The single biggest unlock for closing Christmas light installs online — without ever stepping foot on the property — is showing the homeowner exactly what their house is going to look like lit up. Not a generic stock photo. Not a Pinterest example. Their house. Lit up. Looking magical. In 30 seconds, from one iPhone photo they texted you.
That is what AI house mockups do. And in 2026, two tools matter: Google's Gemini 2.5 Flash Image (nicknamed “Nano Banana”) and OpenAI's GPT-4o image generation inside ChatGPT. Both are good at this exact use case. Both have trade-offs. Below is the contractor's guide to using them — with the exact prompts I use, a real before/after example, and the workflow that gets the mockup into the homeowner's email in under five minutes.
I am Jason Geiman. I scaled a Christmas light installation business from $2,000 to $1M+ with four crews, sold it in 2018, and now run Christmas Lights HQ and the 43,000+ contractor community at Christmas Lights University. We tested both Gemini Nano Banana and ChatGPT image generation across hundreds of real customer photos this spring. Here is what works.
Why AI mockups matter for Christmas light sales
Until 2024, the only way to show a homeowner what their house would look like lit up was to either drive out for an in-person walkthrough, take a guess and email a text quote, or send them a photo of a different house and say “something like this.” All three options leak deals.
The walkthrough wins on close rate but only scales so far — one truck, one route, one job at a time. The text quote loses to anyone who shows up with a visual. The “different house” example is even worse — it tells the homeowner you didn't bother to look at theirs.
AI mockups solve all three problems. You get the homeowner's actual house, lit up, in a 30-second turnaround. You can include it in an online quote you send within an hour of the phone call. That speed-plus-visual combination is what's pushing pro online close rates from the historical 20–30% range up to 40–55% in our community.
The before/after that makes this real
Below is what an AI mockup looks like in practice. Top panel is the kind of iPhone photo a homeowner texts you. Bottom panel is the AI-generated mockup of the same house lit up — what you include in their online quote email.
Mockup quality has gotten close enough to photoreal that homeowners don't think “AI” when they see it — they think “wow, that's our house.” That emotional reaction is what closes the sale. They are not comparing it to a perfect photo. They are imagining their family pulling in on Christmas Eve and feeling something.
Gemini Nano Banana vs ChatGPT: which one for Christmas light mockups
Both tools work. They have different strengths. Here is the head-to-head on the criteria that matter for Christmas light sales mockups specifically.
| Criteria | Gemini Nano Banana (2.5 Flash Image) | ChatGPT (GPT-4o Image) |
|---|---|---|
| Photo fidelity to the actual house | Excellent — windows, doors, roofline, trim color stay accurate | Good but tends to redesign minor architectural details |
| Speed | 5–15 seconds | 15–45 seconds |
| Cost per image | ~$0.04 (API) or free in AI Studio | Included with ChatGPT Plus ($20/mo) up to limits |
| Editing follow-ups (“make the wreath bigger”) | Excellent — iterative edits preserve the rest of the image | Good but sometimes regenerates more than you asked |
| Night-scene lighting accuracy | Excellent — warm C9 glow looks real | Excellent — sometimes too saturated |
| Tree wrapping detail | Good | Better — ChatGPT draws individual branch-wrapped strands more convincingly |
| Workflow ease | Google AI Studio (free), Gemini app, or API | ChatGPT app, web, or API |
| Best for | High-volume contractors quoting 20+ houses/week | Contractors already paying for ChatGPT Plus |
My verdict: start with Gemini Nano Banana. The architectural fidelity is the deciding factor — the homeowner needs to recognize their own house. ChatGPT is fine if you are already a Plus subscriber and don't want a second tool in your stack. Either will close more deals than no mockup at all.
The contractor's AI mockup workflow (under 5 minutes per quote)
- Ask for the photo on the phone. During the qualifier call, say: “Could you text me a photo of the front of the house? Daylight is best. Standing back across the street if you can.” 90% of homeowners send it within 10 minutes.
- Open Google AI Studio at aistudio.google.com (free). Pick the Gemini 2.5 Flash Image model. Upload the homeowner's photo.
- Paste your prompt (templates below). Customize the package — roofline, peaks, trees, wreath — based on what you discussed on the phone.
- Generate. 5–15 seconds. Inspect the result. If a window got changed or the trim color shifted, type a follow-up edit: “Keep the same windows but make the front door wreath 60 inches and add a red bow.”
- Download and email as part of the online quote. Subject line: “Here is what your house could look like this Christmas — quote attached.” Open rate on that subject line runs above 80% in our community.
The five AI mockup prompts that work
Copy-paste these. Customize the bracketed sections based on the package the homeowner is buying. All five are written for Gemini Nano Banana but work in ChatGPT with no changes.
Prompt 1: The full residential package mockup
Use this for most residential walkthroughs — roofline + bushes + wreath + tree.
- Warm white C9 LED bulbs along every roofline edge, 12-inch spacing, evenly spaced. Lights should glow warmly, not blow out.
- Warm white C9 LED bulbs along all peaks, ridges, and dormers.
- A 60-inch lit wreath with a red velvet bow centered on the front gable above the garage.
- Bushes wrapped in warm white mini lights, 6-inch spacing, fully lit.
- One large evergreen tree in the front yard branch-wrapped in warm white mini lights, top to bottom.
- A subtle snowy ground effect for atmosphere.
The result should look like a professional photograph of the same house in early December at dusk, magical and inviting, taken with a high-quality camera. No text or watermarks.
Prompt 2: Roofline-only minimal package
For $1,200–$2,000 starter packages.
Prompt 3: Multi-color C9 (red, green, multi)
When the homeowner specifically asks for traditional multi-color or red-and-green.
Same house, multi-color C9 mockup from Prompt 3.
Prompt 4: Commercial / HOA mockup
For HOA and commercial walkthroughs — rooflines, entry monuments, columns.
Prompt 5: The tree wrapping focus mockup
When the upsell is a 20-foot tree wrap and you want the homeowner to see the price.
Same trees, branch-wrap mockup from Prompt 5.
Iterative edit prompts: how to fix what the first generation got wrong
About 30% of first-pass mockups need one quick edit. The trick is to ask for the edit specifically and tell the AI what to keep:
- Wreath too small. “Keep everything else identical but make the wreath on the front gable 60 inches (much larger) with a bigger red velvet bow.”
- House color drifted. “Keep all the lights exactly as they are. Change the house siding back to the original gray color from my first photo.”
- Lights too bright/blown out. “Same scene, but make the lights look softer and warmer, with a more subtle glow — like a real photograph taken at dusk.”
- Bushes look wrong. “Keep the roofline and trees identical. Re-wrap the bushes with warm white mini lights at 6-inch spacing, more lights per bush, more even coverage.”
- Need a darker sky. “Same image but later in the evening — deeper twilight sky, more contrast between the dark sky and the warm lights.”
Iterate two or three times max. If you can't get it close in four generations, the source photo is probably the problem — ask the homeowner for a clearer or wider photo.
Where the mockup actually goes (the workflow that converts)
The mockup is a tool, not the end product. Where it goes matters more than how pretty it is. Drop it into:
- The proposal email. Subject: “Here is what your house could look like this Christmas.” Image embedded above the line-item pricing. The image is what they see first; the price feels reasonable after.
- The text follow-up. After a phone call, text them the mockup with one line: “Just for fun — here is what we'd put together for the front of your house. Quote sent to your email.” Texts with images get 8x the response rate of plain text.
- The Instagram / Facebook DM. If they messaged you from social, mockup as the second reply (after the price-range question). Closes social leads dramatically faster.
- The walkthrough close. Even on in-person walkthroughs, generate the mockup before you knock on the door. Have it on your phone. Pull it up at the kitchen table. “Here's what we're putting together for you.” See our kitchen table close guide for the full sequence.
- The follow-up email a week later. If they ghost after the first proposal, re-send the mockup with one line: “Wanted to make sure this didn't get lost — spots are filling up for the week of [date], let me know if you want me to hold yours.”
The legal and ethical lines (don't skip this)
AI-generated mockups of someone's home sit in a relatively new ethical space. Three rules I follow and teach in our classes:
- Get verbal consent on the call. “If you send me a photo of the house, I'll put together an AI mockup of what it could look like lit up. Sound good?” 100% of homeowners say yes. Saying it out loud covers you.
- Label the image as a mockup. Don't pretend it's a finished install photo. The email subject and caption should make it clear. “AI mockup of your home” or “Concept rendering — final install may vary slightly.”
- Never use the mockup as marketing without permission. The mockup is for the homeowner who shared the photo. Don't post it to your Instagram or use it on your site without their okay. After the install is done, ask if you can use the real photos.
The gear that lets the install actually look like the mockup
The mockup sells the magic. The gear delivers it. If the mockup shows warm white C9 LEDs at 12-inch spacing on the roofline, the install needs to match. Cheap big-box C9s won't — they get spotty within three weeks of cold weather and the homeowner's first reaction is “this looks nothing like the picture.”
Christmas Lights HQ stocks C9 LED bulbs with a 5-year warranty, Tuff Bulb clips and shingle tab clips, SPT-1 socket wire, Gilbert plugs, and pre-assembled Pro Light Kits ready to ship same-day before 2 PM ET. Free shipping on orders over $349. Use our Christmas light calculator to size any job in 30 seconds. Shop the full catalog.
Related Guides
- The Kitchen Table Close: How Pro Christmas Light Contractors Convert Walkthroughs Into Booked Jobs
- How to Sell Christmas Light Installations: A Pro Contractor's Sales Process
- How to Sell Christmas Lights to HOAs: A Pro Contractor's Sales Playbook
- How to Market Your Christmas Light Installation Business
- How to Rank for “Christmas Light Installation Near Me”
- How to Bid Christmas Light Jobs: Contractor's Pricing Breakdown
- How to Get 200+ Google Reviews for Your Christmas Light Business
FAQ
Is Gemini Nano Banana free for Christmas light contractors?
Yes, with limits. Google AI Studio at aistudio.google.com lets you use Gemini 2.5 Flash Image (Nano Banana) for free at moderate volumes. If you are quoting 20+ houses a week, switch to the paid API at roughly $0.04 per image — cheap enough that one closed deal pays for thousands of mockups.
Does the homeowner need to send me a special kind of photo?
A daylight photo of the front of the house, taken from across the street or from the driveway, works great. Wider is better than closer — the AI needs to see the roofline, peaks, bushes, and trees. Cloudy days actually work better than bright sun (less shadow contrast for the AI to work around).
What about Stable Diffusion or Midjourney?
They both work but neither is the right tool for Christmas light mockups in 2026. Midjourney is excellent for stylized art but doesn't preserve photo identity — the homeowner's house ends up looking like a different house. Stable Diffusion requires more technical setup and tuning. Gemini Nano Banana and ChatGPT GPT-4o image are purpose-built for “edit this real photo” tasks, which is exactly what we need.
Should I tell the homeowner the mockup is AI-generated?
Yes, every time. Both for ethics and for trust. The label doesn't hurt the close rate — if anything, calling it a “quick AI mockup of your house” reinforces speed and modernity. Pretending the mockup is a real photo destroys trust the moment they realize it isn't.
Will AI mockups replace in-person walkthroughs?
Partially. They are replacing in-person walkthroughs for the smaller, simpler jobs — a clean roofline + a couple of bushes is a job a pro can quote online with an AI mockup in 20 minutes. In-person walkthroughs still win on complex rooflines, big tickets ($4K+), HOA/commercial bids, and first-time customers who want to meet the contractor before they commit. The kitchen table close is still the highest-conversion move in the business — AI mockups are how you scale the rest of the funnel around it.
About the author
Jason Geiman is the founder of Christmas Lights HQ (wholesale Christmas light supplies) and Christmas Lights University (43,000+ contractor community). He scaled a Christmas light installation business from $2,000 to $1M+ with four crews before selling in 2018, and now teaches the install playbook on the roof and in the classroom. Jason is a firefighter, ASE/EVT certified technician, EMT, and hazmat responder. The mockup workflow in this guide is the same one taught at Christmas Lights HQ training classes.