A prospect walked into our discovery call last quarter with a very clear brief. He wanted AI. Specifically, he wanted a chatbot on his website, a dashboard with some graphs, and — his words — "something we can show investors." His 28-person trading firm had real operational chaos underneath: sales reps managing pipelines in WhatsApp threads, finance chasing payment confirmations over email, and an ops team doing manual reconciliation every Monday morning. But none of that was on his agenda. The deck was. We turned the project down.
What "AI as Theatre" Actually Looks Like
It's not always as obvious as the example above. Sometimes it's dressed up in the right vocabulary. "We want to automate our workflows." "We're looking to leverage AI for efficiency." Sounds good. Except when you start pulling on the thread, you find that nobody in the business has actually mapped what those workflows are. There's no documentation. The team leads haven't been consulted. And the CEO wants to go live in six weeks because there's a board meeting.
We've started calling this AI theatre — implementations designed to be seen, not used. A chatbot that answers five FAQs and calls it transformation. A dashboard that visualises data nobody acts on. An automation that handles 3% of the actual workload because the other 97% was too messy to map. The vendor gets paid. The slide looks great. The ops team is still on Excel the following Monday.
There are specific signals we listen for in a discovery call. The first is when a client can't name the person responsible for the process they want to automate. If nobody owns it, nobody will adopt the new system — and the project dies in implementation. The second is when the timeline is driven by optics, not operations. "We need this done before our annual dinner" is not a project requirement. The third — and this one is almost always fatal — is when the client says "we don't want to change how we work, we just want AI on top." That's not automation. That's decoration.
We also watch for businesses where the leadership wants the result but hasn't bought the team in. As we've written before, the failure mode for most AI projects isn't the technology — it's the internal resistance that nobody planned for. If the operations manager finds out about the project the same week we're supposed to go live, we already know how this ends.
Why We Say No — And What It Costs Us
Turning down projects is not a comfortable business decision. We're an SME ourselves. Every scoped project that doesn't convert is real revenue we didn't take. But we've learned — the hard way, honestly — that the wrong client doesn't just cost money. It costs credibility, team morale, and the kind of case studies that actually build a business.
When you implement AI theatre for a client who wanted theatre, you get a testimonial that says "great team, very professional." What you don't get is a result you can point to. You don't get the 12-person logistics firm in Tuas that cut its weekly reporting from two days to three hours. You don't get the 35-person distribution company whose sales team stopped losing leads because follow-ups were finally automated. Those outcomes — the ones that make the same team produce bigger output — only happen when the client is genuinely committed to changing how the business operates, not just how it looks.
There's also an ethical dimension that's easy to gloss over. When we take a project, we're asking a business to invest time, money, and internal political capital. We're asking operations staff to learn new systems and change old habits. If we know going in that the brief isn't serious, taking the money anyway is a poor trade — for them and for us. Choosing the right AI partner cuts both ways. We're choosing our clients as much as they're choosing us.
The Client We Actually Want to Work With
Here's what a good discovery call sounds like. The business owner comes in frustrated, not excited. They're not pitching us on how ready they are — they're describing a specific operational pain that's costing them something real. "We're losing two days a week chasing purchase order confirmations and I need that time back." "My best sales rep spends four hours on Friday doing admin instead of selling." That's a problem we can build against.
The clients we say yes to can usually answer three questions clearly: What process breaks the most? Who owns that process? And what does success actually look like in six months — not in a deck, but in operations? When those answers exist, the rest of the engagement has a foundation. Real operational transformation starts with a real diagnosis, not a wishlist.
We work best with businesses that have been running for a while — often 10 to 30 years — where the founders built something genuinely valuable but the systems never scaled with the growth. The manual processes that worked at 8 staff are now strangling the business at 30. The team is capable. The operations are just held together with WhatsApp groups and institutional memory. That's not a broken business. That's a business that's ready. Your best people shouldn't be chasing invoices. And when a business owner says that back to us unprompted, we know we're in the right room.
We're also honest about grant funding. If your project qualifies under the Enterprise Development Grant, we'll tell you. If it doesn't, we'll tell you that too. We don't engineer projects around grant eligibility just to make the number look better. The operational outcome has to justify the investment with or without the subsidy.
If you're reading this and thinking "that first scenario sounds uncomfortably familiar" — that's worth sitting with. And if you're reading this thinking "no, I have a real problem I need fixed," then let's talk. Come with the messy version. That's the one we're actually good at solving.
