How Can Agents Use AI to Draft and Produce Seller Videos?
Noah Haupt explains the AI real estate seller ad production workflow through the examples and operating choices in this video. His walkthrough connects the message, qualification process, and next action while distinguishing reported outcomes from projections.
By Noah Haupt, founder of Peak Presence
What does this video cover?
- —How do you define the seller before prompting AI
- —What structure should the AI-generated script follow
- —What does Haupt put in his ChatGPT prompt
- —Why does Haupt criticize the AI example’s consultation offer
- —How should an agent record and edit the video
- —How do you test AI-assisted scripts without confusing the result
How do you define the seller before prompting AI?
Haupt begins with an ideal customer profile containing demographics, motivations, and pain points. He contrasts an older downsizer struggling with stairs with a younger renter considering a first purchase; the same message will not address both situations.
His Charlotte example includes first-time sellers and upsizers, then explores relocation, maintenance concerns, pricing uncertainty, and coordinating the next purchase. Haupt suggests asking ChatGPT to investigate questions on Reddit and Google so the eventual script addresses a specific situation.
What structure should the AI-generated script follow?
Haupt organizes the message into a hook, a problem, a solution, and a call to action. The opening should immediately identify a relevant concern, while the body makes the homeowner feel understood before presenting the agent’s offer.
He emphasizes benefits rather than features: a guide matters because of what it helps the homeowner accomplish, not because it contains steps. He rejects choosing a short duration simply for its own sake; the script should have enough space to communicate the relevant value.
What does Haupt put in his ChatGPT prompt?
Haupt assigns the model the role of an expert real estate marketer, specifies the market and audience, and asks for a strong opening in the first five to ten seconds. His example requests conversational copy addressing pricing, timing, market concerns, and motivations such as upsizing or relocation.
He asks for short, energetic, simple sentences and a clear next action, then requests three to five variations using different angles. Haupt warns against reading generic AI output unchanged: the agent should refine it into a natural personal voice before recording.
Why does Haupt criticize the AI example’s consultation offer?
Haupt points out that a superficially acceptable script can still contain the wrong offer. In his example, a direct consultation request asks too much of a cold viewer who has not yet developed confidence in the agent.
He recommends a lower-friction entry such as a useful ebook or video guide for that audience. A direct booking invitation can become more appropriate once the viewer is already familiar with the agent; script quality and offer fit must be evaluated separately.
How should an agent record and edit the video?
Haupt recommends a clean kitchen or office, natural light facing the speaker, and centered framing. He warns against distracting backgrounds, harsh facial shadows, and framing that shows only an awkwardly cropped head.
Delivery should sound like a conversation with a friend, not a hard sales pitch. For editing he recommends smooth cuts and captions, mentions CapCut for do-it-yourself work, and favors delegating editing so the agent can spend time on higher-value business activities.
How do you test AI-assisted scripts without confusing the result?
Haupt treats the first five to ten seconds as the highest-leverage variable. The reason for generating several angles is to compare openings that address different seller concerns rather than rely on a single AI draft.
He says to keep the audience, editing approach, and other variables consistent while testing the hook. Performance expectations should reflect the market, and the agent should evaluate actual lead generation rather than assume that automation has produced a winning message.
What else should you know about this video?
How do you define the seller before prompting AI?
Haupt begins with an ideal customer profile containing demographics, motivations, and pain points. He contrasts an older downsizer struggling with stairs with a younger renter considering a first purchase; the same message will not address both situations. His Charlotte example includes first-time sellers and upsizers, then explores relocation, maintenance concerns, pricing uncertainty, and coordinating the next purchase. Haupt suggests asking ChatGPT to investigate questions on Reddit and Google so the eventual script addresses a specific situation.
What structure should the AI-generated script follow?
Haupt organizes the message into a hook, a problem, a solution, and a call to action. The opening should immediately identify a relevant concern, while the body makes the homeowner feel understood before presenting the agent’s offer. He emphasizes benefits rather than features: a guide matters because of what it helps the homeowner accomplish, not because it contains steps. He rejects choosing a short duration simply for its own sake; the script should have enough space to communicate the relevant value.
What does Haupt put in his ChatGPT prompt?
Haupt assigns the model the role of an expert real estate marketer, specifies the market and audience, and asks for a strong opening in the first five to ten seconds. His example requests conversational copy addressing pricing, timing, market concerns, and motivations such as upsizing or relocation. He asks for short, energetic, simple sentences and a clear next action, then requests three to five variations using different angles. Haupt warns against reading generic AI output unchanged: the agent should refine it into a natural personal voice before recording.
Why does Haupt criticize the AI example’s consultation offer?
Haupt points out that a superficially acceptable script can still contain the wrong offer. In his example, a direct consultation request asks too much of a cold viewer who has not yet developed confidence in the agent. He recommends a lower-friction entry such as a useful ebook or video guide for that audience. A direct booking invitation can become more appropriate once the viewer is already familiar with the agent; script quality and offer fit must be evaluated separately.
How should an agent record and edit the video?
Haupt recommends a clean kitchen or office, natural light facing the speaker, and centered framing. He warns against distracting backgrounds, harsh facial shadows, and framing that shows only an awkwardly cropped head. Delivery should sound like a conversation with a friend, not a hard sales pitch. For editing he recommends smooth cuts and captions, mentions CapCut for do-it-yourself work, and favors delegating editing so the agent can spend time on higher-value business activities.
Want this running in your market?
We work with one agent per market. On a 30-minute call we map your market and the listing number we would guarantee you.
Book a strategy call