Your AI Isn’t the Problem. You Hired the Wrong One.
Most interior designers use AI to write emails and captions and to make the odd room image. Let’s call this Level 1 AI: a virtual assistant who’s fast, capable and knows nothing about your business. It’s fine for Level 1 jobs. The trouble starts when you hand that Level 1 AI a Level 2 or Level 3 job.
I want to introduce you to the three levels of AI for interior designers, explained as three different hires, what each one is good for, and the three things every level needs from you before it gives you results worth using.
Are interior designers who only use chatbots and image generators falling behind?
No. Using ChatGPT, Claude or Gemini for emails, captions, summaries and concept images puts you in good company, because it’s where most designers are. Level 1 AI is fine for Level 1 tasks. The problem shows up when you ask a virtual assistant to do a specialist’s or an operations manager’s job, like building your marketing strategy or sorting out your purchasing. You aren’t bad at AI. You’re asking the wrong kind of AI to do the job.
Apple - Youtube Music - Spotify - Youtube
What are the three levels of AI?
Think of AI as a new hire. Level 1 is a virtual assistant: good at everyday tasks like emails, captions, summaries and research, and it knows nothing about you unless you tell it. Level 2 is an expert hired for one role, like renderings, marketing or bookkeeping. It knows its specialty well and still knows nothing about you. Level 3 is an operations manager who’s worked with you for years and knows how you run your studio.
Why does AI give me generic answers to specific questions?
Because Level 1 AI knows nothing about you or your business. Chat apps like ChatGPT, Claude and Gemini, the image generators and the AI buttons inside Canva, Gmail, Outlook and Instagram all run on models trained on a huge amount of public text and images. Add whatever you type or upload, plus some memory, and you have everything it works with. When there’s a gap, it fills it with something which sounds right. It guesses at the answer it thinks will make you happy.
What does every level of AI need from you?
Three things: specification, intent and context. Specification is what you want: the thing to make and what finished looks like, such as a 100-word Instagram caption or three layout options for a 12 by 14 foot bedroom. Intent is why you want it and who it’s for, such as attracting homeowners planning a kitchen renovation, not people looking for decorating tips. Context is who you are: your clients, the project, your style, your way of working and what you’d never do.
You already work this way with clients. You don’t start a project by designing; you start by asking questions. AI loves giving answers and isn’t great at asking questions, so the burden falls on you. Context is the hardest of the three, because almost all of it lives in your head.
Why don’t AI room images look realistic?
Because image generators and other Level 2 tools know their specialty, not your project. Rendering, mood board and floor plan apps produce polished rooms, and designers keep raising two complaints: the rooms aren’t realistic, and clients expect you to build them anyway, physics be damned. Give the tool specification (camera angles, materials, products and what must stay the same between images), intent (what the image should help the client decide) and context (real room photos, measurements, what’s staying, the budget range and the client’s taste).
What would an AI operations manager do for you?
Level 3 AI connects to the software you run your studio on: design studio software such as Studio Designer, Mydoma or Programa, project management tools like Asana, Trello or Notion, a CRM, QuickBooks, email, calendar and file storage. Like an operations manager who’s been with you for years, it uses all of it to spot patterns you’re too busy to see. Where do your profits come from? Which types of projects lose money? Which vendors are always late? Which clients send you referrals?
As of November 2026, this kind of analysis is still rare in design software, and the answers are only as good as the data the AI has to work with.
Should you let AI act on its own?
Not yet. Level 3 AI works in three ways: it answers questions (when is the sofa for the Smith project arriving?), drafts content for your approval (an email to Ms. Smith about her sofa), or acts on its own (sends the email, schedules the install and updates everyone involved). The best AI models still come with a disclaimer: AI makes mistakes. You wouldn’t let a new junior designer loose with the company credit card, and you’re the one accountable for whatever your AI does. Keep a human in the loop.
What is Level 4 AI?
Level 4 AI doesn’t exist yet. If Level 3 is an operations manager who’s been with you for years, Level 4 is going back to day one and building your business with AI at its core, using everything you know today. Wharton professor Ethan Mollick argues the easy path is using AI to do today’s tasks faster, bolted onto the systems you already have, and the better path is rethinking how the work gets done. If you plan to be in business ten years from now, Level 4 is where you want to be.
Nobody has a proven template. Among the designers, coaches and experts I talk with, nobody has rebuilt a business around AI yet, including me, and the forecasts for the AI of 2028 disagree wildly.
Which level of AI should you be using?
All three, each for its own job. The trick is matching the level of AI to the level of the task: “Level one AI for level one tasks, level two AI for level two tasks, and level three AI for level three tasks.” Use a chat app for emails and captions, a specialist tool for renderings or bookkeeping, and connected software for running projects. Whichever level you use, the safest bet is the same: teach your AI the specification, intent and context which make your clients love you.
-
Most interior designers use AI to make their business go faster. Draft emails, social media captions, design concept images, and all that stuff is level one AI, and level one AI is fine for level one tasks. But when you ask level one AI to do a level 2 or 3 task, that's when we run into problems. If you're like 99.9% of AI users, this idea of different levels of AI is going to be brand new to you.
The truth is, I'm a giant AI nerd, and it was new to me up until a few months ago. The best explanation I can come up for it is for us to think of AI like a new hire in your business. Level 1 AI is like a virtual assistant, competent at everyday tasks, emails, captions, summaries, research, but knows nothing about you and your business unless you tell it.
Level 2 AI is an expert at a specific role or task. They still know nothing about you, but they're really good at their specialty, and that's why you've hired them. This could be image renderings, marketing tasks, bookkeeping, etc. This level of AI isn't a generalist like level one. They're a specialist. Level 3 AI is that business partner or operations manager who's worked with you for years. It knows your business.
It's great at helping you keep your projects on track, finding mistakes, and recognizing opportunities for improvement. Level 3 AI could do amazing things for your business. But that doesn't mean that level one AI sucks. The trick here is to use the right level of AI solution for the right task. Level one AI for level one tasks, level two AI for level two tasks, and level three AI for level three tasks.
All right, let's dive into this. Level one is your virtual assistant, AI that's capable and fast but knows nothing about you or your business. We're talking AI chat apps, ChatGPT, Claude, Gemini, AI image generators, and AI that's built inside of apps you already use: Canva, Gmail, Outlook, Instagram. All of these level one AIs are built on top of AI models that have been trained on a huge amount of public text and images.
This training, plus whatever you give it in the chat, words, files, or photos, and some memory. Is what it has to work with. That's it. When there's a gap, it fills in with something that sounds right. In essence, it guesses at the answer that it thinks will make you happy. Now, there is a way to improve the output of any level one AI, and that's by giving it the following three things: specification, intent, and context.
In real life, when you meet a new client, in person. You don't begin by diving into the design process. You start by asking questions to better understand what they want, why they want it, and all of the personal stuff that will make your designs perfect for them. AI is exactly the same. To do a good job for you, AI needs to understand what you want, why you want it, and most importantly, who you are.
Specification is the what you want piece. The thing to make and what finished looks like. Do you want a 100-word Instagram caption or do you want three layout options for a 12 by 14 foot bedroom? Intent: why you want it, what's it's for, and who it's for. To attract homeowners planning a kitchen renovation, not people looking for decorating tips. Context: who you are, about you, your clients, and the project, your style, your way of working, the stuff you do, and the stuff you'd never do.
Of all these three things, context is the most complex and difficult to pin down, because almost all of it lives in your head. And as you know, while AI loves to give answers, it's not so great at asking questions. So the burden for this information falls on you. If you give your level one AI these three things: specification, intent, and context, it will give you better level one results.
Promise. Level 2 is your expert. Level 2 AI has been built or trained for a specific job. It knows its specialty well, but just like level 1 AI, it doesn't know you or your business. We're talking more specific design related AI image generation, renders, mood boards, floor plan apps, AI tools for. A specific business task, like social media marketing, website writing, or bookkeeping. The AI inside QuickBooks, for example, a ChatGPT or Claude project that you load with your own files so it can find past specs, pricing, and what a client agreed to.
Each of these tools does a specific thing, and they do it well. But as with level one AI, if you don't tell them exactly what you want, specification, why you need it. Intent and what they need to know about the project context, the results are not going to make you and your business shine. Let's look specifically at AI image generation. When I speak with designers, I keep hearing two main complaints about AI generated interiors.
One, they're not realistic. Two, clients expect you to take those unrealistic images and make them come to life. Physics be damned. And like level one. Level 2 AI gets better when you give it these things. Specification: we're talking camera angles, materials, products, and what must stay the same from one image to the next. Intent: what should this image help the client decide? A mood to react to, a layout that needs approval.
Context: the real room photographs, measurements, which furniture is staying, which is leaving, budget range, client's taste, all of that stuff. AI image generators and all other level 2 AI tools, when armed with specification, intent, and context, will give you better results. All right, let's move on to level 3. Level 3 AI, like that operations manager or business partner who understands your business on a deep level. Picture an AI that's already connected to the software that you run your studio on.
We're talking design studio software that runs projects, purchasing, and invoicing, such as Studio Designer, Mydoma, Programa. We're talking project management tools for tasks and timelines, Asana, Trello, Notion, a CRM that tracks leads, clients, and follow-ups, accounting software like QuickBooks, also email, calendar, and file storage. Your studio software holds years of data information on every project, purchase order, invoice, vendor, delivery date, and hour billed. An operations manager who's been with you for years uses all of that to spot patterns that maybe you're too busy to see.
In theory, AI can do the same thing, do it 24 hours a day for less money. And in theory, you should be able to ask all of those level 3 AIs questions like, "Where do my profits actually come from? What types of projects lose money? Which vendors are always late? How long does purchasing really take from this vendor?" Which clients send me referrals? Unfortunately, as of today, this kind of analysis is still pretty rare in design software, but it should be coming in the near future.
But just like levels one and two, the answers you get are only as good as the data that AI has to work with. What will make level 3 AI work for you? Is the good technical data that it already has access to, plus this focus on specification, intent, and context that I keep hammering you with. AI at this level can do work in three very different ways.
It can answer questions. When is the sofa for the Smith project arriving? It can draft content for your approval. Draft an email to Ms. Smith about her sofa, then you review it and send the email. Number three, it can act on its own. Send that email to Ms. Smith, schedule the install date, and update everyone involved. This last one is the promise of AI agents that can essentially run your business for you.
And to be honest, as someone who works with AI every day, this still kind of freaks me out. The best AI models in the world still come with a disclaimer: AI makes mistakes, which is fine if you keep a human in the loop. But just like you wouldn't let your new junior designer loose with the company credit card, as of right now, I would still be wary of autonomous AI agents at this level three AI level.
At the end of the day, you're the one accountable for whatever your AI does. Okay, so that's your three levels of AI lesson for today. If you've made it to the end of the video, you're ahead of almost everybody else when it comes to understanding AI in your business. But what if I told you that there could be or should be a fourth level of AI? Level 4 AI doesn't exist today, except in the minds of people like Wharton Professor Ethan Mollick.
If level 3 AI is that operations manager who's been with you for years, level 4 AI is like going back in time and having you and that operations manager applying everything you know about your business today, but right from day one, with AI built in at the core of your business. Professor Mollick is adamant about how businesses of every size should be using AI and that we're all doing it wrong.
Right now, even the most nerdy business owners are using AI to improve their business speed and efficiency by bolting AI on top of their existing systems, which is fine. But according to Mollick, the better, smarter, and to be fair, more difficult path is rethinking how your entire business works, and then committing to put AI at the center of it all. Imagine if the business side of your business was run almost exclusively, instantly, and perfectly by tomorrow's AI.
What if you could focus on designing interiors and interacting with clients and prospective clients? What would you do differently? What would this business look like? That's level 4 AI. And if you're planning on being in business 10 years from now, level 4 AI is probably where you want to be. But the truth is, I spend a lot of time talking about AI with designers, with coaches, with experts, and I haven't yet found anyone who's rebuilding their business around AI, myself included.
Nobody has a proven template; it's just too new. And the target keeps moving. Are you setting up your business for today's AI or the AI of 2028? Nobody knows for sure what that's going to look like, and the forecast, to be quite honest, disagree wildly. The safest bet for me is to assume that AI is eventually going to be able to take over more and more of the business side of your business.
Your job is to provide that AI with your specification, intent, and context. Teach your AI all of the stuff about you that makes your clients love you. All right, that's it for today's podcast. I know it was a lot. As always, if you have any questions, hit me up on Instagram or email me. If I can answer your question right away, I'll just shoot you a message.
If you want help integrating AI into your business, I work one-on-one with business owners. And if you want to learn more about AI, I have created a guide to this three levels of AI. It's a free PDF download. I will have all of the links in the description below. All right, that's it for today. See you next time.
Key Takeaways
You’re not bad at AI. Most designers use Level 1 AI for emails, captions and concept images, and it’s fine for those jobs.
Think of AI as a new hire. A virtual assistant, an expert hired for one job, and an operations manager who knows your studio.
Match the level to the task. Asking a chat app for your marketing strategy is asking a virtual assistant to do a manager’s job.
Every level needs three things from you. Specification (what you want), intent (why you want it) and context (who you are).
Keep a human in the loop. You wouldn’t hand a new junior designer the company credit card, and you’re accountable for whatever your AI does.
Build for what’s coming. Write down how you work and what your clients love about you. It pays off at every level.
Conclusions
Strategic Benefits: Matching the level of AI to the task stops you blaming the tool for a bad hire. Chat apps handle everyday writing, specialist tools handle renderings and bookkeeping, and connected software, as it arrives, answers questions about your own projects, vendors and profits.
Implementation Blueprint: Pick one task you already use AI for and give it all three things: specification, intent and context. Compare the result with what you usually get. Then list your biggest business headaches and note which level of AI each one needs.
Professional Transformation: The goal isn’t using more AI. It’s spending more of your week designing and talking with clients while the right hire handles the business side. Teach your AI what makes your clients love you, and it’ll be ready for whatever comes next.
Join the Interior DesignHer Community
If you enjoyed this guide and podcast, you're going to love all the other business education resources we're bringing to interior designers.
As interior designers ourselves, we created Interior DesignHer to bring top-notch business education to interior designers. We especially have a soft-spot for interior design solopreneurs & micro-businesses.
Subscribe now so you never miss our latest business guidance, guest interviews, and insider strategies specifically for interior designers.
