Riding the AI Shift: Why Tomorrow’s Startups Are Built with Fewer People and Bigger Ideas

When a major technological wave hits, the instinct for many founders is to immediately change their vocabulary. They sprinkle "AI" across their landing pages, bolt a chatbot onto their customer service tab, and hope they’ve done enough to stay relevant.

But according to Henson Tsai, the founder and CEO of SleekFlow, chasing the hype cycle is a fast track to nowhere. Real founders don't look at the technology; they look at the customer.

In a recent deep dive for the Macao Startup Club, Henson shared his raw insights on how he navigated SleekFlow’s massive, bold pivot into a 100% AI-native architecture with AgentFlow. More importantly, he laid out a clear blueprint for how AI today can radically supercharge an entrepreneur—without replacing the human heart of a company.

1. Spotting a Shift vs. Chasing a Trend

"I've started three companies and lived through a few hype cycles," Henson says, "so I've stopped judging these things by the technology and focus on the customer."

To Henson, a trend is superficial—it leaves customer expectations untouched. A shift, however, permanently rewrites what people expect from a business. In the messaging space, customers had been demanding instant, accurate, 24/7 answers for years, but the technology simply couldn't deliver. When AI finally closed that gap, SleekFlow didn't just add a feature; they rebuilt their entire platform.

For founders trying to figure out if they should pivot, Henson offers an unglamorous litmus test:

"Has my customer's definition of acceptable moved, and is there a cheaper way to meet it? If the answer is no, the existing model is already finished."

2. Workforce Design: What to Automate, What to Protect

Henson doesn’t view AI as a threat to human workers. Instead, he views it as an exercise in modern workforce design.

The rule of thumb is simple: Offload the work where the answer already exists. Order statuses, lead qualification, and 2 AM FAQs shouldn't be eating up your team's day. "Nobody grows by answering a question for the thousandth time," Henson points out.

By handing that volume over to AI agents, founders can fiercely protect the human traits that actually compound value:

• Creative problem-solving

• Conceptual strategy (deciding what to build next)

• Deepening genuine customer relationships

"A strong person paired with capable agents outperforms either one alone, by a wide margin," he says.

3. The New Standard of Customer Trust

The market appetite for AI is undeniable—last quarter, 87.6% of new SleekFlow customers chose an AI plan. But with that adoption comes a massive shift in user patience.

"People can feel when they're dealing with AI, and their patience for one that wastes their time is running out," Henson warns. A generic bot that gets stuck, provides useless answers, or hides the exit route back to a human will actively destroy trust.

For early-stage startups, Henson's advice is clear: Only automate outcomes your system can genuinely complete. Be honest that it’s an AI, and always make the route to a human instant and obvious.

4. Defeating Hallucinations and Securing Data

Two major fears hold founders back from adopting AI: hallucinations and data security. Henson believes both are entirely answerable.

To eliminate hallucinations, you must ground your AI agents strictly in verified knowledge bases, make them say "I don't know, let me pull in a colleague," and keep humans in the loop for irreversible actions like refunds.

As for security, Henson advises founders to aggressively interrogate their AI vendors across four fronts:

• Security: How is data isolated and encrypted?

• Governance: Does your data train a model you don’t control?

• Compliance: Look for independent standards (SleekFlow, for instance, went through ISO/IEC 42001 certification).

• Infrastructure: Where does the data physically live and how long is it kept?

5. The 2026 Lean Startup: Hiring "Relentless Builders"

If Henson were starting a brand-new company from scratch today, it would look completely different from his first startups. AI has shrunk the distance between an idea and a working product to virtually zero. Product leads can build rough versions instead of writing specs, and engineers can ship in days what used to take months.

Because of this, the ideal startup hire has changed. "I used to want people who could take a clear brief and execute it well," Henson says. "Now I want people who write their own brief, build the thing, find out it's wrong, and have the next version up by Friday."

In this new era, teams should stay small on purpose to avoid heavy coordination costs. But most importantly, the ultimate macro-decision remains entirely human.

"The one call I'd never give to a machine is which problems the company takes on in the first place. AI finishes the task well. Deciding whether the task should exist stays with the founder."

Team SleekFlow

Over to You, Macao Startup Community!

As we navigate this massive technological shift together, we want to hear from our local ecosystem:

1. Where is your boundary? If you look at your startup's workflow today, what is the single most repetitive task your team does that you could offload to an AI agent by tomorrow morning?

2. Hiring for the future: If you are scaling your team this year, are you looking for traditional project executors, or are you actively hunting for what Henson calls "relentless builders" who can self-direct with AI tools?

3. The trust factor: How is your startup addressing customer transparency and data security as you integrate smarter automation into your product?

Drop your thoughts by sharing this article on LinkedIn or IG and tag @MacaoStartupClub to keep the conversation going!

========================

Interested in joining the Macao Startup Club?

Become a member to connect with founders, innovators, and global startup ecosystems.

👉 Register here: https://www.link-card.xyz/associations/msc/membership