Pop Girl-Group AI Song Mashup

I have three separate pop tracks sung by a virtual three-member girl group. I want them blended into a single, release-ready record running between four and five minutes. Each singer’s part has already been generated, but some words were mangled by the AI vocalist; I need those moments repaired so the lyrics are perfectly intelligible while still sounding natural—no obvious Auto-Tune artifacts or robotic edges.

Here’s the full scope:

• Seamlessly arrange the strongest sections of all three songs into one coherent structure (intro, verses, hook, bridge, outro).
• Replace or re-synthesize every mis-pronounced word so it matches the written lyric sheet, keeping each singer’s unique tone. If AI resynthesis can’t do it cleanly, creative comping or manual re-recording techniques are fine—what matters is that the corrections feel organic.
• Polish the mix: balance three distinct female vocals, layer any supporting harmonies, add tasteful pop production touches, and ensure the track punches on phones, earbuds, and club systems alike.
• Master to modern streaming specs (-14 LUFS integrated, ‑1 dBTP) and provide a Spotify/iTunes-ready 44.1 kHz/24-bit WAV plus a high-quality MP3. Please also deliver instrumental and a cappella stems.

I normally work in Logic Pro and can pass you the stems; if you prefer Ableton Live or another DAW that’s fine—just keep all routing clear so future revisions are simple. Melodyne, iZotope RX, or similar vocal-repair tools are welcome, but however you do it, the final result should sound like a professional studio recording, not an AI experiment.

If this sounds like your sweet spot, let’s talk timing and any questions about the lyric sheet or reference mixes.