Back to works

From AI Try-On tool to AI image & video creation platform: aitryon.art product experience redesign

aitryon.art had outgrown its original try-on positioning. I reframed it as an AI image and video creation platform, redesigned the homepage generation entry, and connected that entry with tool pages, model pages, and the post-login flow. After launch, visit duration rose 83.3%, subscription conversion rose 40%, and monthly net profit was about 2.4x higher.

RoleProduct strategy, UX/UI, SEO content system

ProductAI image & video creation platform

Duration~1 month, shipped via multiple small versions

ScopeHomepage, generation entry, tool pages, model landing pages, SEO structure

+83.3%visit duration lift

+12.8%pages per visit lift

+40%subscription conversion lift

~2.4xmonthly net profit lift

01. Product Context

Before this project, aitryon.art was still mainly perceived as an AI try-on tool. Vertical tools like this could work in the early days: users were drawn in by a single try-on toy, such as clothing, color, or tattoo, completed a trial, and sometimes even paid because it happened to solve their need.

That premise started to break as general image models improved. Many lightweight try-on needs no longer required a dedicated tool. At the same time, higher-intent users cared more about model quality, batch creation, product images, short videos, and social content.

Basis for the judgment The signal did not come from one channel. Users were arriving through Google, TikTok, YouTube, Reddit, Product Hunt, AI directories, and external links. That told me the product was no longer being discovered only as a try-on tool. The market was moving the same way: general image models were absorbing simple try-on use cases, while video models were becoming part of creator workflows. The product needed to move with that shift.

Old growth logicUse vertical try-on tools to catch curiosity traffic and long-tail search, solving one-off small needs.

New user intentTry new models, generate images and videos, produce product imagery, short videos, or social content.

Product pivotFrom a tool collection to a more professional AI image & video creation platform.

02. Core Problem

The old page introduced the product, but it did not explain fast enough what aitryon.art had become. Users still had to read, choose a tool, move to another page, upload, log in, and then generate. The path had too many small handoffs.

I split the problem into three layers: product positioning had not caught up with capability expansion; the homepage lacked a strong enough generation entry; SEO and tool pages drove traffic but content structure, page templates, and downstream generation experience were not unified.

The old homepage was display-first. Users had to find a tool, enter the tool page, then upload, log in, and generate.
Outward narrative still skewed toward try-on, unable to express compound capabilities across image, video, UGC, and avatars.
Homepage generation entry and tool page forms could pass parameters, but UI style and experience rhythm were inconsistent.
History entry existed, but visuals were dated; reviewing older history required jumping between generation and history.
Model pages and tool pages had SEO potential but needed more stable templates and continuous update mechanisms.

03. Product Strategy

A visual refresh would not solve the problem. The homepage needed to become a generation entry. Model pages and tool pages also needed to work as part of the product path instead of stopping at traffic.

I prioritized shifting the product narrative from a single-point try-on tool to an AI image & video creation platform. The homepage now defaults to AI Video, because video model heat and conversion potential are higher; AI Image and Virtual Try On are still preserved so old and new users can find their own entries.

Start generation earlierUsers can upload, write a prompt, choose a type, pick a model, and continue from the first screen.

Make SEO pages useful after the clickModel and tool pages should help users understand what to generate next after they land.

Ship in smaller stepsI changed the homepage and entry stitching first, while leaving the full generation workspace for the next phase.

Keep the Studio direction openThe fragmented form, result, and history experience became the next product problem to solve.

Earlier aitryon.art landing page design draft — Before: the page felt like a lightweight try-on landing page with a content display. Platform capability was not clear.

04. Experience Redesign

The redesigned homepage lets users start before they fully understand the whole site. They can upload an image, write a prompt, choose AI Video, AI Image, or Virtual Try On, then pick a model and template. If they need to log in, their input is preserved and carried into the right tool page afterward.

Homepage defaults to AI Video to capture higher-heat, higher-conversion video generation intent.
Prompts, uploaded images, models, templates, and related generation params from the homepage carry into the tool page.
The generation page was not fully rebuilt in this version. I only adjusted color and entry stitching, because AI Studio comes next.
History entry was preserved as-is, but flagged as the key experience problem to unify in Studio next.

aitryon.art platform page after redesign — After: the product narrative moved toward an AI image and video creation platform. The image shows the structural alignment cut; live copy was adjusted later.

05. SEO Content System

SEO was not a separate content task in this project. It was part of the product path. Model pages captured demand by model name, and tool pages captured task-based intent. Both needed to guide users toward generation.

To let future model and tool pages keep scaling, I built a JSON-based copy generation skill. It can batch-produce baseline English copy while keeping examples, features, advantages, use cases, how-to guides, related tools, and FAQ modules consistent in structure, H2 phrasing, and SEO keywords. I still review the output by hand so the pages read like product copy, not keyword stuffing.

Model landing pagesCapture search intent by model name, introduce model capability, and guide users into the matching generation tool.

Tool pagesCatch task-based needs such as video generation, image generation, and virtual try-on workflows.

ArticlesEditorial content supplements long-tail keywords, tutorials, and topical pieces.

Copy skillJSON content structure improves update efficiency and keeps English SEO copy and translation logic consistent.

Earlier SEO tool page design draft — Before: early tool pages had basic SEO modules, but structure was thin and visual / content templates were not unified.

Updated dark SEO tool page — After: tool pages became a clearer traffic-catching surface, including model, examples, features, related tools, FAQ, and articles.

06. Before / After

Dimension	Before	After
Product positioning	More like a single AI try-on tool	A more professional all-in-one AI image & video creation platform
Homepage role	Display product intro and hero image	Directly captures upload, prompt, model, and generation intent
Generation path	Users had to read the page first, then find a tool entry	After login, inputs are preserved and carried to the matching tool page to continue
SEO system	Page structure and copy update flow were not unified	Model pages, tool pages, articles, and the writing skill now support the same generation path
Next phase	Generation page, tool page, and history experience were fragmented	Plan AI Studio to unify generation forms, results, and history into one workflow

07. Outcome

+83.3%visit duration: 18s → 33s

+12.8%pages per visit: 1.79 → 2.02

+40%subscription conversion: 0.5% → 0.7%

~2.4xmonthly net profit: ~$1,250 → ~$3,000

I checked whether users stayed longer, opened more pages, subscribed more often, and whether the business result moved. All four moved in the right direction after the small releases.

08. Next: AI Studio

This release fixed the positioning and made generation easier to start. It also made the next problem more obvious: generation forms, result preview, and history still lived in separate places. AI Studio is the next step toward putting that workflow in one workspace.

AI Studio interface concept - unified generation workspace — AI Studio concept: a unified generation workspace integrating model selection, parameter setup, and result preview.

AI Studio interface concept - history and workflow management — AI Studio concept: history management and workflow optimization, reducing page-to-page jumps.

Unify model selection, templates, uploads, prompts, result preview, and history into AI Studio.
Let users filter history by model, type, and time to cut down jumps between generation and history.
Recommend the right image / video / virtual try-on workflow based on user source and search keywords.
Keep expanding organic search traffic via model pages, tool pages, and articles.