When people imagine the businesses leading artificial intelligence adoption, cleaning companies, tuition centres, photography studios, and other everyday service providers probably aren’t the first that come to mind.
Jingjing Zhong thinks they should be.
The former investment banker is now Co-Founder of Superbench, a Singapore-based company helping B2C service businesses use AI across sales and operations. Working with these companies has exposed Zhong to an often-overlooked part of the AI economy: businesses where customer relationships live inside messaging apps, workflows remain fragmented, and conventional technology implementation has often been too expensive or cumbersome.
For Zhong, AI could change that equation—not by giving SMEs more technology for technology’s sake, but by helping them improve the numbers that actually matter.
In this interview with NewInAsia, Zhong explains why smaller operators may have an unexpected advantage over large enterprises, where AI is already producing measurable business results, and why the service business of 2030 could look radically different from today’s.
Highlights
The Underdogs May Have an AI Advantage.
Q: Much of the AI conversation revolves around large enterprises and technology companies, yet Superbench focuses on service-based SMEs. Why do you believe these businesses represent one of the biggest untapped opportunities for AI adoption?
A: Depending on which region you look at, service SMEs make up anywhere from 15% to 20% of the economy on the low end, up to 50% in some markets. That scale alone should make this the biggest untapped opportunity in AI, yet it is barely part of the conversation.
These businesses have traditionally been left behind by technology, not because they didn’t want it, but because implementation has always been slow and costly. When you’re a small operator just trying to survive, you don’t have the time or capital to spend six months on a software rollout.
AI changes that equation completely. It gives these businesses additional brainpower, additional hands and legs, to execute faster without needing to hire or build a tech team from scratch.
That actually makes AI more of an underdog’s advantage than a big enterprise’s advantage.
Large enterprises will benefit from this wave too, but later, because they’re dealing with legacy tech stacks, and migrating or connecting that legacy infrastructure to AI takes real time. Service SMEs have no legacy to untangle.
They can move first.
Forget the CRM. The Business Is Running on WhatsApp.
Q: Service industries often deal with fragmented workflows, manual processes, and constant customer interactions. Which operational challenges have surprised you the most, and where have you seen AI deliver the greatest real-world impact?
A: The thing that surprised me most, honestly, is that B2C service businesses don’t run on a CRM. They run on WhatsApp or chat messengers, full stop.
Extracting existing context and workflows out of those conversations was genuinely hard in the early days. We had to do it manually, and back then AI simply wasn’t good enough to extract and map that context on its own.
It’s also not enough to just extract a knowledge base and plug it in. Someone has to clean it up, and that requires a subject matter expert to actually read through it and audit how the context is being managed.
This is where we have a real edge over other companies because we know this industry from operating inside it. Even the manual version of this work was faster and more accurate for us than it would be for an outsider. Now, with more advanced AI, we can do it far faster still.
The real-world impact shows up most clearly at the very first touchpoint with a customer. Our client Firefly Photography has seen sales increase by 20%, with 60% of that customer-facing work now handled by AI. We see similar numbers at Helpling. Both are high-volume B2C service businesses, and both show the same pattern: consistency at the first touchpoint moves the topline.
There’s also a second, less obvious impact. AI closes the speed gap, and humans still close the trust gap.
Our client Bike Hero named their AI agent Dianna, and Dianna has actually started showing up by name in their five-star Google reviews.
Do customers know Dianna is AI? Honestly, I don’t know. But she’s showing up in the reviews all the same.
What If Knowing More About AI Makes You Slower?
Q: Many business owners still view AI as something experimental or too complex for their operations. What misconceptions continue to hold service businesses back, and how should leaders rethink AI as a practical business tool rather than a technology project?
A: We’re seeing two distinct camps.
The first is tech-savvy business owners who are genuinely comfortable with technology and have experimented with consumer AI tools like ChatGPT on their own. But when it comes to bringing AI into their actual business, they get cautious because they understand exactly what liability they’re taking on.
Ironically, we see this group falling behind in AI adoption precisely because they know too much to move fast.
The second camp is less tech-savvy service business owners, and this group has surprised me. They’ve ended up at the frontier, acting more like innovators than laggards, because once they see AI move their topline or cut costs, they don’t overthink it—they just keep going. Quite a few of them have seen real, measurable results, both in revenue and efficiency.
There’s also a third group now emerging: the first wave of AI-traumatised business owners.
I know some who engaged an AI vendor about two years ago. Out of curiosity, I went to that vendor’s website to send an enquiry myself. I waited six hours for a human reply, and when it came, I was told the project had failed.
That kind of experience sticks with people and makes them wary of trying again.
My view is that AI adoption is still an experiment, and people should be careful going in. But if you never try, you never know. The companies that have committed to using AI properly over the last few years are the ones seeing genuinely outsized revenue gains and cost reductions, and that gap is only going to widen.
If AI Doesn’t Move the P&L, What’s the Point?
Q: Before founding Superbench, you worked in investment banking. How has that perspective influenced the way you evaluate business problems and identify where AI can create lasting value for traditional industries?
A: The clearest thing investment banking taught me is to always look at the business outcome.
From day one at Superbench, we were clear that we didn’t want to build a tool for the sake of building a tool. We wanted to build something that moves the P&L.
Every business owner wants more sales and better EBITDA margins, but we also understand it doesn’t always work out that cleanly. So when we’re building something for a client, we make sure it has a direct, measurable impact on specific business numbers, not just a flashy demo.
I also believe that any SaaS tool that only offers productivity gains, without ever proving its actual ROI, won’t survive long term.
That’s why we’re constantly working to improve the ROI our clients get from working with us and pushing to go deeper with them—automating more of the business over time rather than settling for a single-feature win.
The $10 Million Business With Three People
Q: Looking ahead, what do you think the service business of 2030 will look like? Which parts of the customer journey will always require a human touch, and where do you see AI becoming virtually invisible—simply part of how great businesses operate?
A: By 2030, I expect to see a lot of service businesses doing five to ten million dollars in revenue with a team of three.
One person will likely run the entire online experience with the help of AI. Two people will handle the offline experience. The rest of the cost structure, probably around 90%, will sit with the people actually delivering the service, which is exactly where it should be.
The customer journey is going to get genuinely interesting because we’ll start seeing real personal-concierge behaviour from AI.
Picture this: you want to book a massage this Friday at 2 pm, within a certain budget, and your AI personal assistant goes and finds the right business for you and books it, all without you touching a chat window yourself.
When that becomes normal, the customer a business is serving is no longer just a human—it’s also that human’s AI assistant. You’ll need to be able to take orders from both.
The businesses that figure out how to win agent-to-agent transactions, not just human-to-business ones, will have a real first-mover advantage.
But none of that changes where the human touch actually lives, which is in person, at the moment the service is delivered. That never goes away. It’s what the customer is paying for.
What changes is everything around it.
AI closes the speed gap: the response at 11 pm, the follow-up nobody had time to send, the booking that happens the second it’s requested instead of hours later.
Humans close the trust gap, in the room, on the job.
That’s the invisible economy I keep coming back to—not a future where service businesses talk about AI, but one where AI has fully disappeared into the speed of the business, while the human being in front of the customer stays exactly as central as ever.
Final Thoughts: The Best AI May Be the AI Nobody Notices
The AI race is often framed around who has the most advanced technology. Zhong’s experience with service businesses points towards another measure: who can turn that technology into a better-run business?
For SMEs, that means AI still has to prove itself commercially. Faster responses matter because customers are waiting. Automation matters when it reduces costs or frees people to focus elsewhere. And productivity matters when it eventually translates into sales, margins, or a better customer experience.
The more interesting possibility is what happens when all of this becomes ordinary.
Customers may eventually stop noticing whether AI answered an enquiry, followed up at 11 pm, scheduled an appointment, or communicated with their own AI assistant. They’ll simply experience a business that responds faster and works better.
The human interaction doesn’t disappear. Instead, technology recedes into the background so people can remain central where they matter most: delivering the service itself.
If that’s where the next phase of AI adoption is heading, some of its most consequential applications may not come from businesses trying to look more technological.
They may come from everyday businesses making the technology almost invisible.
Highlights
Read the Chinese article here.
