Vyshnevskyi

Case · 06 / Go get'em

Go get'em

  • Competitive intelligence
  • AI tool
  • SaaS
  • Solo build · 8h
  • Next.js

My role

Solo build — narrative, UX, visual, AI pipeline, deployment. 8 hours from blank repo to live URL. The constraint was a product design problem, not a dev one.

Outcome

A working AI competitor-analysis tool, shipped solo in 8 hours. Drops a URL → auto-identifies three rivals → returns scored design recommendations with real screenshots and exact copy diffs. Live URL — not a Figma file.

01

Overview

Most SaaS founders lose conversion to competitors but never know exactly why. Go get'em delivers the diagnosis: drop your URL, get a six-dimension design audit benchmarked against your top three competitors — with exact copy rewrites and screenshot evidence.

02

Process

IdeaSolution

Step 01 / 06

The process

NARRATIVE ARCHITECTURE

Storyboarded the page as one continuous argument: you're losing but don't know why → manual comparison is broken → we show you exactly where → we tell you exactly what to fix → you act now. Every section and line of copy was made to ladder into that arc.

03

Shape

before

code

Shape before code

The page was a five-beat argument before it was a single component. Each section is one move in the arc — drawn as a wireframe first, so the structure carried the story and the code just dressed it.

Single-page scroll architecture · five beats

Swipe through the five beats →

01Hero

See exactly what your competitors do better

Cold open. The promise, a 3D blob, one CTA.

02Problem

See where you fall behind

You're losing but don't know why.

03Solution

Benchmarked against the best

Identify, analyze, benchmark — one pass.

04Result

Changes you can ship today

Not "improve your CTA" — the exact rewrite.

05Input · CTA

Drop your URL

One field gates the whole pipeline.

04

Landing

page

Landing page

A cinematic dark scroll experience — a hero with a 3D blob and three editorial storytelling beats: Problem, Solution, Result. The whole page walks the visitor from "see where you fall behind" to "changes you can ship today" before they ever touch the tool.

Landing page 1
Landing page 2
Landing page 3

05

URL

input

URL input

One field, one gesture. Drop a SaaS landing-page URL and the tool kicks off: Firecrawl scrapes the page, Claude identifies the top three direct competitors, and the vision pipeline benchmarks against them.

06

Analysis

results

Analysis results

Claude Sonnet 4.5 vision scores the page across six dimensions, 1–10, each backed by specific evidence and a concrete rewrite. Real output for Jira: 1/10 on Pricing Transparency (no pricing anywhere on the page), 6/10 on CTA Strength (generic "Sign up" vs Asana's "Get started · View demo"), 7/10 on Hero Clarity vs Linear's sharper positioning.

Analysis results 1
Analysis results 2
Analysis results 3

07

The

full

report

The full report

The first analysis is free — the full multi-page report sits behind a soft gate. "Don't stop at one page": drop your email for the complete breakdown, or subscribe for unlimited analyses. Mock Stripe stayed mock; the paywall was a scoping decision, not a build-out.

Final

thoughts

Final thoughts

The product was defined before it was built. Flows, positioning, monetization, and tech direction were locked upfront — so AI could generate 95%+ of the code while every product decision stayed human.

The lesson is focus, not speed. Every feature was tested against one question: does it make the core experience better? Everything else got cut.

AI handles execution. Humans hold the calls.

Scroll to top