Client presentations used to be my least favorite part of custom furniture work. I’d spend hours sketching, maybe build a rough mock-up, show samples of timber and finish options, and hope the client could visualize what I was describing.

About half the time, they’d say yes and we’d move forward. The other half, they’d be uncertain, ask for revisions, or decide they needed to “think about it”—which usually meant they didn’t trust that the final piece would match what they were imagining.

Over the past year, I’ve started using AI rendering tools to generate photorealistic images of custom pieces before I build them. It’s changed my client presentation process completely, and my conversion rate from quote to commission has jumped from about 50% to close to 80%.

Here’s how I’m using these tools and what I’ve learned about making them work for furniture presentations.

What AI Rendering Actually Does

The tools I’m using—primarily Midjourney and a newer platform called Veras that integrates with SketchUp—take basic descriptions or rough sketches and generate photorealistic images of furniture in realistic settings.

I can describe a piece—“mid-century modern walnut credenza with brass pulls, 180cm wide, in a bright living room with white walls”—and get back an image that looks like a professional furniture catalog photo.

Or I can upload a rough sketch and have the AI generate a finished rendering based on that sketch, maintaining the proportions and design elements while adding realistic materials, lighting, and context.

The results aren’t always perfect on the first try. Sometimes the proportions are off, or the AI adds details I didn’t ask for. But with some iteration—refining the description, regenerating a few times, sometimes combining elements from different outputs—I can usually get something very close to what the final piece will look like.

Why This Works for Client Presentations

The biggest value is reducing uncertainty. Clients who commission custom furniture are often nervous. They’re spending thousands of dollars on something that doesn’t exist yet, based on their trust that I can deliver what they’re imagining.

Showing them a photorealistic render that looks like the finished piece in their actual room dramatically reduces that uncertainty.

I recently worked with a client who wanted a custom dining table in blackbutt with a natural oil finish. I showed them timber samples and talked through the design. They were interested but hesitant.

I generated three AI renders showing the table in different room settings with different lighting conditions. Seeing the table in context—understanding how the grain would look, how the finish would interact with light, how it would fit in a room—gave them confidence to commit.

They told me later that the renders were what convinced them. Without those images, they probably would’ve gone with a ready-made option because it felt less risky.

The Workflow I’m Using

My process now looks like this:

Initial consultation with the client to understand what they want. I take notes, measurements, photos of their space if relevant.

I create a rough sketch or basic SketchUp model of the piece. This doesn’t need to be detailed—just proportions, basic form, key design elements.

I use AI tools to generate 4-6 renders based on the sketch and description. I’ll typically try different angles, different room contexts, different lighting to give the client a full sense of the piece.

I review the renders, select the 2-3 best ones, and sometimes do minor edits in Photoshop if there are small errors the AI introduced.

I present the renders to the client along with timber samples and finish options. The renders become the visual anchor for the conversation.

The whole process from initial sketch to final renders takes me about 2-3 hours for a typical project. That’s comparable to what I used to spend on detailed hand sketches, but the results are much more convincing.

Where AI Rendering Falls Short

These tools aren’t perfect. There are consistent problems I’ve learned to watch for.

Joinery details are often wrong. The AI might show dovetails where I described finger joints, or render hardware that doesn’t exist. I have to either fix this in post-processing or clearly explain to the client that the joinery in the final piece will be different.

Proportions can be off. The AI sometimes exaggerates or minimizes dimensions in ways that look good visually but aren’t accurate to the spec. I’ve learned to call this out proactively and remind clients that the render is for concept visualization, not dimensional accuracy.

Material rendering is inconsistent. Sometimes the AI nails the look of oiled blackbutt. Sometimes it looks more like pine or some generic “wood.” I’ve found that the more specific I am in my descriptions—“Australian blackbutt timber with visible grain, natural oil finish showing honey tones”—the better the results.

The tools also struggle with unusual or complex designs. They’re trained mostly on conventional furniture styles. If I’m designing something genuinely unique or structurally complex, the AI often can’t render it accurately.

For those projects, I still rely on detailed sketches or SketchUp models.

The Client Education Piece

One important thing I’ve learned: clients need to understand what they’re looking at.

When I present AI renders, I explicitly say “this is a visualization tool showing you what the general look and feel will be, but the final piece will differ in details.” I point out areas where the render is conceptual versus accurate.

Without that context, clients sometimes treat the render as a literal representation of what they’ll receive. Then they’re disappointed when the final piece doesn’t exactly match the render.

Setting expectations upfront prevents those issues.

The Cost Question

Most AI rendering tools have subscription pricing. Midjourney is about $30/month for the standard plan. Veras is around $50/month. There are free options like Stable Diffusion, but they’re harder to use and require more technical setup.

For me, spending $50-80/month on rendering tools is easily worth it given the improvement in conversion rates and client satisfaction.

I used to lose maybe one commission per month because clients couldn’t visualize the design and decided not to proceed. Each of those lost commissions was typically $3,000-8,000 in revenue. If the rendering tools help me close even one extra project every couple of months, they pay for themselves many times over.

Integration with Traditional Skills

AI rendering isn’t replacing traditional furniture design skills. I still need to know how to design structurally sound furniture, select appropriate materials, understand joinery techniques, and build the piece.

What AI rendering does is fill a gap in the process between design and fabrication. It makes the design legible to clients who don’t have the experience to visualize furniture from technical drawings.

I think of it as a communication tool, not a design tool. It helps me communicate what I’m already planning to build.

That said, I’ve also found it useful for exploring design variations quickly. I can generate renders of a piece with different leg styles, different hardware, different finishes, and show the client options in minutes rather than hours.

That flexibility in the design conversation has led to better final designs because clients feel more comfortable expressing preferences and experimenting with options.

Where I Got Help

Learning to use these tools effectively took time. The basic use is straightforward—type a description, get an image. But getting consistent, useful results for furniture presentations required understanding how to write effective prompts, how to control style and detail, and how to integrate the tools into my workflow.

I worked with their AI consulting practice to set up the workflow and learn best practices for furniture visualization specifically. That accelerated the learning curve significantly and helped me avoid common mistakes.

What Other Furniture Makers Are Doing

I’ve talked to a few other custom furniture makers about this. Reactions are mixed.

Some are enthusiastic and using AI rendering extensively. Others are skeptical and prefer traditional presentation methods.

The common objection is that AI renders feel impersonal or “fake” compared to hand sketches. There’s something about a hand-drawn sketch that communicates craft and care.

I get that perspective, but I think it’s a false choice. I still do hand sketches for my own design development. But for client presentation, I want the most effective communication tool, and photorealistic renders work better than sketches for most clients.

There’s also some concern that AI rendering will commoditize custom furniture by making it easier for clients to comparison shop or for less-skilled makers to present designs they can’t actually execute well.

That’s possibly true, but I think the solution is to be clear about the connection between design visualization and fabrication skill. The render is just a picture. The value is in building the piece well.

The Practical Reality

AI rendering tools are now good enough and accessible enough that any custom furniture maker can use them. You don’t need technical expertise or expensive software.

The tools aren’t perfect. They require some learning and experimentation. They work better for some design styles than others. They don’t replace traditional skills.

But they solve a real problem in the client presentation process, and they make it significantly easier to close commissions because clients can see what they’re buying.

If you’re doing custom furniture work and you’re not using AI rendering for presentations, you’re probably losing commissions you could be winning.

The tools exist, they’re affordable, and they work. The question is whether you’re willing to integrate them into your workflow.

Based on my experience over the past year, the answer should be yes.