Precision Affiliates Edge Method

Production sheet

Ad 01: You don't have the slightest clue what we do in Norway to never get type 2 diabetes.

Top: what needs your call. Middle: the ad as approved so far, stage by stage. Bottom: the archive, collapsed. Rejected files are off the page. Open the approved script

Needs your review

9 waiting

Everything below is new and not decided. Approve the one to carry forward (one per group), Reject what is out, then Submit.

2. Voice: A: 35 years in the Midwest, mostly American rhythm with a Scandinavian lilt
ElevenLabs A-1candidate
Voice Design v3, 10.8 s.
10.8 s · 252 KB · download
ElevenLabs A-2candidate
Voice Design v3, 13.7 s.
13.7 s · 321 KB · download
ElevenLabs A-3candidate
Voice Design v3, 12.7 s.
12.7 s · 298 KB · download
2. Voice: B: deeper, 70 percent American 30 percent Norwegian
ElevenLabs B-1candidate
Voice Design v3, 11.4 s.
11.4 s · 266 KB · download
ElevenLabs B-2candidate
Voice Design v3, 12.4 s.
12.4 s · 289 KB · download
ElevenLabs B-3candidate
Voice Design v3, 12.3 s. Says 'clue of what' (one extra word), fine for a voice pick.
12.3 s · 287 KB · download
2. Voice: C: mostly American, accent only on some vowels and the rhythm
ElevenLabs C-1candidate
Voice Design v3, 9.3 s.
9.3 s · 218 KB · download
ElevenLabs C-2candidate
Voice Design v3, 10.0 s.
10.0 s · 233 KB · download
ElevenLabs C-3candidate
Voice Design v3, 11.4 s.
11.4 s · 266 KB · download

1. Portrait (Kristian)

The still every clip is animated from. Nano Banana Pro via KIE, 9:16. One pick carries through the whole script.

p2 clean
p2 cleanapproved
Grey fleece, oak kitchen. Chest logo removed by an edit pass, face unchanged. Ryan's pick 12 Oct 2026. This is the face in every clip below.
768x1376 · 269 KB · download

2. Voice

Pick ONE. It is saved as the permanent Kristian voice and every script renders from it. Preview text: hook 1 plus the first body sentences, 10 to 22 s.

9 waiting on you at the top of the page.

3. Video

Lip-synced clips: portrait plus the audio file into Kling AI Avatar via KIE. The voice is whatever audio went in; Kling adds none of its own.

Nothing approved yet.

Note to Claude