Southeast Asia’s startup ecosystem continues to mature rapidly, with founders increasingly building companies that operate across multiple markets from day one. Expansion into hubs such as Singapore, Hong Kong, the UAE, and the United Kingdom is often a natural next step, but navigating incorporation requirements, compliance obligations, accounting standards, and corporate governance can quickly pull focus away from innovation and growth.

As digital business infrastructure evolves globally, companies are looking for streamlined, technology-enabled solutions that reduce administrative overhead while maintaining regulatory confidence. This need is particularly pronounced in Southeast Asia, where regulatory environments vary widely. Osome addresses these challenges by combining AI-enabled software with expert human support, helping businesses manage incorporation, accounting, compliance, and corporate secretarial requirements so founders can focus on building and growing their businesses instead of financial admin.

Meeting Osome at Echelon Singapore 2026 offers founders and operators a chance to explore how streamlined business infrastructure can support faster expansion. Whether entering Singapore for the first time, scaling regionally, or preparing for international growth, practical guidance on compliance and operational setup can make a measurable difference in execution speed and confidence.

Osome: Leading the way in digital business services

Osome is a digital business service provider focused on helping entrepreneurs start, manage, and grow companies across Singapore, Hong Kong, the United Kingdom, and selected international markets. Its mission centres on making entrepreneurship more accessible by removing administrative and compliance barriers that often slow early stage and growth stage companies.

The company combines AI and automation with professional expertise to deliver incorporation support, accounting services, corporate secretarial solutions, and compliance management through an integrated digital platform. This hybrid approach allows businesses to benefit from efficiency and transparency while still having access to experienced professionals when navigating regulatory requirements.

Osome’s regional relevance is particularly strong in Southeast Asia, where cross-border expansion is increasingly common. Partnerships with venture capital firms, accelerators, incubators, co-working spaces, and other ecosystem enablers further strengthen its role as a practical support layer for founders scaling across markets.

Also read: Scaling Southeast Asia: Who to meet at Echelon Singapore 2026

Meet Osome at Echelon Singapore 2026

Osome joins Echelon Singapore 2026 as a preferred incorporation partner, alongside other ecosystem leaders, founders, investors, and innovators gathering at Suntec Singapore CEC on 3-4 June 2026. The two-day event brings together the region’s startup community through content stages, exhibitions, networking sessions, and knowledge-sharing opportunities designed to support Southeast Asia’s innovation economy.

Attendees can connect with the Osome team to discuss incorporation strategies, compliance planning, and cross-border expansion considerations. The company will offer consultation packages and discounted incorporation services tailored for startups and SMEs attending Echelon, creating a practical entry point for businesses preparing to scale regionally or internationally. Book a call with Osome here to get started.

As Southeast Asia’s innovation economy continues to evolve, strong operational foundations are becoming essential to sustainable growth. Collaboration between founders, infrastructure providers, investors, and ecosystem enablers helps ensure that companies can scale efficiently while remaining compliant and resilient. Platforms like Echelon play a key role in enabling these connections and supporting the next phase of regional innovation.

The region is evolving quickly, and Echelon Singapore 2026 offers the right place at the right moment to be part of what comes next. Register here to join the conversation.

—

Want updates like this delivered directly? Join our WhatsApp channel and stay in the loop.

The e27 team produced this article

We can share your story at e27 too! Engage the Southeast Asian tech ecosystem by bringing your story to the world. You can reach out to us here to get started.

Featured Image Credit: Osome

The post How Osome is enabling smoother business expansion for startups appeared first on e27.

When global researchers predicted that speech recognition for low-resource languages wouldn’t be commercially viable until 2030, we had already achieved 97 per cent accuracy in 2020. This isn’t a story about one company – it’s about how developing nations can build AI sovereignty by solving real problems, building infrastructure patiently, and getting creative with talent.

After 25 years of building language technology in Mongolia – from basic character rendering to complex reasoning systems – I’ve learned that the path to AI independence requires three things: sustainable development, pragmatic focus on real missions, and a bigger vision. Here’s what actually works.

Start where problems challenge society

Silicon Valley starts with solutions looking for problems. Developing nations must start with problems demanding solutions. Mongolia’s AI journey began with frustrations that might sound trivial to outsiders but were paralysing daily life:

• The typing crisis: We literally couldn’t type in our traditional vertical script. Eight centuries of written culture were becoming digitally extinct. This wasn’t a market opportunity – it was cultural death in slow motion.
• The dictionary that wasn’t: No comprehensive digital Mongolian-English dictionary existed. Students and professionals relied on dusty paper editions from the 1960s. Every missing translation represented a lost opportunity.
• The spellcheck absence: Mongolian text had no spell-checking. Government documents, business contracts, and academic papers were riddled with errors that undermined credibility and caused actual legal disputes. That was one of the biggest social problem until 2017.
• The language corruption crisis: Here’s something that keeps me awake at night – we’re raising a generation that speaks broken Mongolian. Young people unconsciously rely on Google Translate for homework, social media, and daily communication. But Google Translate’s Mongolian support is so poor that it’s literally corrupting our language. Students write essays in grammatically incorrect Mongolian, thinking it’s proper because “Google said so.”
• The speech recognition gap: Parliament and government operate through meetings, making transcription a critical function. Without speech recognition, protocol-keeping consumed enormous human resources. And solutions had to be on-premise for security.
• The service crisis: Banks and telecoms haemorrhaged money on customer service. Without automated Mongolian language support, every query required human agents.
• The procurement corruption: At Erdenet Mining, Mongolia’s largest company, manual tender evaluation enabled systematic corruption. Millions disappeared into rigged contracts or excessive delays in purchase completion.

Each problem we solved provided data, experience, and credibility for the next challenge. Our fully automatic tender evaluation system at Erdenet didn’t just save money and reduce corruption – it generated the revenue and trust that funded our speech recognition and language model research.

This problem-first approach built our capabilities organically. By the time we tackled large language models, we had accumulated 20 years of linguistic data, domain expertise, and customers who trusted us.

Harness your diaspora’s expertise

Mongolia has more software engineers in Silicon Valley, Berlin, Seoul, and Tokyo than in Ulaanbaatar. This brain drain, common across developing nations, became the cornerstone of our talent strategy.

The diaspora advantage

We couldn’t compete with FAANG salaries. Instead, we offered something money can’t buy: the chance to build their homeland’s technological future. Our pitch was simple: “Your skills + our mission = your legacy.”

AI Engineers from Germany, Switzerland, Italy, and the US were invited to work on-site, fostering knowledge exchange with local talent. They mentored, contributed code, and validated technical decisions, creating a sustainable talent flow instead of a one-way brain drain. This approach assured young engineers that they could venture out, gain experience, and return to impactful work.

Also Read: On-chain data and Web3 security: Insights from industry experts

Build infrastructure with extreme patience

Infrastructure has been our greatest challenge from day one. We move slowly, but we move forward. Our datacenter evolution tells the story of patient building:

We started in 2017, purchasing two gaming GPUs (RTX 2090 Ti). By 2019, we had received angel investment to buy our own L40S GPUs. In 2021, we added A6000 units. Today, we run our own infrastructure with local hardware for sensitive data. We use cloud resources from AWS and Google to train bigger models.

We manage everything ourselves – not by choice, but by necessity. No managed services exist for our use cases. This forced us to develop deep infrastructure expertise that later became a competitive advantage.

The lesson? Start with whatever computer you can access. Our first speech recognition model was trained on two gaming GPUs for one year. Perfect infrastructure is a luxury, developing nations can’t afford to wait for.

Competitive advantage

Cultural localisation

The assumption that AI models can simply be “translated” fundamentally misunderstands how language and culture intertwine.

The mixed language reality

Mongolians don’t speak “pure” Mongolian. Never have, never will. Real conversations flow like this:

• Mongolian grammar structures with English tech terms embedded
• Russian expressions from the Soviet era sprinkled throughout
• Chinese trade phrases when discussing business

Our models had to understand “Kodoo push hiigeed product ownertoo мэдэгдчихлээ.” (I’ve pushed the code and notified the product owner.). When we insisted on linguistic purity, our models failed spectacularly. When we embraced messy reality, they worked.

Release strategy that makes sense

While big tech companies focus on the “top 20 languages” and throw everything else into an “other” bucket, we take a different approach. For each market, we focus on the languages that actually interact. For Mongolia, that means:

• Mongolian (obviously)
• English (global connectivity)
• Russian (historical ties)
• Chinese (trade relationships)

This connected-language approach delivers 10x better results than generic multilingual models.

Also Read: High adoption, low understanding: The Philippines’s blockchain knowledge gap

Data sovereignty

Your data will be securely stored within your country’s territory, allowing you to use it with peace of mind. If governments fail to actively embrace and utilise artificial intelligence, they risk falling behind, potentially leading to significant social inequality within their societies.

Digital sovereignty through pragmatic building

True sovereignty means controlling critical layers while pragmatically using what works.

What we built ourselves:

• Character encoding and input methods (100 per cent local)
• Finate state automata and transducers as language complexity (100 per cent local)
• Speech recognition and synthesis (100 per cent local)
• Core language models (100 per cent local)
• Application frameworks (80 per cent local, 20 per cent open source)
• Infrastructure management (60 per cent local, 40 per cent standard tools)

The honest trade-offs:

• We use NVIDIA hardware (no alternative yet)
• We leverage open-source frameworks (why reinvent PyTorch?)
• We adapt international research (standing on the shoulders of giants)
• But we control all critical paths and data

Economic sustainability:

Sovereignty without economic sustainability is just expensive nationalism. Our model evolved through necessity:

• Government as first customer, not sugar daddy
• Enterprise solutions funding research
• Open source contributions building global goodwill

The uncomfortable truths

Let me be honest about what building AI in a developing nation really means:

• You’ll never have enough resources. We have 1/1000th of OpenAI’s budget. Make constraints drive innovation. Our GPU shortage forced us to optimise models that now run on minimal hardware
• Technical debt is inevitable. Plan refactoring cycles from day one. We still maintain code from 2017 because it serves critical government systems.
• International competition will arrive. Build cultural moats early. By the time OpenAI or Google supports Mongolian properly, we’ll have ten years of local context they can’t replicate.
• Talent will always be scarce. Create compelling missions. We can’t match Silicon Valley salaries, but we offer something better: the chance to preserve their culture through technology.
• Trust is your biggest challenge. When we first claimed 97 per cent accuracy in Mongolian speech recognition, nobody believed us. International researchers said it was impossible. Now, the customers and government say, “Why not just use ChatGPT?” You need rock-solid local success stories. Building credibility took years of consistent delivery.

Also Read: Asia rises in the AI chip race: China to outgrow US by 30 per cent by 2030

The future is distributed

The era where a few companies in a few countries control global AI is ending. The future belongs to distributed, culturally-rooted AI systems serving specific populations with a deep understanding.

Mongolia’s journey from missing keymaps to billion-parameter models proves that developing nations don’t need charity. Don’t blame Silicon Valley for ignoring your market – they don’t owe you anything. Build it yourself.

With sustainable development focused on real problems, pragmatic building strategies, and a vision for digital sovereignty, any nation can achieve AI independence. Don’t be discouraged if someone says you can’t compete with OpenAI or Anthropic. While they are currently reducing their model parameters, we are increasing ours. In the near future, we will converge in the middle. The first to reach that point could be the winner, but nevertheless, business applications using models under 30B parameters will capture 90 per cent of the whole AI market share.

The question isn’t whether to build homegrown AI, but whether your problems are painful enough to sustain the decades-long journey to solve them. For us, watching our youth lose their mother tongue to bad machine translation was painful enough.

But when a nomadic herder can speak to AI in their own language and get help accessing government services? When corruption drops because algorithms can’t be bribed? When does your culture live digitally for future generations? Egune AI is making it possible.

—

Are you ready to join a vibrant community of entrepreneurs and industry experts? Do you have insights, experiences, and knowledge to share?

Join the e27 Contributor Programme and become a valuable voice in our ecosystem.

Image credit: Canva

The post Homegrown AI: Mongolia’s blueprint for developing nations appeared first on e27.

Over the past two quarters, I’ve been teaching digital transformation in the era of 5G and AI, which is something new to me, but deeply rewarding. One of my first questions to every new class is simple: “What brought you here?”

Surprisingly, more than half the class often says they’re here because their SkillsFuture credits are expiring soon. They’re not chasing certificates or career pivots — they simply don’t want to waste the opportunity to learn something relevant in a fast-changing world.

Many of them are staff from SMEs, people who didn’t come in with big expectations but with an open mind. By the end of the course, they often walk away feeling genuinely proud — not just for learning new skills, but for discovering that they can.

It made me reflect on something remarkable about Singapore’s ecosystem. While we often talk about being a “Smart Nation” in terms of infrastructure and innovation, what’s truly smart is how the government has quietly built a learning mindset into the fabric of society. Every citizen has access to resources to up-skill, re-skill, or simply stay curious, and that creates a compounding effect across the workforce.

The ripple effect on startups and SMEs

When we think of innovation, the spotlight naturally falls on the usual suspects — the Y Combinators, the Antlers, the tech founders experimenting with Claude 4.5 for code generation, Meta’s AI glasses, or OpenAI’s agents. These are the pioneers pushing the frontier.

But real transformation happens when AI and digital literacy reach the everyday office worker — the operations executive, the logistics coordinator, the sales manager. The future isn’t just being built in accelerators or venture labs; it’s being sustained by people in SMEs who are now using digital tools, automating manual processes, and thinking differently about their work.

Technology only fulfils its potential when it becomes widely usable. No matter how advanced a model is, if only a few understand it, it remains an elite game. But when the broader population starts to experiment — even in small ways — that’s when innovation scales.

Also Read: Building the future: Up-skilling and empowerment in India’s real estate boom

Giving the “back row” a front-row seat

In every learning journey, there’s a simple truth: we don’t know what we don’t know. AI is evolving so fast that even tech professionals struggle to keep up. For non-tech professionals, programs like SkillsFuture create an entry point — giving them a shared vocabulary, a way to join the conversation instead of being left behind by it.

Owning an AI tool today is like owning a smartphone 15 years ago. You might not use every feature, but it changes how you live, work, and connect. The same applies to AI – it’s becoming the new baseline for digital fluency.

The dual movement of a smart nation

What I find admirable about Singapore’s approach is its two-speed innovation model: one that accelerates frontier technologies through startups and MNCs, and another that ensures the entire society moves along with it. The front-runners may be building the future, but the rest are being equipped to live and thrive in it.

That’s what makes the ecosystem resilient. Startups can scale faster when the broader workforce is digitally competent. SMEs can adopt innovation without fear. And individuals, regardless of age or industry, can adapt rather than resist change.

A shared momentum

Whether you’re at the front driving innovation, at the back learning to catch up, or somewhere in between enabling the system, we’re all part of the same movement.

The next wave of transformation won’t just be about smarter technologies; it’ll be about smarter people using them with confidence. And if there’s one thing Singapore’s story shows us, it’s that when everyone learns, the nation moves forward together.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. Share your opinion by submitting an article, video, podcast, or infographic.

Join us on Instagram, Facebook, X, and LinkedIn to stay connected.

Image credit: Canva

The post When everyone learns, the nation moves forward appeared first on e27.

Today’s workforce is undergoing rapid transformation. As AI reshapes every industry, organisations and employees are discovering new ways of working, creating a need for different professional skills and knowledge. For employees, this has both opened opportunities and sparked anxiety.

Many are experiencing FOBO (Fear of Being Obsolete), with nearly half of professionals worrying their skills will be outdated within five years. At the same time, 94 per cent say they would stay longer at an organisation that invests in career development, showing that up-skilling opportunities are a top priority for professionals.

Yet, despite the necessity and demand for learning, skills gaps remain from digital fluency to problem-solving and adaptability. This work landscape is proving to be particularly challenging for Gen Z, who are newer to the workforce and must learn foundational professional skills at a time of rapid change. Older employees may also find this integration of new technology overwhelming and need support with digital skills.

The challenge for organisations is clear: provide employees with the learning support they need to prepare for the future of work. Edutech offers a pathway to achieving this goal by delivering scalable, engaging, and personalised learning opportunities tailored for modern employees.

Drive continuous up-skilling and adaptability with micro-learning

With AI accelerating the pace of change, organisations can no longer afford to only deliver employee training in all-day sessions once or twice per year. Instead, edutech tools enable managers to embed learning into the flow of everyday work with micro-learning. Mobile-friendly and bite-sized training formats ensure learning can happen anytime, anywhere, and are especially appealing to Gen Z as digital natives. These sessions are easier to fit into employees’ busy schedules without disrupting work routines, while also making learning a continuous process that can be quickly adapted according to changing needs.

Micro-learning can be designed to help employees build skills and knowledge step-by-step as they need it. This lets employees see the relevance of the training right away, which is a key factor in keeping learners engaged and helps employees avoid information overload.

Also Read: The future of edutech: Personalising learning for all

Edutech can also make it easier for employees to track their learning progress, filling knowledge gaps and developing new skills at their own pace. For employers, integrated reporting helps track participation across training sessions, identify skill gaps, and refine future learning strategies. With full visibility of their workforce’s strengths and weaknesses, and how they respond to the training in real time, trainers can continuously adapt their approach to be more efficient and impactful.

Add interactivity to make learning engaging and effective

Traditional training methods like long modules, static slide decks, and one-way seminars often fail to keep employees motivated. With edutech, trainers can design learning experiences that are interactive, using gamification, real-time feedback, and adaptive learning paths that make sessions memorable and effective.

In fact, research has shown that interactive learning lowers employee attrition rate by 37 per cent.

A new innovative approach to training doesn’t need to take a lot of extra time or effort, either. Organisations can use AI to transform existing training materials into interactive or even gamified experiences. AI-enhanced edutech can also make it easier to personalise learning, giving employees highly relevant training that’s suited to their needs and learning styles. With the right tools, even compliance training or technical refreshers can become experiences employees look forward to.

Encouraging collaboration and knowledge sharing

Beyond individual learning, an AI-ready workforce requires collaboration and collective problem-solving. Peer-to-peer learning and knowledge exchange sessions can be a powerful way to connect teams while promoting learning. However, many employees would hesitate to give a presentation to their colleagues even with valuable information, simply for fear of boring their coworkers.

Edutech platforms can help professionals turn these dry sessions into exciting and energising activities by using friendly competition or collaborative challenges. Session hosts can test their coworkers’ knowledge with gamified quiz questions, and use live polls, Q&As, and brainstorming to ensure everyone has a voice, even in hybrid or global teams. Employees can also use AI tools to create these peer learning sessions, helping them build AI skills at the same time.

Also Read: Edutech in SEA is ripe for acceleration. This is why they can help build a more inclusive society

Preparing today for tomorrow’s workforce

As AI transforms industries at a lightning-fast pace, the workforce must evolve just as quickly. Edutech can play a crucial role in this transformation by making learning engaging, accessible, and collaborative, and enabling organisations to deliver it at scale across their workforces.

Organisations that invest now in embedding edutech into their talent strategies will not only be better equipped to future-proof their workforce; they can also unlock new levels of productivity, retention, and innovation. Ultimately, the intersection of AI and edutech is about empowering people with the confidence, curiosity, and collaboration skills they need to thrive in the years ahead.

—

The post The future of work is here: The role of edutech in an AI-ready workforce appeared first on e27.

In the creator and startup world, “community” has become a buzzword.

Everyone wants one. Creators launch Telegram groups. Founders open Slack channels. Professionals start networking circles.

But most communities don’t monetise. And it’s not because of low engagement. It’s because there is no economy running behind them.

At our recent “Community That Converts” session, this pattern became obvious. The room was filled with creators, early-stage founders, and corporate professionals trying to pivot. They didn’t lack ambition. They didn’t lack skill. Many of them already had audiences.

What they lacked was structure.

A group is not an ecosystem

There’s a misconception that if you gather enough people in one place, value will naturally circulate.

It doesn’t. A group is a container. An ecosystem is infrastructure.

An ecosystem has:

• Defined roles.
• Defined value propositions.
• Defined price points.
• Defined outcomes.
• Defined pathways for exchange.

Without these, what you have is a connection. Not commerce.

And connection alone is not sustainable.

Even charities need funding to operate. Even volunteer-driven movements require resources. In today’s economy, value is transacted through money. If no monetary exchange exists, sustainability becomes fragile.

The hard truth is this: If your community has no internal economy, it will eventually stall.

The “more content first” trap

One recurring belief we encountered was this: “I need more content first.” More posts. More reach. More followers. But content amplifies what already exists.

If the value proposition is unclear, more content amplifies confusion. If the offer is undefined, more visibility magnifies the gap. People cannot pay for what they do not understand.

And many community builders avoid pricing not because they are unsure of their audience, but because they are unsure of their own value.

Also Read: Customer churn analysis: How can startups get it right?

Monetisation is not hard selling — it is a measurement

Monetisation is often framed emotionally, as something uncomfortable or aggressive. But monetisation is simply an exchange of value transacted through money. Money measures clarity.

If no one is willing to pay, one of three things is usually happening:

• The outcome is vague.
• The positioning is weak.
• The offer is undefined.

You cannot expect payment without a price tag. You cannot expect sustainability without structure.

Consider this in professional terms. You do not pay an intern a manager’s salary. You do not pay a junior consultant a partner’s rate. Price is determined by positioning, experience, outcome, and brand equity.

The same applies to communities.

If members do not know:

• What they are offering,
• What problem do they solve?
• What outcome do they deliver?
• What they charge,

There can be no internal economy. And without economy, there is no conversion.

From individual hustle to layered ecosystems

A monetisable community is not about everyone paying one central figure. It is about enabling structured value exchange within the ecosystem.

Imagine a business-focused environment that includes:

• A strategist.
• A videographer.
• A photographer.
• A PR consultant.
• A lead generation specialist.
• A copywriter.

Each has a defined value proposition. Each has pricing. Each understands positioning.

When someone in the ecosystem needs support, they do not go outside. They transact within. That is when a community becomes infrastructure. That is when it converts. The platform itself becomes the enabler of value exchange.

Belonging is necessary, but directional

Belonging matters. But belonging without direction becomes noise.

Communities thrive when members share a similar outcome. An investment-focused group works because members aim for capital growth. A health-driven group works because members prioritise wellbeing. A business-focused ecosystem works because members want leverage, revenue, and growth.

The shared ambition defines the economy. Not just shared interests, shared direction.

Also Read: From bridge rounds to global awards: Startups across Asia keep building

Why most communities fail

Most communities fail because they are built emotionally, not structurally.

They prioritise:

• Inclusiveness without tiers
• Engagement without offers
• Networking without pricing
• Connection without defined outcomes

It feels good initially. But without an economic layer, energy dissipates. The most sustainable communities are both inclusive and exclusive. Inclusive in knowledge-sharing and access. Exclusive in-depth, commitment, and paid tiers.

Open the front door. Structured inner rooms. Free distribution. Paid progression.

That structure is what enables revenue, confidence, clarity, and leverage — not just for the founder, but for the members themselves.

The blueprint: What makes a community convert

A converting community requires five foundational elements:

• Clear value proposition: Members must know exactly what they offer and to whom.
• Defined outcomes: The community must rally around a specific ambition or transformation.
• Structured pricing: Offers must be priced clearly. No ambiguity. No hidden discomfort.
• Tiered access: Free access for exposure. Paid tiers for depth and accountability.
• Facilitated exchange: The platform actively enables members to transact within the ecosystem.

When these layers exist, monetisation is not forced. It becomes natural. Because value is visible.

The bigger shift

The creator economy is maturing. Visibility is abundant. Content is easy. AI compresses execution.

What differentiates sustainable founders now is not output. It is ownership. Ownership of:

• Audience channels.
• Pricing structures.
• Ecosystem design.
• Internal economies.

A large audience without structure is fragile. A smaller, well-architected ecosystem is durable.

A community that converts is not about building a room full of people. It is about building a marketplace of defined value.

And once value can be exchanged clearly, sustainably, and repeatedly, conversion is no longer accidental. It is engineered.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. Share your opinion by submitting an article, video, podcast, or infographic.

Enjoyed this read? Don’t miss out on the next insight. Join our WhatsApp channel for real-time drops.

The post Why your community needs an economy, not just engagement appeared first on e27.

Yesterday, while grabbing coffee after church, a friend vented to me, “Aleks, why does my dad’s generation trust AI-generated influencers so much? They treat them like real people with credible opinions, but won’t listen when I tell them to look for actual facts.”

It instantly hit me: the marketing world is facing the exact same problem. There’s a massive disconnect between what we think AI is doing and what the audience actually feels. Recent Braze research highlighted by MarTech proves this point perfectly: while 93 per cent of marketers are convinced AI helps them understand customers better, only 53 per cent of consumers feel brands accurately predict their needs.

We find ourselves in a “high-capability, low-trust” paradox. While marketing teams are sprinting ahead with AI adoption, consumers remain deeply sceptical about how and why brands are using these tools.

The core issue isn’t a technology problem; it’s a trust problem. And for brands looking to scale across diverse markets in Asia and beyond, realising that difference is the key to survival.

The rise of the AI intermediary

The implications of this trust gap are profound. Today, 19 per cent of consumers are already using AI intermediaries to interact with brands, a number expected to surge to 46 per cent in the near future. However, more than half of consumers expect brands to use AI in self-interested ways (like cutting costs) rather than to actually improve the customer experience.

When your customers view your technology as a barrier rather than a benefit, visibility and value clarity matter more than ever. Marketers are no longer just optimising for search results or social algorithms; they are optimising for how an AI assistant interprets, filters, and recommends their brand.

This perception gap cannot be solved with better internal dashboards or more efficient media buying alone. It requires transparent, tangible value demonstration. The customer needs to see and feel how AI makes their experience better.

Enter creative differentiation: The META’s perspective

How do brands bridge this gap? The answer lies in moving away from algorithmic homogenisation and leaning heavily into what Meta defines as creative differentiation.

Consumers don’t want to feel targeted by a machine; they want to feel understood by a brand. For example, in our company, SOMIN, we recognise that as the most powerful application of AI, which isn’t finding cheaper ways to serve the same generic ad — it’s using AI to fuel ad diversification and rapid content expression development.

Also Read: When AI stops being a feature and starts being infrastructure

When brands rely on a single, rigid visual or message, ad fatigue sets in, and the AI comes across as a cold, calculating retargeting tool. By introducing ad diversification, brands can break out of this trap. We use AI-powered insights to analyse audience behaviours, motivations, and cultural nuances, allowing brands to develop a wide array of tailored “content expressions.”

This means a single campaign can dynamically present different visual cues, messaging angles, and ad formats based on what genuinely resonates with a specific, hyper-local audience segment.

When AI is used to drive creative differentiation, the consumer experience shifts dramatically. Instead of feeling stalked by repetitive, self-serving corporate messaging, consumers are served varied, highly relevant content that speaks to their specific pain points and interests. The AI stops feeling algorithmic and starts feeling authentic.

The path forward

The brands that succeed in the next era of digital marketing will be the ones that can explain their AI usage through actions, showing how it improves customer outcomes rather than just internal operational efficiency.

As the MarTech data suggests, moving fast breaks things, and right now, it’s breaking customer trust. Rebuilding that trust requires connecting AI capability with genuine cultural relevance. By embracing diversification and prioritising diverse content expressions, brands can finally prove to their customers that AI isn’t just about the bottom line; it’s about building a better, more personalised relationship.

Are you experiencing the AI trust gap in your customer interactions? It might be time to rethink how your creative speaks to your audience.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. Share your opinion by submitting an article, video, podcast, or infographic.

Enjoyed this read? Don’t miss out on the next insight. Join our WhatsApp channel for real-time drops.

The post Bridging the AI trust gap: Why ad diversification and creative differentiation are the future of customer connections appeared first on e27.

myFirst, a Singapore kids’ tech company selling “first devices” for children, has secured over US$8 million in a Series A round led by Vertex Ventures Southeast Asia & India, as investors chase a category many parents already treat as unavoidable: giving children connected tech, but trying not to hand them the keys to the entire internet.

The company sits in the fast-growing space between two uncomfortable extremes — kids borrowing a parent’s smartphone, or kids getting their own and immediately colliding with adult platforms, adult content, and adult incentives. myFirst’s answer is a tightly controlled ecosystem built around its myFirst Fone watchphone and the myFirst Circle family app, which it says is already used by more than one million families across 60 countries.

Also Read: How myFirst aims to provide a safer social media experience for children

“We started myFirst because our own kids wanted to use technology, but everything out there was built for adults,” said founder and CEO G-Jay Yong.

A safety-first ecosystem built for children, not retrofitted from adult tech

myFirst’s technology stack is designed around a simple idea: children should be able to communicate and create without being exposed to the open social web.

At the hardware level, the flagship watchphone is positioned as a compromise product: kids can call and message, but parents aren’t effectively giving them a pocket computer with an app store. The company highlights GPS tracking, safety zones (geofencing), and an SOS button—features aimed at reassurance and rapid response rather than entertainment.

The ecosystem approach goes further than a single device. myFirst sells a range of kid-targeted gadgets (watchphones, cameras, headphones, drawing tablets and more) paired with connected services. The point is less about piling up screens and more about creating controlled entry points into digital behaviour: communication, creativity, and self-expression.

The core safety design principles, as described by the company, include:

• Closed networks over open platforms (children interact only within approved circles)
• Parental controls baked into the product, not bolted on after the fact
• Encrypted environments for sharing and messaging
• No ads, reducing the risk of behavioural targeting and engagement traps
• Reduced exposure to strangers by limiting who can contact or view a child’s content

Taken together, myFirst is trying to sell parents something the mainstream internet struggles to offer: a smaller, safer digital world that’s still functional.

myFirst Circle: social networking without the stranger-danger business model

The biggest differentiator is myFirst Circle, which myFirst describes as a “closed, ad-free social network designed specifically for children”.

Also Read: GeckoLife wants to protect the digital privacy of you and your children

That sets it apart from traditional adult social media in three key ways:

• Access is permissioned: kids share moments only with approved family and friends, rather than broadcasting to the public or algorithmic discovery feeds.
• The environment is parent-controlled: the platform is designed around guardians managing contacts and boundaries, rather than expecting children to self-moderate.
• It’s not ad-driven: without advertising as the engine, there’s less incentive to maximise time spent, provoke engagement, or collect extensive behavioural data.

In other words, myFirst Circle is attempting to keep the “social” part (sharing, reacting, staying connected) while stripping out the parts that have made adult social platforms a headache even for adults.

Where the money goes: retail, telcos, and a push outside Southeast Asia

The Series A funding will be used to deepen the kidstech ecosystem and expand distribution, with myFirst pointing to retail and telco partnerships as its main route into new markets.

The company says it plans to scale across North Asia, the Middle East, the US, and Europe

It also namechecked large retailers Walmart and Best Buy as partnership examples, signalling a strategy built around mainstream consumer shelves and carrier bundles rather than niche online-only sales. For a kids’ device category—where trust, warranties, and physical retail visibility matter—that approach could be decisive.

Jessica Koh, Senior Executive Director at Vertex Ventures SEA & India, framed the bet as a shift in consumer behaviour: “We see a structural shift in how families introduce children to technology, driving demand for purpose-built products rather than adaptations of adult platforms.”

The real battle: defining “first tech” before the smartphone does

myFirst is effectively trying to claim the moment when a child moves from offline to online before a smartphone becomes inevitable. The company’s wager is that parents will pay for a safer on-ramp if the products are credible, not clunky, and don’t feel like a punishment device.

Also Read: Unlikely mentors: What kids can teach you about entrepreneurship

The challenge, as always in children’s tech, is balancing freedom and safety without turning the experience into surveillance theatre. But with fresh capital and an explicit global expansion plan, myFirst is pushing its case that kids shouldn’t have to start their digital lives on platforms built for adults—and parents shouldn’t have to choose between connection and chaos.

myFirst as one of the first 15 semifinalists of e27‘s TOP100 programme in 2023

The post myFirst bags US$8M to build a safer first internet for kids appeared first on e27.

Agents Stack positions itself as an advisory board for the C-suite at a time when AI is reshaping how enterprises operate. Founded in February 2025 by Aanchal Gupta, the consultancy works with senior leaders navigating AI transformation, offering expertise across AI, digital transformation, cybersecurity, sustainability, ethics, and governance.

Rather than pushing tools, Agents Stack focuses on tightly defined, strongly guardrailed use cases designed to deliver measurable outcomes. The company works with clients across banking, financial services and insurance, manufacturing, maritime and real estate, among others.

Its premise is straightforward: efficiency gains from AI are only sustainable when governance comes first.

“When you deploy AI, how do you make sure that it’s secure and has ethical practices and governance built into it by design?” Gupta said in a call with e27.

For Agents Stack, governance is not an afterthought but the foundation upon which digital efficiency is built. It works closely with customers not only on strategy but also on change management and implementation, helping organisations embed AI responsibly within their operations.

Also Read: Bridging the AI trust gap: Why ad diversification and creative differentiation are the future of customer connections

Before formally launching, Gupta spent much of 2025 meeting more than 1,000 C-level executives across the Philippines, Singapore, Malaysia, Thailand, the Middle East, Europe, and India. The objective was not to pitch, but to listen.

The most consistent feedback was not a lack of tools, but too many of them. Executives cited an overload of cybersecurity products and information, coupled with limited visibility over actual risk exposure. What they sought was a trusted partner that could consolidate and simplify the landscape, cut through vendor noise and provide clear oversight of cyber threats. That insight now shapes the company’s strategic direction.

In 2026, Agents Stack’s priorities centre on critical infrastructure, operations and supply chain risk — areas Gupta sees as the most pressing vulnerabilities in the coming year. Global expansion is also high on the agenda, alongside helping customers consolidate vendors to reduce complexity and cost.

Geographically, the company plans to extend its footprint into the Middle East, the European Union, Japan, South Korea and Africa. Its service scope is expanding in parallel, with end-to-end risk advisory spanning climate and nature risk, operations risk, supply chain risk and AI risk.

The emphasis on operations technology and supply chain exposure reflects a broader shift in the cybersecurity landscape. As hardware and software become more tightly integrated in sectors such as energy, railways and defence, the attack surface widens.

Also Read: Purpose or drift? America’s Iran strategy, global shockwaves, and Tehran’s post-strike trajectory

According to Gupta, operations and supply chain risk in critical infrastructure will be among the most significant threat vectors in 2026, particularly as AI-driven systems become more prevalent.

In the Asia Pacific, governments and regulators are sharpening their focus on these vulnerabilities. Policies targeting data protection, operational resilience and supply chain transparency are gaining traction. Gupta expects regulatory attention on operations and supply chain risk to intensify, given the growing exposure created by digital transformation.

At the same time, she argues that long-term resilience will depend on capacity building. Agents Stack is prioritising public-private-education integration across Asia Pacific to strengthen talent pipelines and foster collaboration. With limited cybersecurity talent and rising costs, enterprises cannot address every risk simultaneously. Strategic prioritisation is essential.

As generative AI and automation accelerate, Gupta advises organisations to adopt a calculated approach. Security leaders must prioritise high-risk parameters and accept that not every vulnerability can be mitigated at once. “There is too much noise with a lot of tools and vendors on data visibility and protection,” she said. What matters is meaningful coverage of the most critical risks.

In an environment defined by rapid technological change and regulatory complexity, Agents Stack is positioning itself as a steady hand for the C-suite — one that combines governance by design with practical risk prioritisation. As it expands globally in 2026, the company’s bet is that clarity, consolidation and trust will matter more than ever in the age of AI-driven transformation.

—

Image Credit: Agents Stack

The post Governance before efficiency: How Agents Stack guides AI adoption for businesses appeared first on e27.

South Korean President Lee Jae Myung’s stop in Singapore before heading to the Philippines was more than ceremonial diplomacy — it was a strategic signal.

At the Korea-Singapore AI Connect Summit, Lee positioned artificial intelligence not merely as a technology trend but as a national competitiveness imperative. His announcement of a US$300M global fund to be established in Singapore by 2030 marks Korea’s first offshore master fund dedicated to AI and deeptech startups, underscoring Seoul’s intent to anchor itself deeper within Southeast Asia’s innovation ecosystem.

The proposed Korea-Singapore AI Alliance adds institutional weight to that ambition, linking policymakers, researchers, and industry leaders in what Lee described as an open innovation framework. The emphasis on joint research to address global challenges suggests a bid to move beyond capital flows toward co-development and shared intellectual leadership.

For two resource-scarce but talent-rich economies, collaboration in AI is both pragmatic and symbolic. As the sector remains dominated by the US and China, Seoul and Singapore appear to be betting that middle powers can still shape emerging technologies — not alone, but together.

REGIONAL

S Korea to launch US$300M AI startup fund in Singapore by 2030: The fund will invest in promising AI and deep‑tech startups from both countries. The fund would be Seoul’s first offshore global master fund and proposed a Korea‑Singapore AI Alliance.

myFirst bags US$8M to build a safer first internet for kids: Vertex Ventures is the lead investor. myFirst’s technology stack is designed around a simple idea: children should be able to communicate and create without being exposed to the open social web.

Dyna.Ai lands eight-figure Series A to move banks beyond AI pilots: Investors include Lion X Ventures and ADATA. Singapore-based Dyna.Ai aims to deploy agentic AI inside banks, betting enterprises are shifting from pilots to governed, results-driven systems built for measurable operational execution.

Toku emerges stronger from IPO year with AI suite driving monetisation: Toku grew FY2025 revenue 9.3% to US$34.8M, with usage revenue up 21%, improved EBITDA, reduced debt, and strengthened finances post-IPO, positioning its AI platform for margin expansion and global growth.

Indonesia launches platform to combat rising gaming addiction: DARA is a multiplatform private counselling service designed to help children and families manage gaming addiction. Studies show about 33% to 39% of high school students in some samples fall into moderate to severe gaming addiction.

FEATURES & INTERVIEWS

‘Profitability is an inflexion point, not the finish line’: PolicyStreet CEO: Khazanah-backed PolicyStreet posted over US$1 million in profit after 2.5x revenue growth, with CEO Yen Ming Lee outlining disciplined capital management and a sustainable embedded insurance model.

Strait of Hormuz closure: A potential chokepoint for the Southeast Asian tech startup ecosystem: Skyrocketing energy prices might erode profit margins for startups in SEA hubs, where data centres guzzle electricity for AI training and cloud computing.

From bridge rounds to global awards: Startups across Asia keep building: Asian startups are securing funding, partnerships, awards, and leadership hires — but visibility drives credibility. On e27, dynamic profiles turn milestones into lasting signals of traction, momentum, and growth.

INTERNATIONAL

Amazon to invest US$50B in OpenAI, expands cloud deal: They’re expanding their multi-year contract from US$38B to US$100B over eight years, and OpenAI will secure about 2 gigawatts of Trainium capacity on AWS. They’ll work to customise OpenAI models specifically for Amazon developers to integrate more deeply with AWS infrastructure.

Amazon to invest US$9B in Korea, eyes public sector AI push: As part of the deal, it plans to invest over US$5 billion by 2031 in data centres. Ham Kee-ho, AWS Korea’s managing director, predicts that AI-native models could allow Korean startups with fewer than 12 employees to generate US$100B in annual revenue by 2029.

China unveils first national standards for humanoid robots: The system is built on six pillars – foundational and common standards, neuromorphic and intelligent computing, limbs and components, full-system integration, application, and safety and ethics.

ChatGPT uninstalls surged by 295% after the deal with the US gov: CEO Last week, the AI giant reached an agreement allowing the Department of Defense to use its AI models in the department’s classified network. This follows a high-profile standoff between the DoD and OpenAI’s rival Anthropic.

Pine Labs to launch stablecoin prepaid cards in 9 countries: The prepaid card will be funded from consumers’ digital wallets with stablecoins and will convert them into local currencies at the point of sale in real time. It does not plan to launch the product in India or China.

CYBERSECURITY

‘Cybersecurity must move at the speed of AI development’: ArmourZero CEO: Tho Kit Hoong warns AI-driven development is accelerating application and supply chain risks, urging enterprises to adopt real-time, contextual risk intelligence to secure modern software ecosystems across Southeast Asia and beyond.

Beyond the audit: Why risk management is the secret engine of 2026 growth: In a permacrisis economy, risk management is no longer compliance theatre but a growth engine. Founders must tackle non-linear scaling, correlated expansion risks, and AI obsolescence to build resilient, sustainable startups.

Cyber risk is moving upstream but we’re still defending downstream: Cyber defence once focused on firewalls and incident response. Today, risk forms upstream—in architecture, code, identity, and supply chains—long before attacks occur, demanding security by design, not detection alone.

Cybersecurity is not an IT problem: It is a trust architecture crisis: Southeast Asia’s digital economy risks fragility as cybersecurity remains siloed. Trust depends on empowered people, coordinated governance, and resilient infrastructure — not perimeter defence — to sustain confidence at scale.

SEMICONDUCTOR

China chip firms post big 2025 gains as Nvidia H200 blocked: Cambricon, Moore Threads, and MetaX posted large revenue rises for 2025 as domestic demand for semiconductors climbed amid Beijing’s tech-sufficiency push. Cambricon posted its first full-year profit of US$306.02M in 2025, while revenue jumped 450% to US$947.23M.

India chip market to hit US$103B as AI, EV drive demand: Domestic semiconductor market could nearly double to US$103B by 2030 from US$52B in 2024. The next 18 months are key as the ecosystem transitions from policy announcements to operational readiness.

US may cap Nvidia H200 exports to Chinese firms at 75,000 units: Bloomberg reported that Alibaba and ByteDance privately sought orders larger than 75,000 units, which per-customer limits would restrict. The US rules would restrict Chinese companies from using H200s to build data centres overseas.

AI

71% of APAC firms see AI as top data security risk: report: Only 35% of Asia Pacific firms know where all their data resides, 40% can fully classify it, and 47% of sensitive cloud data remains unencrypted. Credential theft leads attacks on cloud management infrastructure in Asia Pacific at 69%, compared to 67% globally.

Singapore targets 100K workers in AI upskilling programme: The city-state last month unveiled a US$122B budget to help it navigate a “more dangerous world,” marked by technological challenges and trade fractures. The PM has made AI a key plank in his economic strategy and pledged there’ll be “no jobless growth” amid this push.

When AI stops being a feature and starts being infrastructure: As AI quietly shifts from experiment to infrastructure, reliability, governance, and accountability matter more than novelty. Once embedded in workflows, AI demands ownership, maintenance, and resilience — because business-critical systems cannot afford failure.

Nvidia, telecom firms partner to develop AI-powered 6G networks: Nvidia said the members will commit to building 6G networks using software-controlled radios running on general-purpose computers that use AI to help direct radio traffic safely and efficiently.

THOUGHT LEADERSHIP

Extreme fear grips crypto: What 15 Fear Index reading means for your portfolio: US–Israel strikes on Iran triggered a global risk-off selloff, with crypto plunging alongside equities as correlation hit 78%, extreme fear spiked, leverage unwound, and oil-driven inflation fears clouded near-term recovery prospects.

How to use Bayesian thinking to pick a winning startups investment: Startup investing isn’t about bold vision, but Bayesian humility—updating beliefs with evidence, not hype. In a game of frequent failure, disciplined probability beats conviction, protecting investors from bias, overconfidence, and self-deception.

Why I’m trading bytes for atoms: The 65-year-old investor breaking the climate tech silos: After 33 years building and backing startups, this founder is “refiring” into FOAK climate-tech and nature-based solutions, launching Sherpa Alpha to back asset-heavy ventures bridging innovation, capital, and infrastructure gaps.

The investor data room: The final battlefield for trust: Startup deals often stall at due diligence: investors scrutinise data rooms, governance, and documentation. Institutional capital flows to ventures proving operational discipline, transparency, and readiness — not just compelling growth narratives.

The hidden risk most Founders don’t plan for: When everything looks “fine”: Success can mask structural risk. Overreliance on key people and weak systems stay invisible until disruption hits — and by then, damage is costly. Founders must diversify keyman risk early, while calm prevails.

Why perfect carbon audits could cripple climate finance — and what to fix instead: Blaming auditors for junk carbon credits misses the real problem: bloated methodologies. Lean baselines, satellite MRV, and smarter oversight can protect integrity without pricing developing-country climate projects out.

Transition climate risk: Navigating the future of sustainable real estate: As economies decarbonise, real estate faces rising transition risks from regulation, market repricing, technology shifts, and reputational pressure, demanding proactive decarbonisation, green financing, and data-driven resilience strategies.

Throwaway gold: How data can tap into the unrealised potential in plastic waste: Amid a worsening plastic crisis, poor data blocks a US$120B recycling opportunity. Better insights, inclusive systems, and circular innovation could transform waste into economic and environmental value.

The shifting geopolitics of sustainability, energy, and climate: Policy shifts in the EU, UK, and Canada reveal sustainability, energy security, and geopolitics converging — reshaping ESG standards, supply chains, Indigenous governance, and strategic business decision-making worldwide.

Can Bitcoin help us in the fight against climate change?: Bitcoin’s energy use draws criticism, but experts argue it consumes less than traditional banking and increasingly relies on renewables, potentially incentivising clean energy adoption while reshaping ESG and institutional investment conversations.

Revolutionise your business operations: A smarter alternative to lengthy paper processes: Paper-based processes slow businesses and cost up to US$36 per agreement. E-signatures boost productivity, enhance security, reduce costs, and cut environmental impact.

Why investors and customers are betting on ESG-aligned startups: Embedding ESG early helps startups accelerate growth, attract investors and talent, manage risk, drive innovation, and build resilient, purpose-driven businesses aligned with evolving market, regulatory, and consumer expectations.

The post Ecosystem Roundup: S Korea’s US$300M AI fund in Singapore; Amazon’s US$50B OpenAI bet; Cybersecurity races AI in APAC appeared first on e27.

The post Southeast Asia’s rare-earth moment: How China’s export controls could spark the region’s next industrial leap appeared first on e27.