
Spent hours fighting a 406 error that made zero sense. Lovable said the linter was showing cached results, Supabase wouldn’t budge, and I had no clue what a “view” or a “query” even meant at this point.
Tried deleting stuff, re-running code, poking around the DB... nothing. Eventually, Lovable fixed it via tool call. Apparently the "security definer view" had to be updated. I still don’t fully get it — but hey, it’s gone. I’ll take the win.
ENTRY 014 - SCAN INSTAGRAM TO STOREFRONT MVP
This was fun.
Here’s the thing — while the big vision behind Pairsonna is rooted in storytelling and building something not purely transactional, I wanted to test how fast I could validate a quick, sticky idea. This one came to mind first: what if someone could build a shoppable storefront from their IG, instantly, without needing to be a product creator themselves?
User enters IG handle → gets dropped into a storefront preview with 3 recommended products → can claim the store → dashboard updates.
Even added a little dice icon so users can shuffle product matches. Like a mini “roll the dice” moment for curation. Also added the ability to weigh in on tags. That one took a few prompts, but it works. Love how easy Lovable makes tiny UI interactions feel.
It felt light, fun, and still connected to the bigger goal. So I ran with it. Check out the parts that were built in the images below.
The UX is still kind of a mess.
That’s the downside of building in Lovable — once you start editing things live, the working flow gets shaky. Feels like I’m tripping over my own shoes.
So I’m slowing things down. Gonna zoom out, focus on the UX and architecture first. Less back and forth later, cleaner builds ahead.
ENTRY 015 - CLAIM THIS STORE FLOW
When the user clicks “Claim This Store”... what now?
That’s where I got stuck. Lovable helped scaffold a confirmation modal and button logic, but I realized fast — we hadn’t thought through what happens after the claim.
Should they manage the store? Customize it? How many stores can they have? What if they want to skip Instagram and build manually? I wrote a bunch of these questions down. Most of them still open.
For now, I’ve locked it down to one store per user, and we pushed a dashboard update that surfaces your claimed store if it exists, and the “Claim” button if it doesn’t.
Still not perfect — but forward motion.
ENTRY 016 - RE-RE-ALIGNING THE FLOW
Big reset today.
I’ve been trying to figure out how Scan-to-Storefront really fits. At first, I wanted it as one of the main dashboard tabs, but digging deeper into the architecture made me realize it was pulling too much weight. Pairsonna isn’t supposed to be a dropshipping tool. It’s not about passive storefronts. The whole point is creative partnerships — shared ownership, collaboration, story.
So Scan-to-Storefront got reframed. It’s not the product. It’s the entry point. A way for someone to get a taste, to see something generated quickly, and then get nudged into building a real project with a maker. More like an onboarding spark than a full tab.
That realignment felt important. It sharpens the focus: every core feature in the dashboard has to deepen collaboration or amplify the shared story. Anything else is an experiment for later.
Now I’m back to the basics: what actually belongs in each part of the dashboard. Starting with the overview panel.
ENTRY 017 - BREAKTHROUGH MOMENT
Been circling the dashboard for days now. Not designing it yet, just trying to understand it. What even is a dashboard when the whole platform is about collaboration, but everyone’s bringing different roles and goals to the table?
The big unlock came today: it’s not about people. It’s about projects. Everything orbits around them. Matching happens project-to-project, not profile-to-profile. That clears up so much.
Before this, I was trying to cram too much into profiles. But a maker might work on five different projects, each one needing a different type of creator. A creator might spin up three ideas, all asking for totally different kinds of support. So tying everything back to a profile just... didn’t fit. Projects are the real heartbeat. That’s where the story is.
So here’s the pivot:
The dashboard becomes a contextual workspace.
You choose the project you’re working on — and the whole dashboard shifts into that context. Like Slack switching workspaces. Or Notion switching pages. The tabs on the left stay the same, but they now only show content tied to the active project. Contracts for Mushroom Mug Drop. Tasks for Mushroom Mug Drop. Revenue for Mushroom Mug Drop. You get the idea.
We’ll still keep a "Home" or "My Projects" view where you can see everything zoomed out. That’s the place to create new ones, check on overall progress, or switch context.
I love how clean and scalable this feels. I’m not adding more screens, I’m just shifting how to frame them. It’s more intuitive. It also aligns with how I want people to communicate — not profile to profile, but project to project. All interactions, all collaboration, is centered around the thing being built.
Feels like a huge unlock.
ENTRY 018 - PIECES FALLING INTO PLACE
There were questions around communication. Should people chat freely like a social platform? Nah. We’re keeping it focused. All interactions happen through a project. That keeps the intent clear, the context shared, and hopefully cuts down on random noise. You don’t message a person — you respond to an invite or a shared idea tied to a project.
One big UX thread I untangled: ROLES. You can’t be both a creator and a maker on the same project. That clears up a lot of ambiguity and keeps the structure solid. It also makes it easier to design the invite flow and permissions.
I’m still holding off on deeper storefront features, but made a note to revisit how they’ll live under Settings or maybe evolve into their own thing later. The important bit was locking in the first experience inside “My Projects” — seeing all your builds, a snapshot of activity, and smart nudges like pending invites or new messages. Even defined the empty state with a friendly nudge to get started.