Your Questions on Gendo — Answered

Over the course of our recent Gendo Live sessions — From Sketch to Concept, Styling with Intent, Enhancing Your Work, and Workflows in Practise — you asked us nearly 250 questions. Some were short and sharp, others more exploratory

I’ve gone through every question and grouped them here. Below are my answers honest, direct, and from the perspective of someone who has been building Gendo alongside your feedback.

Materials & Rendering

“Can I have precise control of geometry and material surfaces?”

The platform is built to stick with your design as tightly as possible, whilst still giving you freedom to test surface finishes. You’re the designer, and we assume geometry changes will be made by you and re-uploaded. If you say “vertical cedar cladding with black zinc roof”, that’s exactly what Gendo will aim for. Precision in input + precision in language = precise results. Using the material tool you can select and make change and also add material references to ensure alignment with your vision.

“How do I apply materials, textures, and lighting effectively?”

So, I would handle these 3 things in 2 steps.

Step 1 would be to get the atmosphere and lighting right in the first iteration of the image. So, in accordance with our prompt formula, I would add keywords and descriptions for ‘atmosphere & lighting’ - these would be things like ‘blue sky, sunny day’ or ‘soft ambient interior light, low lighting’ or ‘warm evening light, golden hour’.

You can and should add reference to specific materials you want here either as part of ‘key features’ or ‘extra keywords’ - again following our prompt formula. Finally, before hitting generate, I would then always enhance my prompt with the prompt enhancer, and double check the output.

If you prompt well, use a good input and a good reference image you should get beautiful output - but it is unlikely to be 100% perfect - and this is what the material tool is for.

So once you have an image which has good atmosphere and lighting and is 80% of the way there, I would move onto step 2 - material tool.

Here we can adjust surfaces for specific material finishes you want to see - which will also be in line with the lighting in the rest of the image.

Add a good reference image, a close up texture sample would be ideal, and prompt for that same material. So if I’ve given a corten steel texture map, I would then prompt something like ‘red rusty metal, corten steel, metal sheet, exterior’ - and then enhance prompt.

This way you get a great overall image, making some controlled material edits to get to a beautiful final result to your specification and sticking with your design.

“How do I make images look more real and less like sketches?”

Reference images and prompting are key.

Your reference image is ideally something which closely aligns with what you’re hoping to generate - this should include interior/exterior, lighting, materiality, architectural style as well as the image aesthetic (photographic, cartoon, watercolour…). Once you’ve got a good reference image it’s then about prompting. If you’re hoping for something photoreal then there are some good keywords to include, like architecture photography or HDR. It can also be helpful to reference film or artists with similar styles.

If you’re still getting sketchy results it may be that you need to reduce the composition slider, to give Gendo more room to deviate from your input, this is often the case with Sketchup inputs which have ‘edge’s and ‘profiles’ turned on in the style panel.

“How do I make Gendo respect specific patterns, like wood floors or cladding orientation?”

Currently the best thing to do is upload a reference image which has the correct pattern in, as well as the material itself - such as brick bond patterns. Then prompt for the matching pattern as well, such as ‘red brick in stretcher bond.’

For typical patterns like that, or vertical/horizontal orientation you should be able to control as described. However, we’re working on further control to specify material orientation with a greater degree of control, like rotating a tile pattern on the floor to a specific orientation - but this is an upcoming feature.

“Can Gendo generate photorealistic renderings?”

Absolutely, while other styles are possible we assume photoreal as a baseline. Reference and prompting are key for this.

“How do I improve my renders overall?”

Keep inputs clean, instructions clear, and don’t try to change everything at once. Work in cycles: generate, review, adjust one parameter, generate again. Save prompts that give you the look you like.

Reference images and prompting are key.

Lastly you could try the Enhance tool, which is made just to add subtle detail and realism to an image without changing the essential design.

Workflow & Integration

“At what stage of design should I implement Gendo?”

Anywhere. I’ve seen people use it on napkin sketches, massing studies, near-finished renders, and even photos of existing buildings. Early in the process, it helps with ideas, bringing along stakeholders and alignment with clients. Midway, it’s about materials and style. Later on, it’s a presentation polish and winning over planning with a compelling and representative image.

The settings you should use will vary with the kind of input you supply. E.g. if you have a napkin sketch or low fidelty massing model, you’ll want to give the AI more creative room by reducing the composition slider - letting it fill in more gaps.

“Can I bring in 2D floor plans? Can Gendo extrude them into 3D?”

Gendo assumes that you’re uploading a perspective view of your design as you expect to see it. So we would recommend making a 3d model or hand sketch of your design based on the floor plan and upload that.

However we are working on some new tools which will help make textured/rendered floor plans and are looking into how we could get accurate views based on that! What it also won’t do (yet) is generate editable 3D models from plans. Gendo is an image tool, not a BIM tool.

“When will Gendo integrate with Revit, SketchUp, or Blender?”

Right now, the best workflow is to export scenes or views and bring them into Gendo. Direct plugins are on our roadmap, but they take time we want to get them right.

“How do I maintain consistency across different scenes in one project?”

Consistency comes from reusing the same “style language.” Many practices I’ve spoken to create a short style prompt: three or four lines describing materials, light, and atmosphere - and they also will have key reference images used across generations. We also see users using a previously generated image as the reference image to get consistency that way.

“How do I add entourage (trees, people, furniture)? Do I need to model them first?”

Excitingly we’ve just released a couple of new features to help with people - Quick Populate and Custom Cutouts.

Quick Populate will generate a group of people into your scene, matching the existing perspective and lighting to your high level specification of density, age and ethnicity.

Custom Cutouts is for when you need a very specific person for your scene. For example, you could use it to generate a chef in chefs whites and hat, with the right pose, rotation to the camera, ethnicity and age.

We’ll be adding in more tools for landscaping and furniture, as well as sketch edits soon! Otherwise you could take the image out of Gendo and quickly draw something like a tree over it and re-upload to Gendo and carry on from there.

“How does Gendo fit into urban planning workflows?”

Speed is everything. Being able to show multiple massing or streetscape options quickly makes collaboration with engineers, city officials, or community groups much easier. People respond to images, not CAD drawings.

We have users who are getting quick images out of Gendo to align with stakeholders like planners, colleagues or public consultations - so there are loads of ways to leverage this into Urban Planning. You just need a vision you’re trying to communicate and 5 minutes!

“Can Gendo handle CGI montage into video or AR?”

Not yet. For now, we’re focusing on still imagery - but this will be coming soon!

Vision & Differentiation

“How does Gendo compare to MidJourney, DALL·E, or ChatGPT image tools?”

Those are general image generators. They’re great for moodboards, but they don’t understand your design. Gendo is trained and tuned for architecture and crucially will work hard to preserve your intent, respects scale, and applies materials intelligently. If MidJourney is for inspiration, Gendo is for your actual project.

“How much creative control do I keep when using Gendo?”

All of it. Think of Gendo as speeding up your hands, not replacing your head. You set the geometry, intent, and style. Gendo just accelerates the image-making part. You can then set off on the usual iterative cycle of edits, tweaks and options using a combination of the functions on the platform. All with out the days and $$$ of a traditional visualisation process.

“How do clients respond to the ability to quickly change graphics?”

They love it. Sometimes it does mean they ask for more variations, but that’s easier to manage when you can generate them instantly. It shifts the tone from “defending a finished render” to “co-exploring design options.” We’ll be adding some collaborative features soon, so you’ll be able to sit down with your client and try out ideas and arrive at clear alignment faster than ever possible before.

“What’s next for Gendo?”

We’re focused on three things: consistency (better control across projects), integration (plugins into your modeling tools), and collaboration (team and enterprise features). Longer-term, we’ll move into richer media like text, video and 3D.

“What’s your stance on AI’s environmental impact?”

We take it seriously. Running AI in the cloud uses energy and water, and we’re actively working on efficiency improvements. This is something the entire industry needs to tackle. Catch this blog to read more:

“What’s your favourite colour?”

Gendo black. Aaha no, I think I’m obliged to say green - just based on my current wardrobe!

Business & Pricing

“Is Gendo worth the cost?”

If visuals help to drive your decisions, yes. The ability to turn “what if we tried it this way?” into an image in minutes can save hours of manual rework. Most practices tell us the ROI is clear after a single client meeting. We’ve also had clients tell us that a Gendo image has cinched a planning application: "We've been stuck for six to eight months with the planning department at the preapplication stage, making slow hesitant progress. Client took a rendered image from Gendo from one of our hand sketches to the meeting, the planners took one look, Boom. Planners on board. Off to planning, we go." Dan, Origin Design Studio

“How much does Gendo cost? Are there student or faculty discounts?”

We offer a free plan, paid plans, and discounted access for students and faculty. Larger studios can opt into enterprise pricing. Students get an educational licence which gives them lots of monthly credits for free! Just sign up with your .ac or .edu email and you’ll be automatically assigned!

“Is there an unlimited organizational subscription?”

Yes, with our Enterprise Plans we can arrange as much as you would like.

Beyond Gendo: Questions About AI in Architecture

One of the most common threads in every Q&A isn’t about Gendo specifically, but about AI in architecture as a whole. People want to know: What tools matter? What should I be paying attention to? Where does AI actually fit into a studio workflow, and where is it just hype?

Here’s how I usually answer:

“What are the best AI tools for architecture?”

There isn’t a single “best” tool. It depends on the job. Broadly, I see four categories:

Visualization AI – Tools like Gendo, MidJourney, or Stable Diffusion. These help turn sketches, renders, and ideas into compelling visuals. The difference is whether they respect your geometry (Gendo) or just generate moodboards (MidJourney).
Productivity AI – Large language models like ChatGPT or Claude that can help with writing project descriptions, bids, or even preparing presentations. Quiet but powerful.
Analysis AI – Early-stage tools exploring adjacency studies, zoning, and environmental optimization. These are experimental but worth watching.
Specialist plugins – Integrations inside Revit, Rhino, or Grasshopper that take repetitive drafting or modeling tasks off your plate.
Optimisation - There are also some great tools like Finch and Rayon to help with floor plan layout optimisations!

“Should I be using MidJourney or DALL·E as an architect?”

They’re fantastic for moodboards and inspiration. But if you need accuracy tied to your geometry, you’ll quickly find the limits. That’s where architecture-focused tools Gendo step in.

“Will AI replace architects?”

No. AI replaces repetitive workflows, not creative judgment. The architect’s role is still to define the problem, weigh trade-offs, and set intent. AI accelerates the way you test and communicate ideas.

“Where should I start with AI in my practice?”

Start small. Choose one friction point maybe generating concept boards or polishing quick renders and test AI there. Measure the hours saved. If it works, expand. Gendo shines in early stage design work where you need to bring a client, colleague or other stake holder along with your vision. You can take a quick sketchup model through to a photoreal image in minutes and really sell your idea.

“How much should I trust AI in client-facing work?”

Use it as a conversation starter, not as gospel. AI is brilliant for opening options, but it’s your responsibility to filter, select, and present what aligns with the project. It’s also up to you to make it clear what it is you’re presenting. If you want sign off on the overall massing, maybe photo real isn’t the way to go, you don’t want them getting hung up on the curtains. 🤓 Which is what out style tool is for, keep the essential information in the image there, the design, materiality and lighting, but adjust the aesthetic to a watercolour, to make it clear to the viewer that this is not a final design but a concept.

“Is AI environmentally sustainable?”

This question comes up a lot. Running AI models does use energy and water. The industry — ourselves included — is pushing hard on efficiency, but it’s important to be honest: this is an area where architecture must lead by demanding sustainable solutions. We have as much of our hardware on renewable energy supplies as we can, but there is more to be done.

“What skills should young architects focus on as AI becomes more common?”

Two things: clarity of intent (knowing what you want and how to describe it), and integration skills (knowing how to fit AI outputs back into real workflows). AI won’t reward button-pushing — it will reward designers who can direct it. It will undoubtably be those who have embraced the tech early who will be the biggest benefactors, either by carrying those skills into practices, or by starting your own and punching well above your weight with AI-enhanced processes.

“Will every architecture firm need an ‘AI strategy’?”

Yes, in the same way every firm eventually needed a BIM strategy. It doesn’t mean you’ll be automating everything, but it does mean deciding where AI accelerates you, and where you stick to manual craft.

Closing

Every question you asked — from the serious to the playful — helps shape what we build. Gendo is still young, and it’s growing fast, but it grows in response to you. Keep challenging us, keep sending your sketches, keep asking the difficult questions.

Until the next time,

George