At the time I’m writing this \u2014 May 2026 \u2014 most Singapore ID firm owners I talk to are quietly worried. Not panicking, not closing down, but worried in that low-grade, persistent way that makes you check your QuickBooks dashboard at 11pm for no real reason. Revenue is flat. Senior headcount costs more than it used to. And somewhere out there, AI is doing things that six months ago required a junior designer and two days of work.<\/p>\n
I want to make a specific prediction here, because I think the polite version of this conversation is doing more harm than good: by the end of 2028, a meaningful share of Singapore ID firms that have not restructured around AI will be gone. Not struggling. Gone \u2014 closed, merged into a larger competitor, or quietly wound down when the founding director decides the math no longer works. My estimate is somewhere between 35% and 50% of firms operating today. If I’m wrong, you’ll know by mid-2028 when the Singapore Business Federation’s SME Committee publishes its next design sector survey and the attrition numbers tell a different story.<\/p>\n
I’m not writing this to scare anyone. I’m writing it because I think the window to do something is shorter than most firm owners realise \u2014 and because I’ve had enough uncomfortable conversations in the last six months to feel like someone should say this directly.<\/p>\n
Here’s the math as I understand it in May 2026, drawn from conversations with around thirty Singapore ID firm owners over the past year. A senior designer in Singapore \u2014 someone with 5-7 years of experience who can run a project from brief to handover \u2014 costs a firm between SGD $5,500 and $6,500 a month fully loaded. That includes CPF, AWS, software licenses (SketchUp, Revit, AutoCAD, V-Ray rendering subscriptions), and whatever informal benefits keep them from going to the firm down the road in Bukit Timah.<\/p>\n
Junior designers \u2014 the people who used to do moodboards, material research, basic 3D modelling \u2014 cost SGD $3,200 to $4,200 a month fully loaded. And this is where the structural problem starts. AI tools like Midjourney, Stable Diffusion, and now more specialised platforms like Maket.ai and ARCHITEChTURES are doing the junior-tier work faster, more cheaply, and without a two-week learning curve. According to a Q1 2026 survey by CNA, 67% of Singapore design professionals report that AI is already handling tasks they previously spent at least 40% of their junior hours on.<\/p>\n
So what does an ID firm do? The junior tier is getting squeezed out. The senior tier is still irreplaceable for client-facing work, site coordination, and design judgment. But the revenue model that funded itself on a pyramid \u2014 two or three seniors supported by a wide base of juniors \u2014 is collapsing. Firms that haven’t restructured are now running with the cost profile of a senior-heavy team and the revenue pipeline of a firm that expected cheaper junior leverage.<\/p>\n
That’s a margin squeeze. And it’s not temporary.<\/p>\n
I hear this from firm owners regularly. “We’re watching the space.” “We’ll adopt when it’s more mature.” “Our clients expect the human touch \u2014 AI doesn’t fit our brand.” I understand the instinct. Genuinely. And there’s a version of it that’s reasonable \u2014 you shouldn’t rush AI adoption without thinking through quality control, client expectations, and workflow disruption.<\/p>\n
But let me put it differently: the firms that are winning right now aren’t adopting AI slowly. They’re already 12-18 months into a restructured workflow where AI handles the first-draft moodboarding, the material option generation, the initial space planning iterations \u2014 and their senior designers are spending their hours on the 20% of the work that AI can’t do yet. Those firms are effectively running with 60-70% of the junior headcount they had in 2023, at better margin, with faster project turnaround.<\/p>\n
A Tiong Bahru-based ID firm we’ve had indirect visibility into (not naming them, they didn’t ask us to) went through this restructuring in late 2024. They cut their junior headcount from six to three over eight months \u2014 not through redundancies, but through natural attrition they stopped backfilling. By March 2025, their project delivery time on a standard HDB renovation brief had dropped by about 22%. Revenue stayed flat. But margin went from approximately 14% to 21%. That’s not a small number for a firm their size.<\/p>\n
The firms “watching the space” in 2026 are 18 months behind those firms. In an industry where margins are thin and senior talent is scarce, 18 months is a long time to be behind.<\/p>\n
Here’s something I want to say carefully, because it could be misread as pessimism: the HDB Minimum Occupation Period (MOP) wave of 2026-2027 is both a real demand tailwind and a dangerous distraction.<\/p>\n
HDB data from 2024 showed approximately 38,400 HDB flats reaching MOP in 2026 alone \u2014 significantly more than the 2024 average of around 22,000 per year. This is a genuine demand surge for residential ID firms. The phones are ringing. The inquiry volumes are up. Firm owners who should be worried about their cost structure are instead enjoying a temporarily full pipeline and telling themselves the business is fine.<\/p>\n
It’s not fine. The MOP wave peaks in 2027, then tapers. And when it tapers, firms that have spent the boom years not restructuring will face both a demand slowdown and a cost structure that was only sustainable when inquiry volume was artificially high. The AI-restructured firms will be able to take on that lower volume profitably. The non-restructured firms won’t.<\/p>\n
Aiyo. It’s the kind of problem that’s very obvious in retrospect. Right now, it’s mostly invisible because the revenue is coming in.<\/p>\n
When I ask firm owners why they haven’t restructured, the answers cluster into three categories. Worth naming them directly, because I recognise all three from conversations in the last few months.<\/p>\n
The first is client perception anxiety.<\/strong> “Our clients are paying for human creativity \u2014 if we tell them we used AI for the moodboard, they’ll feel cheated.” I have some sympathy for this. But I’d push back: most clients don’t care how the moodboard was generated. They care whether it reflects their taste, their budget, and their space. Firms are conflating the tool with the output quality. A moodboard generated in twelve minutes by Midjourney, then curated and contextualised by a senior designer who knows the client’s Bishan flat and their specific aesthetic preferences, is more valuable than a generic moodboard produced by a junior spending two days on Pinterest.<\/p>\n The second is senior designer resistance.<\/strong> Some senior designers genuinely feel AI threatens their professional identity. I get it \u2014 these are people who spent years developing skills in rendering, in material specification, in space planning software. Being told that a machine can now do parts of their job is a real psychological blow. But firms that let this internal resistance dictate technology strategy are handing the decision to the wrong person. It’s not different from a bakery letting its head baker decide whether to buy an industrial mixer because he’s proud of his hand-kneading technique. The customers don’t care. The margin does.<\/p>\n The third \u2014 and this is the one I find hardest \u2014 is founder inertia.<\/strong> The firm owner who’s been running the business for fifteen years, who built it on a specific set of skills and relationships, who doesn’t fully understand the AI tools and doesn’t have time to learn them. This is the most human version of AI resistance, and honestly, it’s the most dangerous. Because it’s not irrational, it’s just slow. And slow, in 2026, is its own kind of risk.<\/p>\n I want to be honest about something: I don’t have a tidy “five steps to survival” answer. Anyone who’s selling you a clean framework for this is probably selling something. What I can say is that the firms I’ve seen navigate this well share three specific characteristics.<\/p>\n They’ve restructured their workflow so that AI handles the first 40-60% of the creative generation work \u2014 moodboards, material options, initial renders, space planning iterations. They’ve redeployed their senior time toward the client-facing, judgment-intensive work that AI genuinely can’t replicate yet. And they’ve replaced the junior headcount they shed with either AI-augmented Filipino remote designers (who cost SGD $1,050 to $1,350 a month all-in through arrangements like Kaizenaire’s offshore staffing service<\/a>) or with AI tooling subscriptions that cost a fraction of a junior hire.<\/p>\n That third point \u2014 the offshore augmentation model \u2014 is obviously something I have a commercial interest in. I want to name that clearly. But it’s also genuinely what the math supports for firms that need design capacity beyond what AI alone can provide, without going back to the full cost of a Singapore junior hire. The difference between SGD $3,500 a month for a Singapore junior and SGD $1,200 a month for an AI-augmented Filipino remote designer with equivalent capability at routine tasks is not a rounding error over twelve months.<\/p>\n What doesn’t work is doing nothing and hoping the MOP wave carries you through. It won’t carry you far enough.<\/p>\n Before I go further \u2014 if you want to understand how Kaizenaire actually operates before deciding whether this is relevant to your firm, check out our bad reviews (PS: this is not a typo)<\/a>. It’s the most honest page on our site about what we do, who we work with, and why some people have had a poor experience with us. I’d rather you read that first than make a decision based on marketing copy.<\/p>\n I’ve been making predictions about the Singapore ID sector for about three years now. Charlotte will tell you I get roughly 60% of them directionally right, which is better than a coin flip but not good enough to be smug about. I’ve been wrong about timing more than I’ve been wrong about direction \u2014 things tend to happen slower than I expect, then faster all at once.<\/p>\n My reading of the current signals is that 2027 is the inflection year. The HDB MOP demand starts to taper. The AI tools that are already capable in mid-2026 will be measurably more capable by late 2026, making the productivity gap between AI-restructured and non-restructured firms wider. And the senior designer talent pool \u2014 already tight \u2014 will become even more contested as restructured firms, needing fewer people per project, can afford to pay more for the good ones.<\/p>\n The firms that close or merge won’t do it dramatically. They’ll just quietly stop taking new projects. The founder will say he’s “taking a break” or “stepping back to consult.” The team will scatter. The portfolio will go offline. In the Straits Times and Business Times, it’ll be invisible \u2014 these aren’t firms that generate news. But in the design industry group chats, people will notice.<\/p>\n If my 35-50% closure estimate is wrong on the high side, I’d expect the SBF design sector attrition data by mid-2028 to show something closer to 20-30%. That would still represent a significant structural contraction. The directional call \u2014 that non-restructuring firms face existential risk by 2028 \u2014 is what I’m confident about. The exact percentage is the part I hold lightly.<\/p>\n I genuinely don’t know if your specific firm is in the at-risk group. That depends on your current cost structure, your pipeline quality, your senior team’s adaptability, and honestly whether you’re the kind of founder who can move fast when the situation requires it. Some aren’t, and that’s not a moral judgment \u2014 it’s just a capability profile. Boh pian, as they say. But knowing which one you are matters.<\/p>\n 90 days is specific deliberately. Not “start thinking about AI.” Not “build a strategy.” 90 days, with a decision at the end.<\/p>\n In the first 30 days: map your junior workload. Specifically, account for every task your junior designers do in a typical project week. Not the categories \u2014 the actual tasks. Moodboard iteration. Material research. 3D base modelling. Supplier coordination emails. Site visit scheduling. Then ask, honestly, which of those tasks an AI tool or an AI-augmented remote designer could handle at equivalent quality. My guess is it’s more than 50%.<\/p>\n In days 31-60: run one AI tool on one live project. Not a pilot \u2014 an actual project, with an actual client, with actual stakes. The only way to know whether AI moodboarding or AI-assisted renders work for your firm’s aesthetic standards is to test them under real conditions. This is the step most firm owners skip. They read about AI, they watch YouTube demos, they go to one industry talk, and then they go back to their existing workflow. That’s not learning. That’s postponing.<\/p>\n In days 61-90: make the headcount decision. Based on what you learned in the first two phases, decide whether your next junior vacancy \u2014 when someone leaves \u2014 gets backfilled with a Singapore hire, an offshore AI-augmented hire, or no hire because the AI and your existing team can absorb the work. That decision, made with data from your own firm rather than in the abstract, is worth more than any framework I can give you.<\/p>\n And if you want to talk through the offshore piece specifically \u2014 what an AI-augmented Filipino remote designer can and can’t do for a Singapore ID firm, what the realistic cost looks like, and whether we’re even the right fit for your situation \u2014 reach out to Kaizenaire at our WhatsApp Business Number +65 9636 2204<\/a>. Our team will be ready to serve you. No aggressive follow-up. No pitch deck. Just a straight conversation about whether the math works for your firm.<\/p>\nThe Survival Path Is Specific and Narrow<\/h2>\n
My Honest Assessment of the Timeline<\/h2>\n
What to Do in the Next 90 Days<\/h2>\n