When enterprises plan expansion across the Asia Pacific, connectivity is rarely treated as a strategic decision. It is assumed to be available, stable, and good enough to support daily operations. That assumption is increasingly costly.

Recent regional research shows that more than 30 per cent of organisations in APAC say inadequate network connectivity is actively threatening their growth plans. Nearly 44 per cent report that network limitations are restricting their ability to scale core initiatives such as cloud, data, and AI deployments. These are not edge cases. They signal that connectivity has shifted from a background utility to a critical operational dependency.

Yet many enterprises only realise this after expansion is underway.

Why APAC exposes the problem earlier

APAC is one of the most operationally fragmented regions in the world. Network quality, carrier behaviour, roaming performance, and access reliability vary widely across countries and, in some cases, within cities. For enterprises operating across multiple markets, this creates conditions where assumptions are tested immediately. 

Consider a regional enterprise team rolling out a new workflow across Singapore, Indonesia, and Thailand. On paper, the process is identical. The tools are the same. The timelines are aligned. On day one, however, execution begins to diverge. Team members in one market access systems without issue, while others experience intermittent connectivity, delayed authentication, or partial access to critical tools. Work still gets done, but not in the same way or at the same speed.

Connectivity is exercised from the first moment of execution. Employees land and need access. Systems activate. Customers interact in real time. When access behaves differently than expected, teams adapt quickly to keep work moving. 

What makes this challenging is not severity, but frequency. These are rarely full outages. They are short interruptions, inconsistent performance, or partial access that do not justify escalation on their own. Over time, teams build informal workarounds. Meetings start later. Tasks are deferred. Processes vary by market. Execution appears intact, but consistency quietly erodes.

This pattern is not anecdotal. It is increasingly visible in enterprise data.

Also Read: How eSIM can cut costs, boost CX, and simplify global operations for APAC startups

Quiet failure is a documented enterprise risk

Industry research consistently shows that network and connectivity issues account for a significant share of unplanned IT incidents, with the average cost of downtime running into thousands of dollars per minute for large organisations. In APAC specifically, more than half of surveyed enterprises report multi-million-dollar revenue losses linked to network outages or poor connectivity performance. 

However, the more common impact is not headline-grabbing outages. It is operational drag. 

The risk for enterprises is that this drag rarely appears in formal reporting. Quiet failure does not trigger outage alerts or incident reviews. Performance metrics often remain within acceptable thresholds even as execution consistency degrades across markets. What looks like normal variance at a regional level is often the cumulative effect of access issues that were never designed for or owned. 

This is why connectivity issues are often detected late. By the time leadership sees inconsistent performance across markets or slower decision cycles, the behaviour has already normalised, making the root cause harder to identify and correct.

The planning gap that enterprises underestimate

Most enterprise planning frameworks assume that connectivity will be broadly consistent across markets. Risk assessments focus on regulation, supply chains, talent, and cost. Access is treated as an environmental constant, even though tools such as eSIM already exist to manage connectivity variability more deliberately across regions.

APAC challenges that assumption. Fragmentation means connectivity behaves as a variable, not a given. When this variable is not explicitly accounted for, execution gaps appear early and compound quietly. 

This is not a technology problem. It is a planning problem. 

Enterprises design processes that depend on continuous access without stress-testing how those processes behave under uneven conditions. When access falters, execution does not fail loudly. It bends. 

Also Read: The impact of eSIM on international roaming and travel

Why connectivity shapes execution before other factors 

Other expansion challenges emerge gradually. Pricing models can be adjusted. Regulatory gaps surface over time. Localisation issues are identified through feedback cycles. Connectivity does not offer that margin for correction.

Connectivity is immediate. There is no grace period. Every workflow depends on it from day one. 

Because of this, connectivity becomes the first operational dependency to reveal flawed assumptions. In APAC, that revelation happens faster due to fragmentation. In more uniform regions, it may take longer, but the underlying dynamic is the same. 

What APAC teaches global enterprises 

APAC does not create this risk. It exposes it. 

As enterprises become more distributed and mobile, execution increasingly depends on continuous access across locations, devices, and teams. Connectivity is no longer peripheral to operations. It shapes how work flows, how decisions are made, and how consistently processes perform across markets. 

The takeaway for enterprise leaders is not about adopting a specific technology. It is about recognising connectivity as an operational variable that must be designed for, not assumed away. In practice, this is why enterprises are increasingly looking at approaches such as eSIM, not as a travel feature, but as a way to introduce greater predictability into how connectivity behaves across markets.

In APAC, businesses feel the impact first because the region accelerates the gap between planning and reality. Enterprises that acknowledge this early avoid quiet failure, becoming standard practice. Those who do not often discover the issue later, when inconsistency has already taken root.

—

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.

Featured image courtesy: Canva

The post APAC is not a single market, and connectivity is where most businesses feel the impact first appeared first on e27.

Sea Limited is deepening its relationship with Google, signing a memorandum of understanding to advance AI across Shopee (commerce), Garena (gaming), and Monee (digital financial services).

While the announcement reads like a standard partnership update, the specific areas named (agentic commerce, agentic payments, and AI-assisted game operations) signal something more pointed: both companies want AI systems that don’t just recommend, but act.

That distinction matters. Most consumer AI in 2024-2025 was about chat and content. “Agentic” systems are about execution: software that can navigate interfaces, compare options, apply rules, and complete transactions with minimal human input. Done well, it removes friction.

Poorly done, it becomes an expensive layer of confusion, or worse, a security liability.

Also Read: Agentic AI is powerful – but power isn’t product-market fit

This expanded partnership builds on existing Sea-Google collaborations, such as the YouTube Shopping Affiliate Program with Shopee and Free Fire League on Google Play with Garena. The new element is explicit: Sea wants to operationalise AI at scale across its ecosystem, and Google wants distribution in some of the world’s most mobile-first markets.

Forrest Li, Sea’s Chairman and CEO, framed it as the next platform shift:
“At Sea, we have always believed in the fundamental power of technology to improve lives and create long-lasting value for the communities we serve.

AI is the next big technology revolution, and we believe that it has huge potential to positively transform our business and create value in our markets. This partnership with Google on AI will drive innovation in the business application of the technology at scale, and allow us to make AI more accessible to the digitally underserved in our markets,” Li added.

Google APAC President Sanjay Gupta struck a similar tone. “By combining Google’s AI leadership with Sea’s innovative ecosystem, we’re building products that don’t just solve today’s challenges but define the future of gaming, commerce, and financial services. And we’re developing these solutions responsibly, with user privacy and safety at the core. Together, we are accelerating the adoption of this transformative technology and unlocking the immense economic potential of Southeast Asia’s digital landscape.”

The real story is what “agentic” implies for shoppers, gamers, and people who still sit on the edge of formal finance.

1) AI-powered agents and the future of online shopping: from browsing to delegation

Shopee and Google say they will “jointly explore the building of an AI agentic shopping prototype” that “can seamlessly integrate across Shopee and Google platforms”.

If that prototype becomes a product, it could shift online shopping from search-and-scroll to goal-driven delegation. Instead of a user typing keywords, filtering, opening ten tabs, checking delivery dates, and messaging sellers, an agent could:

• Translate a vague intent (“cheap, reliable phone for my mum”) into specific constraints
• Compare sellers, shipping times, warranty terms, and return policies
• Watch prices over time, alert on drops, and execute purchases within a budget
• Bundle items to optimise shipping or apply the proper vouchers automatically
• Handle post-purchase steps: tracking, rescheduling delivery, filing returns

Southeast Asia is an unusually fertile market for this because commerce is already messy in the real world: multiple languages, informal sellers, heavy promotion mechanics, and a wide range of logistics reliability. An agent that can actually navigate those trade-offs could become the new front door to shopping.

Also Read: Why agentic AI isn’t what the hype suggests

But it also raises uncomfortable questions. Who does the agent really serve — buyer, seller, or platform margin? If an AI agent becomes the shopping interface, then ranking, sponsored placements, and “recommended” choices become even more consequential. Platforms will need to show that agents are not simply optimised to maximise take rate while wearing a friendly chatbot mask.

2) AI innovation in gaming: not just smarter NPCs, but faster live ops and globalisation

On the gaming side, Garena and Google are looking to use Google’s AI solutions to “enhance gamer experiences” and “transform the productivity of game development and operations”, with a line about early access pilots for Google’s latest AI research.

The obvious consumer-facing play is richer worlds: better non-player characters, more adaptive matchmaking, personalised onboarding, and dynamic content. The less glamorous—but likely more valuable—angle is operations:

• Faster content production (assets, localisation, event scripting) to keep live-service titles fresh
• Better moderation and trust-and-safety tooling for voice and chat, where toxicity kills communities
• Anti-cheat systems that can detect novel patterns rather than just known signatures
• Smarter A/B testing loops that tune difficulty and retention without breaking fairness

If Garena can shorten content cycles and improve trust and safety, it can scale globally with less operational drag. That matters because “global gaming experience” is often less about graphics and more about whether a game feels fair, stable, and culturally native in Bangkok, Manila, São Paulo, and Riyadh at the same time.

There’s a catch: generative tooling can also turbocharge bad actors—cheat creation, scams, and automated harassment. Any AI advantage in gaming will be matched by AI-powered abuse, and publishers will have to budget for that arms race.

3) AI and financial inclusion: fewer forms, more approvals, but also new kinds of exclusion

Sea’s financial arm, Monee, will work with Google on an “open, shared Agent Payments Protocol (AP2)”, where Monee will provide feedback to ensure it is “robust, secure, and suitable” for Southeast Asia, with an intention to later explore pilot experiences across platforms.

If AP2 evolves into something widely adopted, it could reduce one of the biggest blockers to financial inclusion: complexity. Many underbanked users don’t struggle with the idea of digital money; they struggle with onboarding steps, confusing user interfaces, and customer support that doesn’t speak their language or understand their context.

AI could help by:

• Turning onboarding into a guided, multilingual flow that adapts to user capability
• Automating dispute handling and customer service at lower cost
• Improving fraud detection to protect first-time users (who are prime scam targets)
• Enabling small merchants to accept digital payments and reconcile accounts without accounting expertise

For SMEs, inclusion is not philosophical; it is operational. If agents can reconcile transactions, chase invoices, or manage cashflow nudges, that’s not “AI magic”; it is time returned to a shop owner.

Still, AI-driven finance comes with a risk the industry often underplays: automated denial. Models can quietly exclude people with thin files, unstable device histories, or non-standard income patterns; the exact users inclusion efforts claim to prioritise. Any “agentic payments” system that touches identity, fraud, or credit will need strong controls, auditability, and clear recourse when automation gets it wrong.

Also Read: Agentic AI: The next frontier in technology

Sea and Google are calling this partnership “strategic”. The test will be whether these projects become everyday tools that work in the region’s real conditions: inconsistent connectivity, diverse languages, scam-heavy environments, and users who will not tolerate extra steps just because the system is “smart”. Agentic AI only wins if it makes life simpler — and doesn’t create a new category of problems that humans then have to clean up.

The post Shopee, Garena, Monee: Sea’s AI ambition gets serious appeared first on e27.

As Ramadan retail continues to evolve across Southeast Asia, brands are being urged to rethink how they plan, activate and optimise campaigns during the holy month. New insights from Criteo’s analysis of Ramadan 2025 reveal that shoppers are starting earlier, taking longer to decide and converting closer to peak festive moments–a behavioural shift that will intensify as Ramadan and Chinese New Year converge in 2026.

Retail sales across Southeast Asia rose 13 per cent year-on-year during Ramadan 2025, underlining the growing commercial significance of the season. Yet the headline growth masks a deeper transformation in consumer behaviour. For purchases made in the final two weeks of Ramadan, the average time between a shopper’s first product visit and completed purchase stretched to 19 days, with some journeys extending beyond 50 days.

Early discovery, however, did not eliminate late conversion. Indonesia recorded a 35 per cent uplift in retail sales during the last two weeks of Ramadan, peaking at 57 per cent on March 16. Malaysia saw sales climb 26 per cent, with a 52 per cent peak on March 23. Singapore’s sales pattern was more stable, reflecting its diversified retail calendar.

The takeaway for Ramadan retail strategies is clear: shoppers are browsing earlier but still buying closer to Eid.

“Ramadan 2025 underscored a fundamental shift in how consumers plan and purchase–discovery is happening earlier, and shopping journeys are becoming increasingly fluid across channels,” said Sukesh Singh, managing director, Southeast Asia at Criteo.

“As festive moments begin to overlap, this behaviour will only accelerate. At Criteo, our AI-powered intelligence helps brands identify the right audiences at the right time and place, adding relevance across touchpoints and strengthening full-funnel, cross-channel outcomes. Brands that anticipate demand and stay relevant across the entire journey will be far better positioned to earn attention that drives higher conversion.”

Also Read: AI shopping adoption surges 39 per cent in APAC, fueling retail tech investments

Preparing for a more compressed festive calendar

With Chinese New Year and Ramadan set to fall in the same week in 2026, businesses face a tighter and more competitive festive window. Criteo’s advice to brands centres on five strategic shifts:

Plan Ramadan as a multi-stage season

Ramadan retail can no longer be treated as a short, promotion-led sprint. With discovery beginning weeks ahead of purchase while conversion clusters around peak moments, brands must structure campaigns in phases. Early weeks should focus on awareness and consideration, capturing shoppers during research and comparison. As Eid approaches, messaging should pivot towards urgency, promotions and conversion-led tactics.

Prepare for demand compression

The convergence of major cultural moments is likely to compress demand into shorter timeframes. Shorter decision windows and heightened competition mean brands must be ready for sharper spikes in traffic and transactions. Budgets, inventory and activation plans should be flexible enough to scale quickly during high-intent moments, rather than being distributed evenly across the month.

Align with cultural and daily rhythms

Ramadan retail activity closely follows daily routines. While afternoons generate the highest overall sales, the largest uplift compared with pre-Ramadan levels occurs during Suhoor. In Indonesia, this surge is most pronounced between 3 AM and 5 AM, while in Malaysia it shifts later, between 4 AM and 7 AM. Campaign timing, creative and offers that reflect these culturally relevant windows can significantly improve engagement and conversion.

Also Read: Multimodal AI: Reshaping search and discovery in retail and travel

Design for non-linear purchase journeys

Ramadan 2025 demonstrated that there is no single path to purchase. Some consumers act quickly on high intent, while others deliberate for weeks. Brands must maintain visibility across multiple touchpoints, from early discovery to final checkout, adapting messaging as intent strengthens. Retail media becomes particularly valuable closer to purchase moments, where intent signals are clearer and performance outcomes are easier to measure.

Rely on data-led optimisation and automation

As festive calendars grow more crowded, static campaign plans struggle to keep pace. Data-driven optimisation and automation enable brands to anticipate demand peaks, detect emerging intent signals and adjust spend, messaging and targeting in real time. This shift towards more adaptive, AI-supported execution allows Ramadan retail campaigns to move from reactive planning to continuous optimisation.

For businesses across Southeast Asia, the message is unequivocal. Ramadan retail is expanding in scale and complexity. Success will not hinge solely on promotional intensity in the final days before Eid, but on sustained relevance throughout a longer, more fluid shopper journey. Brands that anticipate demand, respect cultural rhythms and harness data intelligently will be best placed to win attention — and sales — in the seasons ahead.

—

Image Credit: Rauf Alvi on Unsplash

The post How brands can win Ramadan retail sales as consumer journeys grow longer appeared first on e27.

At the start of 2025, speed felt like the only answer. Faster growth. Faster launches. Faster decisions. Like many founders, I believed momentum was something you either kept feeding or risked losing entirely. In hindsight, that belief shaped several decisions I wouldn’t make again today.

One of the biggest was expanding capacity before demand had fully stabilised. We hired ahead of confirmed pipelines, stretched teams across regions, and said yes to opportunities that looked good on paper but came with hidden complexity. At the time, it felt like the responsible thing to do was to prepare for growth before it arrived. But what I didn’t fully appreciate was how fragile early momentum can be when it isn’t supported by repeatable systems.

The cost wasn’t just financial. It showed up in scattered focus, stretched leadership bandwidth, and teams that were busy but not always effective. Growth happened, but it wasn’t clean. And some of it didn’t age well.

Another lesson came from the metrics we chose to celebrate. In 2025, we tracked volume obsessively, such as the number of leads, the number of projects, and the number of touchpoints. It felt reassuring to see activity increase. But over time, I realised we were mistaking motion for progress. We were moving fast, but not always in the right direction.

What we didn’t pay enough attention to were quieter indicators: operational strain, client fit, team energy, and the effort required to maintain momentum. Those didn’t show up neatly on dashboards, but they told a much more accurate story about sustainability. By the time we acknowledged them, some damage had already been done.

Also Read: Why Asian startups should focus on Southeast Asia in 2026

As budgets tightened towards the end of the year, reality forced a reset. We had to get sharper about how we spent, who we hired, and what we built. That constraint, uncomfortable as it was, became a turning point.

We stopped defaulting to headcount as a solution. Instead of asking who else we needed, we asked what we could simplify. We questioned whether a feature actually solved a problem or just made us feel innovative. We rewrote processes that had grown bloated under pressure and slowly, clarity returned.

Slowing down revealed things that rapid scaling had hidden. It exposed inefficiencies we had been compensating for with effort. It highlighted roles that lacked clear ownership. It showed where culture had been diluted by speed rather than strengthened by intention.

Most importantly, it reminded me that growth without coherence is not progress, but it is postponement. You can delay reckoning by moving fast, but eventually the business asks harder questions.

Going into 2026, my priorities look very different. I’m less interested in how quickly something can scale and more interested in whether it can be repeated without exhaustion. I think more about resilience than reach. I ask whether a decision gives us optionality or locks us into constant acceleration.

Also Read: Why the tech world is heading to Hong Kong in April 2026

If I were optimising purely for sustainability now, I’d do a few things differently. I’d build fewer things, but build them properly. I’d hire later, but onboard better. I’d choose clients and partners more carefully, even if it meant slower revenue in the short term. And I’d pay closer attention to the signals that don’t scream for attention because those are often the ones that matter most.

2025 wasn’t a failure. It was a necessary stress test. It showed me where ambition had outpaced structure and where optimism had overridden discipline. But it also clarified what kind of founder I want to be going forward.

I don’t want to build businesses that only work when everything goes right. I want to build ones that can withstand uncertainty, fatigue, and change. Ones that leave room for people to think, not just react. Ones that grow because they’re solid, not because they’re sprinting.

The market has matured. Expectations have shifted. And maybe that’s a good thing. Because sustainability isn’t about doing less, but it’s about doing what matters, for longer.

If there’s one question I’m carrying into 2026, it’s this. If growth stopped tomorrow, would what we’ve built still be worth sustaining?

That answer matters more to me now than any headline number ever did.

—

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.

Featured image courtesy: Canva

The post Resetting for sustainable growth: What 2025 taught me about building businesses that last appeared first on e27.

Every founder knows this moment. You log into your marketing dashboard, and it feels like you’re staring into a maze. A maze built with your own tools. The marketing technology landscape keeps expanding, yet most teams still struggle to turn those tools into real growth.

In my years working with startups and scaling marketing engines, I’ve seen the same pattern again and again. Small teams drown in tool sprawl, spend more time managing dashboards than driving demand, and fail to build the visibility they need to scale. Much of what they pay for simply goes unused because the stack is fragmented and disconnected.

For early-stage startups, this isn’t just inefficiency. It is a lost runway. Every hour spent toggling between tools is an hour not spent creating meaningful customer experiences or validating product-market fit. In this article, I will break down what a lean AI marketing stack should look like, how to build it first, and why fewer, smarter systems consistently outperform bloated setups when every resource counts.

What lean AI marketing really means for early teams

Lean AI marketing starts with a simple shift in mindset. Early teams do not need more tools or heavier automation. They need a system that helps a small group execute meaningful work consistently without operational drag.

At this stage, marketing responsibilities are straightforward but demanding. Founders and small teams are expected to wear multiple hats every day. The focus should stay on the few activities that directly influence growth:

• Understand what customers are searching for, asking, and comparing
• Create useful, trustworthy content that answers those needs
• Show up where buyers discover solutions, across search and AI-driven channels
• Distribute consistently without manual repetition
• Measure the signals that connect marketing efforts to revenue

AI works best when it supports these fundamentals quietly in the background. The right stack reduces repetitive tasks, connects workflows, and keeps research, content, and visibility moving together as one system.

With that foundation in place, marketing feels lighter, faster, and more predictable. And for startups, the team that ships consistently is usually the team that pulls ahead.

Also Read: Is your business stuck in manual mode? It’s time to automate with AI

The five core jobs every startup must solve first

Once you strip away the noise, startup marketing comes down to a handful of repeatable jobs. Get these right, and growth compounds. Miss them, and no tool stack can compensate.

• Customer research: Identify what your audience is searching, comparing, and struggling with so messaging aligns with real demand.
• Content creation: Publish helpful, high-intent content that answers questions and builds trust at every stage of the buyer journey.
• Visibility across search and AI discovery: Ensure your brand appears consistently in Google results, AI answers, and emerging generative engines where decisions are increasingly shaped.
• Distribution and repurposing: Extend the life of every asset across channels without recreating work from scratch.
• Measurement and optimisation: Track what influences pipeline and revenue, so effort flows toward what actually drives growth.

Everything in a lean AI stack should support these five jobs. If a tool doesn’t make one of them faster or easier, it’s likely adding noise.

The lean AI marketing stack blueprint

Once these five jobs are clear, the stack becomes easier to design. Instead of collecting tools randomly, map each tool to a specific outcome. Every layer should remove manual effort and help a small team move faster with fewer handoffs.

Many early teams try to solve each row with a separate tool. Over time, that creates fragmented workflows and rising costs. Increasingly, startups are consolidating these functions into unified AI platforms that handle multiple jobs in one place, keeping the stack lean and easier to manage.

Also Read: AI is making wealth management feel like concierge service

How should startups build a lean AI marketing stack step by step?

A lean stack works best when built in layers. Trying to set up everything at once usually leads to tool overload, scattered workflows, and stalled execution. A phased approach keeps the team focused and shows results faster.

• Step 1: Start with the customer and search insight. Understand what your audience is actively searching, comparing, and asking. Ground every decision in real demand so your content has direction from day one.
• Step 2: Build a consistent content engine. Set up AI-assisted workflows to draft, optimise, and publish regularly. Consistency creates momentum and compounds visibility over time.
• Step 3: Optimise for discovery. Structure content for both search engines and AI-driven answers. Strong visibility reduces dependence on paid acquisition.
• Step 4: Automate distribution. Repurpose each asset into multiple formats and channels so one piece of work delivers wider reach.
• Step 5: Measure and refine continuously. Track what drives traffic, leads, and pipeline. Reinvest in what performs and eliminate what doesn’t.

Done in this order, marketing stays manageable, measurable, and scalable for even the smallest teams.

Common mistakes to avoid when building your AI marketing stack

Even strong teams lose momentum when the stack grows faster than their strategy. A few early missteps can quietly drain time, budget, and focus.

• Adding tools before defining outcomes: Software should support a clear job. Without that clarity, dashboards multiply, but results don’t.
• Chasing every new AI trend: Not every feature needs adoption. Stability and consistency usually outperform constant experimentation.
• Publishing without a visibility plan: Content that isn’t optimised for search or AI discovery rarely gets seen, no matter how well written it is.
• Working in disconnected systems: Copying data between platforms slows execution and creates avoidable errors.
• Measuring vanity metrics: Traffic and impressions mean little if they don’t translate into leads or pipeline.

A lean stack stays focused, simple, and tied directly to growth.

Build for focus, not complexity

Early-stage startups don’t win with bigger stacks. They win with clearer priorities and faster execution.

When customer insight, content, visibility, and measurement work together smoothly, marketing stops feeling chaotic and starts feeling predictable. Progress compounds. Teams ship more. Decisions get easier.

AI should support that rhythm quietly in the background, reducing manual effort and freeing time for higher-impact work. Keep the system simple. Keep the stack lean. Focus on what directly drives growth.

Because at this stage, clarity and consistency beat complexity every time.

—

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.

Featured image courtesy: Canva

The post The lean AI marketing stack every startup should build first appeared first on e27.