Insurance agent runs a HeyGen clone of themselves with a strict "face budget"

What they did

They made a HeyGen photo avatar from one headshot, used the design-a-look feature for believable settings (kitchen table, home office, truck) with deliberately plain prompts like "candid phone selfie, arms length, uneven lighting", and cloned their voice from a recording. Motion prompts only describe what they want, because writing "no hands" or "no people" produced the opposite. Claude writes the script from real sources and builds each video in a cloud sandbox, in self-contained 60-to-90-second chunks, so one render becomes a long video plus six to eight clips. The cloned voice covers most of the runtime and the face appears only where it earns it. Claude then schedules posts through their browser and emails a report twice a day.

How it went

One render with about two and a half minutes of face cost 86 credits and became an 8.5-minute video plus six clips. Their own reach numbers stayed small, and a plain LinkedIn text post with real prices beat the clone videos.

Worth knowing

• They ran out of credits twice in one month. An early build drifted almost 1.5 seconds out of lip sync until the audio stitching was fixed, and one look invented people in the background. They label every upload as AI, and moved insurance-advice videos off YouTube because of its rules on AI personas giving financial advice.
• Price check: HeyGen: free plan; Creator: US$29/month, or US$24/month on a yearly plan (US$288/year) (checked Oct 26, 2026); Claude: free plan; Pro: US$20/month, or US$17/month on a yearly plan (US$200/year) (checked Nov 1, 2026).

Paying for Claude? See cheaper Claude alternatives.

Try it yourself

Paste this into a chatbot, then add your own details underneath.

Write a script for a [length]-minute video about [topic] for my [business] audience, built from self-contained 60-to-90-second sections that each make sense as a standalone clip. Use short spoken sentences, spell out numbers, one idea per section, and mark the three moments where my on-camera avatar should appear; everything else is voiceover over screens.

Original post

See Tyler Mougey's original post on YouTube → (Sep 30, 2026)