Small and mid-sized enterprises (SMEs) across Southeast Asia are entering one of their toughest hiring cycles in recent years. LinkedIn’s latest research suggests the talent crunch is becoming structural rather than cyclical. Nearly three in four Singapore-based SMEs say it has become harder to find qualified talent compared to last year.

The challenge goes beyond a shortage of candidates. Businesses are facing a widening skills mismatch, intensified competition for in-demand capabilities, and a surge of AI-generated job applications that add noise to hiring pipelines and increase screening workloads.

Also Read: How SMEs can become learning organisations, without the corporate bureaucracy

At the same time, AI is reshaping what companies look for in talent: from technical expertise to broader AI literacy.

In the first part of this interview, Elsie Ng, Director of Talent Solutions for Singapore and Malaysia at LinkedIn, shares insights on how SMEs can adapt to the evolving talent landscape.

Edited excerpts:

Singapore’s SME talent challenges are increasingly seen as “structural”, per LinkedIn data showing 71 per cent of respondents reporting greater hiring difficulty than last year. Beyond global trends, what regional factors may be contributing to more persistent talent constraints for startups?

LinkedIn data shows 71 per cent of hirers in small businesses say it’s harder to find qualified talent, yet 58 per cent of professionals report actively job hunting. This tells us that the labour market is still moving, i.e. people are looking and businesses are hiring, but the alignment isn’t landing.

To understand why, we need to look at the broader context. We’re seeing the labour market rotate toward a new era of work. In Singapore, hiring has slowed to about 20 per cent below pre-pandemic levels, shaped largely by economic uncertainty and monetary policy shifts. But there are pockets of opportunities, driven by AI.

We’re entering what I’d call a “new collar” era of work, one where the workforce increasingly blends knowledge work, advanced technical skills, and distinctly human strengths. AI is at the centre of this shift.

In Singapore, AI engineering roles now make up 4.2 per cent of all job postings on LinkedIn, up 40 per cent year-on-year, while AI engineering talent represents just 1.5 per cent of our member base and is growing at only 10 per cent annually. Demand is outstripping supply by a significant margin.

But it’s not just about engineering. Demand for AI literacy skills has surged over 70 per cent year-on-year and is now spreading into traditionally non-technical roles like marketing. AI is becoming a baseline expectation across the organisation, not just within technical teams.

As AI literacy becomes table stakes, human capabilities are gaining even more prominence. In Singapore, soft skills like communication, teamwork, leadership, and problem-solving are among the top 10 in-demand skills.

We also know that small businesses are growing and still hiring, albeit at a slower pace. SMEs grew 4.97 per cent in company numbers and 3.56 per cent in headcount year-on-year in October 2025, outpacing large enterprises.

And while hiring overall is down, large enterprises are driving the decline more sharply — down 42 per cent compared to 26 per cent for small businesses. The real challenge is that small businesses are hiring into a fundamentally different labour market.

Also Read: Talents remain an issue in AI proliferation, but here are 6 steps that businesses can do to tackle it

In this new era of work, skills matter more than titles, and many traditional hiring approaches haven’t kept pace with how quickly that’s changing.

For Singapore specifically, a few regional dynamics are adding to the structural challenge. Singapore’s position as a regional tech hub, combined with strong government support for AI adoption, creates significant momentum and opportunity, but also intensifies competition for in-demand skills.

Competition for in-demand skills tops SME pain points at 44 per cent. Which specific tech and AI roles/skills are SMEs in Singapore, struggling the most to fill?

The most acute gaps for small businesses lie in AI engineering and AI literacy skills.

Today, 7.7 per cent of employees in SMEs have AI engineering skills, compared to 20 per cent in large enterprises. Put simply, small businesses are operating at roughly one-third of the AI capacity of larger companies, which limits their ability to build, deploy, and scale AI solutions internally.

The gap is even wider for AI literacy, the foundational ability to understand and work effectively with AI tools. Over the past year, AI literacy in SMEs has grown five times slower than in larger enterprises. As AI spreads across industries and roles, this gap risks compounding over time, with real implications for competitiveness, productivity, and long-term resilience. If left unaddressed, the growing gap will widen existing inequalities in access to technology and opportunity.

But there’s a critical counterpoint: employees in small businesses are highly motivated to learn. Nearly half (49 per cent) are learning AI with employer-provided guidance or training. What’s more telling is the initiative they’re taking independently: 67 per cent are learning on their own time using free resources, and 53 per cent are paying for courses themselves.

When it comes to how they prefer to learn, the pattern is clear: employees want practical, hands-on experience. The top three preferences are learning through real-life projects and assignments (35 per cent), using AI tools to practice real scenarios (34 per cent), and virtual training and tutorials (34 per cent). The demand for upskilling is there. The challenge for SMEs is creating the structure and opportunity to channel that motivation effectively.

35 per cent of SMEs cite a sheer lack of qualified applicants. What factors tend to draw candidate attention toward larger employers, and how does LinkedIn help SMEs and startups in Singapore and Malaysia improve visibility and reach candidates with the right skills?

Many candidates are drawn to larger employers because of structured learning opportunities, especially as AI reshapes roles and expectations. In an era of rapidly evolving skills, access to upskilling has become a key deciding factor.

The data is clear: professionals want support from management when navigating AI. Two-thirds (67 per cent) of employees at small businesses in Singapore believe access to lifelong learning resources would boost their confidence in adapting to AI changes, and 66 per cent are actively looking for helpful content (resources, tools, and courses) to learn AI better. More than half (55 per cent) want leadership support to navigate AI-related changes at work. And critically, 65 per cent believe they can successfully reskill in AI regardless of age, with the right support.

Also Read: How startups can overcome the AI talent death

This is where small businesses can compete. While they may not have the scale of large enterprises, small businesses can offer something equally valuable: direct access to hands-on learning, clearer pathways to applying new skills, and leadership that’s closer to the work.

AI-generated applications now plague 40 per cent of SME hiring pipelines, bloating workloads. How is this “noise” disproportionately hammering resource-strapped startups versus larger firms, and what’s the real cost in time and missed hires?

In Singapore, 40 per cent of recruiters say they feel pressure to hire faster, while the same proportion say uncovering hidden-gem candidates is a top priority. For small businesses with limited hiring resources, higher application volumes quickly turn into longer screening hours and slower decisions. Reducing noise and surfacing a genuine fit early can make the difference between moving forward with confidence and missing out on the right hire.

AI-powered tools like Hiring Pro are designed to bring more clarity to that process. Rather than relying heavily on keyword matches or credentials alone, it evaluates candidates against the actual skills and criteria a business sets, using real-time data to surface stronger-fit shortlists.

For small teams, having that kind of support, almost like a hiring partner that’s embedded in the workflow, helps shift time away from manual filtering and toward meaningful conversations with the right people.

With AI adoption exploding, why should SMEs bet on “talent resilience” through tools when upskilling their existing teams might be cheaper and faster than chasing unicorns in a tight market?

It’s not either-or; upskilling and talent resilience need to work in tandem.

Upskilling is essential. In fact, small business employees are already showing strong initiative — learning AI through on-the-job guidance and training, while also investing their own time and money to stay relevant.

What we’re seeing, however, is a widening gap between how quickly AI capabilities are advancing and how slowly organisational systems and workflows are adapting and evolving around them. In that environment, training alone doesn’t always translate to real impact.

About 43 per cent of small business employees say they feel overwhelmed integrating AI into their work, and more than half (53 per cent) feel they’re not using it to its fullest capability. That tells us the challenge isn’t just access to courses; it’s how AI needs to be embedded into day-to-day roles.

Talent resilience means redesigning how capability is built and deployed at the organisational level:

• Embedding AI into everyday workflows, not treating it as a side project
• Segmenting capability — deciding what to buy, what to build, and what to raise literacy on
• Rotating employees through AI-enabled projects to build judgment and domain expertise over time

Upskilling keeps people relevant. Talent resilience ensures the business itself can continuously adapt. Businesses that combine both will move from experimentation to real enterprise productivity gains.

The post Singapore SMEs must prepare for the ‘new collar’ workforce: LinkedIn’s Elsie Ng appeared first on e27.

The crypto market’s recent 0.67 per cent decline to a total capitalisation of US$2.29 trillion reflects more than routine volatility. It signals a decisive macro-driven repricing, with digital assets now moving in lockstep with traditional risk indicators. Over the past week, Bitcoin and the broader crypto complex have maintained a 64 per cent correlation with the S&P 500, a clear signal that rates-sensitive capital is treating crypto as part of the same risk bucket as equities. This is not a crypto-specific story. It is a story about liquidity, inflation expectations, and how geopolitical shocks transmit through every corner of the global financial system.

The primary catalyst for this selloff stems from a sharp spike in oil prices and a surging US dollar. Escalating Middle East tensions, including direct US–Iran conflict, pushed Brent crude above US$113.7 per barrel, its highest level since 2022. West Texas Intermediate followed, surging as much as 22 per cent to over US$111 a barrel at the open. Simultaneously, the US Dollar Index gained 0.6 per cent as investors fled to safety. This dual shock creates a powerful headwind for risk assets. Higher energy costs feed inflation expectations just as labour market data shows unexpected weakness, with 92,000 jobs lost in February. A stronger dollar tightens global liquidity conditions, making dollar-denominated assets more expensive for international holders and pressuring valuations across the board. Crypto, with its high beta and sensitivity to liquidity flows, feels this pressure acutely.

Bitcoin itself fell 2.03 per cent, contributing over half of the total decline in market cap. This move was not random. Large holders, often called whales, distributed coins they had recently accumulated, adding supply to an already nervous market. Spot Bitcoin ETFs saw net outflows, compounding the selling pressure. The Fear and Greed Index reading of 18, labeled Extreme Fear, confirms that sentiment has turned decisively negative. When sentiment reaches these extremes, technical levels gain outsized importance. Bitcoin now tests the US$66,000 to US$66,500 support zone. A sustained break below this range opens the path toward US$63,700. Bitcoin dominance holding above 58 per cent suggests capital is not rotating aggressively into altcoins, which typically underperform in risk-off environments. This concentration of weakness in Bitcoin, the market’s anchor, drags the entire ecosystem lower.

Also Read: While S&P 500 struggles, crypto’s low correlation to gold and stocks attracts institutional attention

The crypto selloff did not occur in isolation. Global markets moved in tandem, confirming the macro nature of the move. US equity futures plunged at the open, with Dow futures dropping over 800 points, roughly 1.8 per cent, and Nasdaq 100 futures sliding 1.9 per cent. Asian markets reflected similar stress, with the Nikkei 225 tumbling 6 per cent toward the 52,000 level, hitting an eight-week low amid Japan’s high dependence on Middle Eastern oil. Even gold, traditionally a safe haven, fell 1.4 per cent to US$5,099 an ounce in early spot trading, suggesting that liquidity needs are forcing investors to sell what they can, not just what they want to. This broad-based risk-off move underscores that crypto is no longer an island. It trades as part of a global macro tape, where oil, the dollar, and equity volatility set the tone.

Behind these price moves lie concrete geopolitical and economic fundamentals. Escalating hostilities involving Iran have effectively halted traffic through the Strait of Hormuz, a critical chokepoint for 20 per cent of global oil consumption. This disruption threatens to rekindle inflation fears just as central banks weigh their next moves. The market now prices in a 97 per cent chance that the Federal Reserve will hold interest rates steady at its March 18 meeting, with any potential cuts pushed back toward late 2026. This shift in expectations matters profoundly for crypto, which thrives in environments of easy money and declining real yields.

Adding to the uncertainty, corporate developments, such as BlackRock limiting withdrawals from its US$26 billion private credit fund, sparked contagion fears, causing its shares to tumble seven per cent. While Broadcom’s 4.8 per cent jump on bullish AI chip forecasts offered a rare bright spot, it was not enough to offset the broader risk aversion. Meanwhile, China’s decision to set its 2026 GDP growth target at 4.5 per cent to five per cent, the lowest in decades, signals ongoing deflationary pressures and trade tensions that further complicate the global outlook.

Also Read: Wallets, not smart contracts, were crypto’s biggest risk in 2025

Looking ahead, the near-term path for crypto hinges on two factors: oil price stability and the Federal Reserve’s tone on March 18. If energy markets calm and the Fed maintains a dovish stance despite inflationary pressures, crypto could find a floor near current levels. A sustained move above US$113 per barrel for oil would keep inflation expectations elevated, likely delaying rate cuts and maintaining pressure on risk assets.

Technically, Bitcoin’s ability to hold above US$66,000 remains the key level to watch. A decisive break below would likely trigger algorithmic selling and force leveraged positions to unwind, accelerating the move toward US$63,700. Traders should also monitor ETF flow data for signs of institutional accumulation or distribution, as these flows have become a reliable proxy for smart money sentiment in the current market structure.

This moment tests a core question for the crypto ecosystem: does it retain its narrative as an uncorrelated alternative asset, or has it matured into a risk-on instrument that trades with tech stocks and macro liquidity? Tell me about it. 

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

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

The post Crypto’s wake-up call: How a stronger dollar and US$113 oil are crushing risk assets appeared first on e27.

Kampanat “Jom” Vimolnoht strode off the Singapore FinTech Festival stage in late 2024 with the kind of polish that sells trust: tailored blazer, crisp black T‑shirt, urbane charm.

To an audience of founders and investors, he was a familiar archetype: a crypto‑savvy venture capitalist with a UK Master’s in Investment Analysis, a history of venture capital roles and government advisory work, and a new post at KXVC, the corporate VC arm of Thailand’s leading Kasikornbank that had announced a US$100 million Web3 and AI fund the year before.

Also Read: Thailand’s corporate capital era: How big business became the startup banker

Only it was a mirage. Over the following year, a string of investors across Bangkok, Singapore, Ho Chi Minh City, and California discovered that the allocations, contracts, and deals Jom sold them were, in many cases, fabricated. The tale — reconstructed from interviews, bank transfers, blockchain traces, and KXVC’s own public warning — is both ordinary and devastating: ordinary because it follows familiar fraud mechanics; devastating because it exploited social capital — the halo effect of reputation — to drain life savings and corrode trust in a nascent investment ecosystem.

How the con unfolded

In private crypto markets, allocations to pre‑launch token sales are valuable and opaque. These deals typically circulate in invite‑only channels (Telegram, WhatsApp, and private investor lists) and access is concentrated among founders, funds, and a few insider intermediaries.

Vimolnoht played the insider role convincingly. He supplied professional‑looking decks, allocation agreements, and payment instructions, and invited friends and colleagues into deals in projects he claimed to be connected with: Monad, Babylon, Linera, and others.

Victims presented a consistent pattern: small initial transfers, followed by larger sums as trust deepened. One Bangkok executive, “Mark”, said he invested “in total more than a million dollars”. Another victim from the US, “Steven”, who paid in USDC, believes he lost about US$130,000, his life savings in crypto. Scamurai’s reporting identified about two dozen alleged victims so far, with individual losses ranging from roughly US$20,000 to over US$1,000,000. A blockchain wallet linked to some of the flows shows about US$1.71 million moving through it between July and October 2025.

When vesting milestones approached, and tokens were supposed to unlock, the excuses began: administrative delays, counterparty issues, even that Vimolnoht himself had been scammed.

Also Read: Inside the dark economy of crypto scams: 2024’s most lucrative fraud tactics

Communications then stopped. Project founders contacted directly by investors denied any affiliation or said they had only spoken informally with him. KXVC posted a terse advisory: it never raises external funds and has “never authorised any individual to act on behalf of KXVC” to solicit investor transfers to personal accounts. The firm has confirmed Vimolnoht left the company in March 2025.

A regional phenomenon

Thailand’s scams are not unique. What makes this episode instructive is how it exposes systemic vulnerabilities present across Southeast Asia: close‑knit networks, the prestige economy of panels and advisory roles, and a high appetite for outsized returns combined with uneven due diligence.

“It does not feel like being deceived. It feels like being trusted with an opportunity,” observed Dr Pun‑Arj Chairatana, former executive director at Thailand’s National Innovation Agency, warning that such schemes are often structured around curated deal‑flow circles and private chat groups that carry an air of exclusivity.

Prominent local voices corroborate the wider pattern. Yod Chinsupakul, CEO of local e-commerce company LINE Wongnai, noted the region’s “halo effect” for charismatic figures and urged a culture of whistleblowing to surface wrongdoing.

He also pointed to several other suspected malfeasances in Thailand’s tech sector: an e‑commerce enabler that collapsed amid alleged CEO fraud and tragic consequences for employees; a loyalty‑points startup accused of unlimited minting that reportedly cost a strategic partner “hundreds of millions of baht” (roughly US$5-15 million); and payment companies that allegedly engaged with illicit betting websites and opaque loans with potential conflicts of interest.

Chinsupakul stressed that while the proportion of bad actors has fallen since the industry’s early, “fluffy” years, the remaining wrongdoing is severe and often hidden.

The contagion effect

Dr Pun‑Arj cautions that the damage extends beyond direct victims. Reputation (the currency of fundraisings and syndications) erodes swiftly. He points to the Silicon Valley Bank collapse in 2023, not as an analogue in causation but as a lesson in contagion: a single failure can chill capital across markets. For small funds and emerging managers in Thailand and across Southeast Asia, the reputational fallout from prominent scams risks hamstringing legitimate teams still building track records.

This reputational contagion has practical consequences. Limited Partners revisit commitments; fundraising conversations stall; partnerships are re‑evaluated. Restoring confidence is not a matter of statements, Dr Pun‑Arj said, but of “consistency of conduct over time: transparent reporting, governance that is substantive rather than performative, and sound judgement exercised even when no one is watching.”

Enforcement, verification and the limits of charisma

Scammers exploit a familiar mix: technical plausibility (token deals, private allocations), social proof (panel appearances, advisory roles), and operational friction (the difficulty of verifying vesting schedules and off‑chain processes). That combination renders even seasoned professionals vulnerable.

What can change the calculus? Firms like KXVC have already issued public warnings; victims have filed police reports; journalists and whistleblowers are amplifying patterns. But systemic improvements are needed: clearer industry standards on syndication and disclosure, better investor education, escrow‑style mechanisms for pre‑sale allocations, and more robust checks on people representing institutional brands.

Also Read: From pig butchering to work-from-home scams, crypto crime has become more professional and global

Recent precedents in the region underline the point. In 2023-2024, Southeast Asia saw several high‑profile investment failures where founders or executives were accused of misappropriating funds or falsifying metrics, cases that left partner companies and retail investors nursing heavy losses and reputational wounds. These episodes reinforce Dr Pun‑Arj’s argument that governing conduct matters as much as technical sophistication.

Quotes that matter

KXVC’s warning is blunt and instructive: “Beware of imposters… KXVC has never raised funds from external sources and has not authorised any individual to act on its behalf in such manner.” It is a reminder that institutional identity can be weaponised in private markets.

Mark, one of the victims, captured the personal betrayal succinctly: “He worked at multiple VCs. He spoke on panels in Thailand, the US and Europe. It’s hard to understand.”

Opinion

Southeast Asia’s startup ecosystem has matured rapidly, but maturity requires not only capital and talent but also institutional hygiene. The Vimolnoht affair is a wake‑up call: charisma should never substitute for verification. Investors (institutional and retail alike) must demand paperwork that can be independently verified, insist on escrow or custodied arrangements for allocations, and treat personal introductions as the start, not the conclusion, of due diligence.

Regulators and industry bodies should tighten identity‑and‑representation norms for funds and require clearer disclosures on fundraising and allocation processes. Equally, platforms that facilitate private market deals must build safer rails: standardised contracts, provenance tracking for allocations and stronger remedies for victims.

Also Read: With new US$100M fund, KXVC aims to help global AI, deeptech, Web3 founders win APAC market

In short, the cure for confidence eroded by bad actors is not fewer deals, but smarter markets. Southeast Asia’s innovation boom can survive and thrive if stakeholders harden the institutional scaffolding that supports its convivial networking culture. Charm sells; proof secures. The region needs both, and, crucially, the latter must trump the former.

The post How a crypto ‘insider’ in Thailand sold deals that never existed appeared first on e27.

In Malaysia, the government’s push for e-invoicing arrived with a timetable that reads like a procurement schedule. The Inland Revenue Board set a phased rollout starting 1 August 2024 for companies above RM100 million (US$21 million) in annual turnover or revenue, with later waves expanding coverage. By December 2025, after pushback from smaller firms, Prime Minister Anwar Ibrahim said the exemption threshold would be raised to RM1 million (US$210,000) from RM500,000 (US$105,000) beginning in 2026.

For Southeast Asia’s startups, that kind of regulatory drumbeat has become a growth engine for a class of products that rarely make the headlines: payroll, compliance, invoicing, procurement, expense controls, and the glue code that connects messy SME operations to increasingly digital state systems.

This is the quiet boom in enterprise back office. It is not powered by hype cycles, but by the daily mechanics of small business, where a missed invoice number can trigger a tax problem, and a payroll mistake can cost an employee’s trust.

Why is this working now

Southeast Asia’s economies are built on small firms, and they are structurally fragmented. MSMEs account for an average of 98.7 per cent of all businesses in Southeast Asia and contribute to 64.6 per cent of total employment, according to a 2024 regional snapshot.

That fragmentation is exactly why enterprise software adoption has historically been uneven. Many SMEs run on WhatsApp, Excel, paper receipts, and informal workarounds. They do not have dedicated finance teams. They do not have procurement departments. They often outsource compliance to external accountants who are juggling dozens of clients.

Yet the same fragmentation creates a large addressable market when the state digitises tax and reporting. Once a government mandates electronic invoices or tightens VAT clearance, optional software starts to look like basic infrastructure.

The buyer psychology shifts. Invoicing and compliance tools were long sold as efficiency. Increasingly, they are sold as risk reduction, and sometimes simply as the easiest way to keep operating.

Invoicing as policy, software as response

E-invoicing has become one of the most direct policy levers shaping SME software across the region, and it rarely moves in a straight line.

Malaysia’s MyInvois rollout is the most visible current example because it combines a phase-based timetable with political adjustment when compliance costs hit smaller firms. For software vendors, these rollouts create a predictable pattern: a rush of integration work at the top end, followed by a long tail of smaller businesses looking for low-cost tools, simple onboarding, and accountant-friendly workflows.

Indonesia is moving on a larger scale. The Directorate General of Taxes has been shifting VAT administration toward a more centralised platform known as Coretax, including changes to how tax invoice numbers are generated and managed as the system transitions. Advisory and compliance vendors describe Coretax going live in January 2025, with VAT reporting and invoice clearance increasingly centred in the new system.

Also Read: Building for fragmentation: How ASEAN SaaS leaders architect optionality into a paradox

Vietnam has already been through an earlier version of this transformation. It mandated e-invoicing nationwide from July 2022, backed by decrees and implementing circulars that pushed businesses onto electronic invoices.

Thailand’s approach is different. Its e-Tax Invoice and e-Receipt system exists and is promoted by the Revenue Department, but adoption remains largely voluntary, with incentives and programmes used to encourage usage rather than a blanket mandate.

For startups building invoicing and accounting tools, this diversity matters. A product that works in one country can fail in the next because the regulatory interface is different: API requirements, invoice schemas, digital signature rules, retention rules, and the practical realities of how small firms issue receipts.

That is one reason winners tend to be local, or deeply localised. It is also why many invoicing startups quietly become compliance companies. Their defensibility is not the UI. It is the regulatory plumbing and the support operation behind it.

Payroll and HR: The other unavoidable system

Payroll looks simple until it meets reality. Minimum wage variations, statutory contributions, tax filing requirements, overtime rules, contractor classification, and multi-entity groups turn “pay people” into a recurring compliance cycle.

In Southeast Asia, the payroll opportunity is amplified by informality and high SME churn. Many firms are formalising for the first time, and they want tooling that makes compliance feel manageable: templates, auto-calculation, reminders, and filings that do not require specialist knowledge.

The best payroll products in the region tend to win less through feature breadth than through trust. They need to be accurate, locally current, and supported by people who can answer questions in plain language. For investors, that can be attractive because revenue is recurring and the product is sticky, even if sales cycles are slower than consumer apps.

Procurement and spend: where leakage hides

If invoicing is about revenue and payroll is about people, procurement is where costs quietly escape, particularly in sectors with messy supply chains like construction, food services, and light manufacturing.

Singapore-based Doxa has built around that logic, positioning itself as a procure-to-pay platform for contractors, subcontractors and suppliers, combining workflow digitisation with payment and financing hooks. Its proposition is a useful guide to why back-office software can still be ambitious in Southeast Asia: procurement software can become a gateway to working capital, because visibility into purchase orders and invoices reduces underwriting uncertainty.

This is also where the next set of enterprise startups may differentiate. SMEs often do not have the discipline or headcount to enforce procurement controls. Software that embeds controls, approvals, supplier vetting, three-way matching, and budget policies can produce savings that feel immediate, which makes pricing easier.

Globally, investors have started paying closer attention to procurement automation, even calling it an unsexy problem worth funding. Southeast Asia’s version may be less about large-enterprise vendor sprawl and more about bringing order to informal supplier networks.

Also Read: SaaS isn’t always the answer: The case for physical innovation in developing economies

Distribution: accountants, banks, and the WhatsApp layer

The most important feature of this enterprise wave is not the category. It is distribution.

Many SME software companies in Southeast Asia do not sell directly to owners first. They sell through accountants, bookkeepers, payroll bureaus, and increasingly through banks and fintechs that want SME deposits and lending relationships.

Regulatory change strengthens these channels. When Malaysia tightens invoice rules or Indonesia shifts VAT systems, accountants become the front line of implementation, and the software that fits into their workflow spreads faster.

In practice, the best products accept a simple reality: SMEs will still use WhatsApp and spreadsheets. Winning tools integrate rather than replace. They pull data in, generate compliant outputs, and leave owners feeling like they did not have to become accountants to stay compliant.

What to watch next

This boom will not produce as many consumer-facing household names, but it is building durable businesses.

Two fault lines will shape outcomes:

• Regulatory interfaces keep moving. Vendors that invest early in integrations, documentation, and rapid updates will outlast those that treat compliance as a one-time build.
• SME willingness to pay is real but limited. Products that bundle value, such as invoicing plus financing, or payroll plus compliance reporting, tend to justify pricing better than standalone tools.

The pitch is simple. Southeast Asia’s SMEs are being pulled into a more digital relationship with the state and the formal financial system. The startups that make that transition less painful are building the region’s next layer of enterprise infrastructure, one invoice and one payroll run at a time.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

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

The post Why payroll, invoicing, and procurement are SEA’s hottest startups appeared first on e27.

We have spent a full decade trying to kill friction.

Guided by growth gurus and endless A/B tests, we built the seamless experience. We wanted one-click checkouts and onboarding so simple a toddler could do it. We assumed that if a tool disappeared into the background, we won. The data at the point of sale usually backed us up.

But this obsession is a trap. By making everything effortless, we created a user base that is soft and uncommitted. We solved for the quick transaction, but killed the long-term value of the relationship.

The most expensive mistake a founder can make today is making their product too easy to use.

The problem with easy value

The current obsession with cognitive ease is a disaster for brand loyalty. When a user feels zero resistance, they invest zero mental energy. That leads to two major failures:

• The value disappears: The human brain is wired to think that if something is easy, it isn’t worth much. Behavioural science is clear on this: if you don’t have to work for a result, you don’t value the result. When onboarding is instant, the user achieves their goal without earning it. They get the benefit, but they don’t respect the tool. When the bill comes due, leaving is just as effortless as joining. There is no memory of a struggle or a win to keep them around.
• You aren’t building memories: Loyalty requires memory, and memory requires action. We remember the things that challenged us. By letting a user slide through your product like they are on a greased chute, you prevent them from forming a real connection to the work. They are staying because of convenience, not conviction. This is the hidden cost of perfect UX: your churn looks great in month one, but it falls off a cliff by month six because the customer has no deep reason to stay.

Using speed bumps to keep customers

Smart Founders should stop trying to erase friction and start using “Intentional Friction.” I call this the Speed Bump Theory. It isn’t about making a bad product. It is about identifying the specific moments where a little bit of work creates a lot of commitment.

Also Read: Elevating your e-commerce strategies with livestreaming and hero products

Try placing these speed bumps at four specific stages:

• Hard onboarding: Don’t let them glide in. Force them to spend five minutes configuring a vital piece of the system. Maybe they have to map out a complex business process or upload a messy historical dataset. This creates an immediate sunk cost. Because they put in the work up front, they are psychologically anchored to the platform. They can’t leave easily because they already did the heavy lifting.
• The mastery gap: Your best feature should not be obvious. It should require a tutorial or a brief training session. This shifts the focus from time to value to time to mastery. When a user finally learns how to use a complex tool, that feeling of achievement is linked to your brand. They aren’t just using an app anymore; they have become experts in a specialised skill.
• Honest pricing: Stop hiding your price in a friendly little table. Force the user to actually look at the cost and justify it. If your product is actually worth the money, making them think about the price reinforces its worth. If the decision is too easy, they will never see the product as a serious investment.
• The exit warning: When someone tries to cancel, don’t just let them click a button. Remind them exactly what they are walking away from: their data, the skills they learned, and the momentum they built. This isn’t about being annoying. It is a final reminder of the value they are about to lose.

The race for “zero friction” is a race to the bottom. True winners are the Founders who realise that a strategic speed bump isn’t a barrier to entry. It is a barrier to leaving.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

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

The post The speed bump theory: How strategic friction creates loyal customers appeared first on e27.

In enterprise cybersecurity, the most dangerous moment rarely looks dramatic. It looks routine: a “normal” login at an unusual hour, a legitimate tool used in an unusual sequence, a small configuration drift that quietly widens access, a patch delayed because production can’t pause. Instead of announcing themselves out loud, breaches often blend in.

That reality is precisely why AI-powered cybersecurity solutions are becoming central to next-generation enterprise resilience, & are helping security teams respond with greater precision, recognise suspicious patterns faster, and reduce risk across sprawling cloud and hybrid environments.   

Today, organisations are facing an uneven battlefield where attackers automate reconnaissance & exploitation at scale, while defenders contend with alert overload, fragmented toolsets, and an expanding attack surface across endpoints, identities, applications, APIs, & third-party connections.

Traditional controls remain essential, but speed and correlation now determine outcomes. Modern enterprise cybersecurity programs increasingly rely on AI to connect signals across logs, network traffic, identity events, endpoint telemetry, & application behaviour to turn raw data into prioritised actions.   

“Enterprises aren’t short on security data; they’re short on time. The goal of AI in security isn’t to replace proven controls. It’s to make them smarter and faster so that teams can focus on what matters, reduce noise, and strengthen response readiness across the organisation.”

Why AI is shaping cybersecurity for enterprises 

As digital operations scale, security complexity grows nonlinearly. Multi-cloud adoption, SaaS sprawl, remote work, and increasingly modular application architectures create more identities, more configurations, and more potential missteps. For many businesses, the challenge isn’t visibility but interpretation and speed. AI helps address that gap through: 

• Smarter detection: identifying anomalous behaviour that traditional rule-based alerts miss 
• Contextual correlation: linking scattered signals across systems into a single incident narrative 
• Prioritised triage: ranking threats by potential impact and likelihood 
• Faster response: triggering automated workflows for containment, remediation, and escalation 
• Continuous learning: adapting to evolving attack patterns and shifting baselines 

These capabilities are increasingly critical for cybersecurity for enterprises, where the cost of false positives is high , and the cost of missed signals is higher. 

Also Read: From grid to code: Why good cybersecurity will help deliver net zero

What next-gen enterprise security solutions look like 

Modern AI-led security programs typically bring together multiple layers of protection & orchestration: 

• Threat detection across the attack surface 

AI-Powered cybersecurity solutions strengthen detection across identities, endpoints, cloud infrastructure, networks, & application layers. They can surface subtle threats such as lateral movement, suspicious privilege escalation patterns, and anomalous data access behaviour in environments where attackers aim to “live off the land.” 

• Automated incident response and containment 

Enterprise resilience depends on reducing response time. AI-assisted playbooks can support actions such as isolating endpoints, rotating credentials, blocking risky sessions, and enforcing policy controls while keeping humans in the loop for high-impact decisions. 

• Security posture management for cloud and hybrid 

Misconfigurations remain a leading cause of exposure. AI can help prioritise misconfiguration risk based on context, enabling smarter remediation sequencing within broader enterprise security solutions. 

• Governance, auditability, and compliance readiness 

For regulated industries, security is inseparable from evidence. AI-enabled workflows can support audit trails, policy verification, and continuous monitoring, helping security programs demonstrate control maturity without slowing operational teams. 

—

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 AI-powered cybersecurity solutions driving next-gen enterprise resilience appeared first on e27.

The Southeast Asian region has developed one of the most vibrant digital payment systems in the world. Real-time payments have become a daily routine in most markets due to mobile-first adoption, high wallet penetration, and fast innovation. However, payments continue to be made in a fragmented system throughout the region, with wallets, QR standards, regulators, and closed systems. And this fragmentation may be a logical consequence of developing financial infrastructure in different markets, policy regimes and financial maturity levels.

The real-time payments process in India started at a highly different point. It did not start with any tremendous amount of digital abundance, but was influenced by institutional constraints: unequal connectivity, distrust in formal finance, and the necessity to serve users at a population scale in the first place. Real-time payments were meant to be public infrastructure, and not a premium layer, starting to be reliable across banks, geographies, and use cases on day one.

With Southeast Asia on the path of increasing interoperability and cross-border real-time payment, the experience of India can be learned not only in the field of technical architecture. The principle of scale fundamentally alters the nature of risk, governance and
trust. Failure becomes systemic and not single. Conflicts, cheating and turnover have to be resolved in the present rather than in the past. The challenges can be seen only after the real-time payments have become a daily infrastructure.

The two territories started at disparate limitations, though they are drawing near to similar questions. It is more important to know what to embrace, as well as what not to copy.

Dissimilar origins, common goals

The real-time payments environment of Southeast Asia has been developing in a diverse environment. The presence of several sovereign markets, different regulators and different degrees of banking maturity has stimulated wallet-based innovation and blistering experimentation. In this regard, fragmentation has been a virtue, not a vice — enabling local ecosystems to optimise speed, incentives and user experience.

India, in its turn, treated real-time payments as one population-level infrastructure problem when it launched the real-time payment mechanism, Unified Payments Interface (UPI). Having little space to run parallel systems, and a lack of standardisation, interoperability was a governance option and not a market event. It was not so much about competition among networks but rather ensuring that any participant who met the requirements could access any user.

Also Read: Digital payments: Adapting to a changing world

Take the case of a local neighbourhood store that takes in QR payments. That QR can take a path through a particular wallet or closed-loop system (often tied to platforms such as GrabPay, ShopeePay, or similar super-app environments) and is optimised to be faster, more loyal and user experience in that ecosystem, which is the case in much of Southeast Asia.

If you look at the Indian situation, the QR has been tested to work across banks and apps, irrespective of the origin of the customer. Both are solutions to inclusion, albeit in different aspects, one adopting competition as the means of quick adoption, the other imposing interoperability to make it universally applicable to begin with.

Scale deranges everything, and what fractures

Once the real-time payments leave the niche use and scale to the population, the weakest assumptions of the system become apparent soon. The operation of processes which worked satisfactorily at smaller volumes starts to fail with velocity, resulting in delayed reconciliations, manual inspection, or post-factum dispute processing. The scale level transforms error cases into edge cases, but they are actual recurring experiences of regular users.

Take an example of a failed transaction at peak time. A payment is recorded in real-time, but credit confirmation is delayed or not done. In low-volume systems, these cases can be fixed by batch reconciliation or customer support processes in days. When scaled, that latency is a trust problem in minutes. Users demand immediate transparency: the payment is made or not. Confidence is lost much quicker than failure.

Scale also reshapes fraud. With rising volumes of transactions, trends change to organised, high-velocity exploitation. Limits, blacklists or rule-based flags are examples of static controls that can not keep up with the transaction settlement of transactions that settle immediately. Risk, refunds and remediation should thus work as fast as the payments themselves. These dynamics are already visible in markets across Southeast Asia and in India, as real-time payments become default rather than optional.

Speed without trust is incomplete infra

Since real-time payments are no longer an exception but a default, speed is no longer a differentiator. More important is how the systems respond to situations of uncertainty, failed transactions, delays in credits, debits in question, or even suspected fraud.

Under these circumstances, user trust is not determined by whether a system is perfect or not, but by whether results are transparent, prompt and responsible. In the UPI platform in India, where there are more than 700 million transactions daily, only 0.7 per cent get declined for all these reasons these days, which used to be more than 10 per cent in 2016, in their early days, which is an indicator of the maturity of digital infra over time.

Also Read: Optimising cross-border payments for seamless APAC expansion

Dispute resolution in high-velocity settings can not be managed as a back-office activity any more. In the case of instantaneous money transfer, resolution times have to shrink. Ambiguity, even for a few hours, can destroy trust more quickly than an outright failure, especially to users who depend on digital payments as part of their everyday business.

Being able to see the status of transactions, the ability to reverse them predictably, and the ability to point to a clearly defined responsibility among the participants are as important as throughput and uptime. In their absence, quick payments will pose a danger of increasing frustration instead of convenience. Infrastructure, which is not able to resolve failures in real-time, is not complete.

What SEA can learn to borrow, but not copy

The experience of real-time payments in India does not provide a template to be emulated, but it does uncover principles to travel across markets. The most prominent of them is the fact that the moment the payments turn into important public infrastructure, the governance decisions become as significant as innovation. Interoperability, dispute resolution and accountability are not optimisations that can be added in later stages; they determine user trust initially.

In the case of Southeast Asia, it is not the gains of experimentation through the market that should be reversed, but rather to understand when coordination should be prioritised over differentiation. With increased volume and complexity of use, the same fragmentation that had facilitated speed can start to limit reliability. To design to scale, we need to have clarity of responsibility, predictable redress, and system-wide awareness of failure even in multi-market environments.

—

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 Fragmentation to scale: What the payment journey of India portends to Southeast Asia appeared first on e27.

DevOps gave us speed without chaos. SecOps gave us visibility and response. MLOps gave us repeatability for models. We’ve learned to operationalise entire disciplines once they become core to how products scale.

Yet one of the most damaging categories of risk on modern platforms still has no consistent operating model: abuse.

Not “cyberattacks” in the traditional sense. Abuse is what happens when systems are used exactly as designed, just not by the kind of actors the designer imagined. Its referral loops turned into cash machines, reputation systems turned into influence markets, recommender algorithms turned into distribution hacks, and onboarding flows turned into factories for fake identity.

We have names for almost every operational maturity curve. But we still don’t have a widely formalised equivalent for adversarial misuse. If we did, we might call it AbuseOps. Or more precisely: abuse engineering.

Why abuse doesn’t fit traditional cybersecurity

Cybersecurity has historically focused on preventing unauthorised access and protecting confidentiality, integrity, and availability. That worldview assumes clear lines: an attacker is “outside” trying to get “in.”

Abuse blurs those lines. Often the actor is technically a user. Often the activity is technically permitted. And the “exploit” isn’t a software vulnerability, it’s a business rule, incentive, or algorithm that can be manipulated at scale.

That’s why many organisations struggle to place abuse. Customer support sees it as an operational nuisance. The product team sees it as edge cases. Security sees it as adjacent but not quite security. Fraud teams handle parts of it, but usually in narrow domains like payments or chargebacks.

Meanwhile, adversaries treat abuse like a profession.

Abuse is an economic game, not just a technical one

The most important shift is this: abuse is driven by ROI.

Attackers don’t just break systems. They farm them. They test small variations, share playbooks, outsource pieces of the workflow, and iterate until they find a repeatable profit engine. Entire ecosystems now exist to supply the building blocks: account creation, SIM farms, bot tooling, CAPTCHA solving, reputation boosting, mule networks, document forgeries, and even deepfake services. What used to require expertise is now packaged like infrastructure.

Also Read: The banking revolution: Balancing convenience and security in the digital era

This is not a “patch it and move on” environment. It’s an adversarial market.

And that is why abuse is best understood as adversarial economics: actors respond to incentives, constraints, and friction the way businesses respond to price signals.

Where abuse shows up first

If you run a platform with distribution, reputation, or rewards, abuse will show up, usually long before a breach does.

It appears in incentive systems: referrals, credits, cashbacks, promotions, loyalty points, and free trials. These mechanisms are designed to accelerate growth, but they can also manufacture value out of thin air when adversaries loop them.

It appears in algorithms: search ranking, recommendations, review systems, “verified” badges, trust scores, and content feeds. The goal isn’t access; it’s advantage. Distribution is currency.

And it appears at the system level: the quiet assumptions embedded into onboarding, rate limits, verification, payout rules, and enforcement logic. Attackers aren’t only probing your code. They’re probing what your product believes about users.

The real problem: Abuse is everyone’s responsibility, so it becomes no one’s

Many companies only take abuse seriously after it distorts metrics or triggers a visible incident. Until then, it gets handled through scattered mitigations: a rule here, a manual review there, an emergency blocklist, a “temporary” policy exception that becomes permanent.

This creates the same pattern: whack-a-mole responses, inconsistent decisions across teams, and rising operational load. Detection grows noisier, enforcement becomes more brittle, and the user experience suffers because friction is added broadly instead of precisely.

AbuseOps isn’t about creating a new label. It’s about admitting that abuse has a lifecycle and needs ownership, tooling, measurement, and governance just like delivery, incident response, or ML deployment.

What abuse engineering actually does

Abuse engineering starts by treating misuse as a design input, not an afterthought.

It asks a different kind of threat-model question: not “how do we prevent intrusion?” but “how do we prevent profitable exploitation?” That changes the work from chasing individual bad actors to redesigning the conditions that make abuse viable.

It then builds the foundation most abuse programs lack: observability. You can’t control what you can’t see. Abuse detection depends on understanding entities and relationships across accounts, devices, payment instruments, content, networks, and behaviour over time. Without that, enforcement becomes guesswork, and guesswork creates either high false positives or low deterrence.

Also Read: From back office to frontline: How fraud teams became revenue drivers

Finally, abuse engineering becomes the discipline of targeted friction by adding resistance where risk concentrates, not where everyone pays the cost. The objective isn’t to make the platform “more secure” in the abstract. It’s to make abuse expensive, unreliable, and difficult to scale while keeping legitimate users moving smoothly.

The north star: Make abuse unprofitable

A useful mental model is simple: adversaries optimise for ROI, so defence should attack ROI.

That usually means doing some combination of:

• Raising the cost of exploitation (verification, throttling, adaptive challenges)
• Lowering the payoff (caps, delayed payouts, clawbacks, reputation decay)
• Increasing uncertainty (controls that adapt, not static rules)
• Increasing consequence (consistent enforcement that’s hard to evade)

Most teams default to blocking. Abuse engineering focuses on economics: cost, payoff, and repeatability.

Why product leaders should treat this as a core strategy

Abuse isn’t only a security problem. It’s a product integrity problem.

Unchecked abuse degrades trust, pollutes datasets, distorts growth metrics, and creates a hidden tax in operational workload. In some businesses, it becomes existential because once users stop trusting the platform, your strongest moat turns into your biggest liability.

That’s why AbuseOps belongs upstream, close to product and engineering, not as a downstream cleanup crew.

A realistic starting point

The first step is not building a massive team. It’s choosing ownership and defining the system.

Create a shared abuse taxonomy that your org can use consistently. Agree on metrics beyond “how many did we block,” including loss, user friction, false positives, and time-to-mitigate. Introduce an abuse review loop for new features with incentives or distribution effects. And invest early in identity, telemetry, and entity resolution because every mature abuse program eventually realises those are the real primitives.

The shift

We’ve matured in how we build, ship, and operate software. But modern platforms aren’t only attacked, they’re manipulated.

That manipulation is not traditional cybersecurity. It is adversarial economics implemented through product mechanics.

If DevOps made delivery a discipline, and SecOps made defence operational, then the next discipline to formalise is abuse engineering because, at scale, the most damaging threats often come from people playing your system better than you expected.

—

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 Abuse engineering: The discipline security teams still don’t formalise appeared first on e27.

The digital economy in the Asia Pacific is like a fast-growing teenager: growing taller every month, moving into everything, and constantly being told, “Be careful on the internet.” Everyone wants more AI, more automation, more apps that magically know what we want before we do—but no one wants their data ending up in a breach, a scam, or a very awkward headline.  

So here we are, trying to build a future where we trust systems we don’t understand, run by algorithms we’ve never met, guarded by cybersecurity policies we definitely didn’t read, in a scam-increasing online environment with all sorts of tried and tested scams.

Cybersecurity: From “annoying IT thing” to trust superhero  

Not too long ago, cybersecurity was that department you only met when something went wrong—like the fire brigade, but with more acronyms and less water. Now, boards treat it as a strategic issue, and CISOs get invited to important meetings instead of being called only when someone clicks “Enable Macros” on a mysterious attachment.  

Think of cybersecurity as the “trust layer” of the digital economy: the invisible flooring that keeps everyone from falling straight into the basement of ransomware, fraud, and reputational disaster. Encryption, identity systems, zero‑trust architectures—they’re the unglamorous steel beams holding up your favourite fintech app, your government portal, and the AI chatbot you yell at when it hallucinates.  

When this trust layer works, no one notices. When it doesn’t, everyone suddenly becomes a security expert on social media.

APAC: So much growth, so many ways to panic  

In Southeast Asia and the broader APAC region, governments and businesses are in a hurry to digitise everything—payments, healthcare, transport, public services, you name it. That’s great for efficiency, inclusion, and impressive keynote slides. It’s also fantastic news for cybercriminals, who treat this region like a rapidly expanding buffet of poorly defended systems and distracted users.  

Cyber incidents and fraud losses have been surging, with some markets reporting eye‑watering growth in cyber-enabled scams and identity theft. People love the convenience of one‑tap everything, but they’re increasingly anxious about whether their data is safe, who can see it, and which OTP they just accidentally shared with a “bank officer” on WhatsApp.  

So yes, technical security matters—but here’s the twist: feeling safe is just as important as being safe.

Also Read: The trust layer: How cybersecurity became hospitality’s most valuable asset

Trust is a feeling, not a patch level  

Humans don’t walk around thinking, “I trust this platform because of its robust zero‑trust architecture and end‑to‑end encryption.” We think, “Does this thing look sketchy?” and “Will I regret clicking this later?”  

Psychology tells us that trust rides on a few simple things:

• Consistency: Does this service behave predictably, or does it randomly log me out and ask for 47 forms of ID?
• Transparency: Are you telling me what’s happening with my data, or hoping I never ask?
• Control: Do I feel I have choices, or am I being dragged through your consent funnel like luggage at an airport?
• Social proof: Who else trusts you—and did they survive?  

You can have world‑class security, but if your login page looks like it was designed in a hurry by a caffeinated intern, people will hesitate. Conversely, plenty of scams work precisely because they imitate the calm, polished look of something trustworthy. Our brains are wired to rely on signals and shortcuts, not security certification numbers.

Behavioural nudges: Jedi mind tricks for good  

Enter behavioural science and nudges—the gentle psychological steering that tech platforms already use to make you watch one more episode, add one more item to your cart, or accept one more cookie. The same techniques can make people more secure without turning them into full‑time security analysts.  

Some of the smartest “nudges” in cybersecurity look delightfully simple:

• Just‑in‑time warnings: A tiny banner that appears right when you’re about to click that suspicious email link, basically whispering, “Are you sure about this life choice?”  
• Secure‑by‑default settings: Multi‑factor authentication quietly switched on by default, so you’re safer before you’ve even finished complaining about the extra step.  
• Positive reinforcement: A small “Nice catch!” message when you report a phishing email, turning security from chore into a minor personal victory.  
• Human‑readable explanations: Instead of “Session terminated due to anomalous authentication behaviour,” try “We logged you out because something didn’t look right with your sign‑in—here’s what we did and what you can do.”  

Also Read: The unseen link: How cybersecurity and sustainability converge on Earth Day

These tiny tweaks don’t require users to become experts; they just make the safe path the easy, obvious one. Clever experiments in organisations show that such nudges can meaningfully reduce risky clicks and increase reporting of suspicious activity—without the usual cocktail of shame, blame, and twelve-page policy PDFs.

The awkward dance between humans and machines  

There’s an uncomfortable truth at the heart of the digital economy: we’re asking people to put enormous trust in systems they can’t see, run by companies they vaguely recognise, governed by policies they never read, secured by teams they’ll never meet.  

So if you’re designing that digital future in APAC—or anywhere—here’s the cheat code:

• Treat cybersecurity not as a cost centre, but as your reputation firewall and growth engine.  
• Pair strong technical controls with strong human signals: clear language, honest incident response, understandable controls.  
• Use behavioural nudges to make the secure behaviour feel natural, not heroic. Nobody should need willpower just to avoid being scammed.  

In the end, cybersecurity as a trust layer is less about scaring people into compliance and more about designing systems that quietly say: “We’ve got you. And we’ll prove it, not just in our architecture diagrams, but in every interaction you have with us.”  

If we get that right, people won’t just use the digital economy because they have to. They’ll use it because, somehow, in a world of bots and breaches and endless notifications, it actually feels like something rare: trustworthy.

—

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 Trust me, I’m (not) a robot: Cybersecurity, psychology, and our awkward digital relationship appeared first on e27.

There’s a widening gap in the AI automation space, and it’s not the one most people talk about.

It’s not the gap between those who have AI and those who don’t. It’s not about access to technology or understanding of capabilities. The real gap—the one that actually matters for business outcomes—is the execution gap.

On one side, you have SMEs with genuine operational problems. Real bottlenecks. Workflows that consume disproportionate resources, create delays, and limit growth. These aren’t hypothetical challenges invented for a case study—they’re the daily friction that prevents good businesses from becoming great ones.

On the other side, you have builders with technical capability. Developers, automation engineers, AI consultants who understand LLMs, RAG systems, API integrations, and workflow orchestration. People who can architect solutions, write code, and deploy systems.

The gap isn’t technical knowledge. The gap is execution in production environments against real business constraints.

Why most AI automation never makes it to production

The AI automation space is filled with proof-of-concepts that never ship, demos that never scale, and innovations that never deliver ROI. The pattern is familiar: a builder creates an impressive prototype, demonstrates capability in controlled conditions, and then… nothing. The solution never makes it into actual business operations.

This happens because building for real business environments requires more than technical skill. It requires understanding operational context, handling edge cases that emerge only in production, designing for maintainability by non-technical teams, and delivering measurable outcomes that justify the disruption of changing workflows.

Most builders optimize for impressive demos. The market needs builders who optimize for deployable solutions.

The AI Workflow Competition at Echelon Singapore 2026 exists to surface and celebrate the builders who understand this distinction—and to prove that a different model of collaboration between SMEs and technical talent can close the execution gap.

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

What makes this model different

Traditional approaches to SME automation follow predictable patterns. SMEs hire consultants who conduct discovery, propose solutions, and deliver implementations that may or may not align with actual operational needs. Or they adopt off-the-shelf tools that promise automation but require businesses to conform to rigid templates that don’t match how they actually work.

Both approaches treat automation as a product transaction rather than a problem-solving collaboration.

The AI Workflow Competition operates differently. It starts with real SME challenges—not consultant-interpreted problems, but actual operational bottlenecks described by the people who experience them daily. These challenges fall into three categories that represent genuine business priorities:

• Save-a-Hire challenges focus on reducing manual labor to free team members for higher-value work. The metric is hours saved per week. These are problems where automation doesn’t just improve efficiency—it fundamentally changes what a small team can accomplish.
• Revenue Rocket challenges enable new revenue streams or increase capacity to process more orders. The metric is additional revenue or order volume. These are problems where operational constraints are directly limiting business growth.
• Cash Flow Guardian challenges reduce operational costs, minimize waste, and optimize spending. The metric is cost savings per month. These are problems where inefficiency has a direct line item on the P&L.

Builders don’t pitch solutions to hypothetical problems. They build working automations for specific, measurable business challenges. The entire programme—from qualification through live demonstration—is designed to filter for execution capability, not presentation skills.

Why builders should care about solving real SME problems

For builders early in their careers or transitioning into AI automation, the challenge is often proving capability beyond GitHub repositories and side projects. Employers and clients want evidence of production experience—solutions that worked in real business environments, handled actual edge cases, and delivered measurable outcomes.

Working on genuine SME challenges provides exactly this proof. You’re not building a demo for a hackathon that gets archived after judging. You’re creating automation that an actual business might implement, solving problems that have real costs and real impact.

The programme structure reinforces this. Before you even work on an SME challenge, you complete a qualification task proving you can execute within constraints. During the 5-day build sprint, you develop working workflows with real logic, error handling, and functional outputs—not wireframes or mockups. At Echelon Singapore, you demonstrate your solution running live, showing how it handles standard cases, edge cases, and recovers from errors.

This isn’t about adding another line to your resume. It’s about building a portfolio that proves you can deliver in production environments.

For experienced builders—AI consultants, automation engineers, startup founders—the value proposition is different but equally compelling. The competition provides structured access to real SME challenges that represent common patterns across industries. Solve one well, and you have a repeatable solution applicable to dozens of similar businesses. The live showcase at Echelon Singapore puts your work in front of 10,000 tech professionals, investors, and business decision-makers. The ecosystem connections create direct pipelines to clients, partnerships, and commercial opportunities.

Most importantly, it positions you as a builder who ships, not just someone who talks about what’s possible.

Also read: Join 150+ builders creating AI workflows that solve real SME problems

What this means for the future of SME automation

Southeast Asia has thousands of SMEs facing operational challenges that AI workflow automation could solve. What’s missing isn’t technology—the tools exist, the platforms are accessible, the models are available. What’s missing is the execution layer: builders who can translate business problems into working solutions that non-technical teams can operate.

The current model doesn’t scale. SMEs can’t afford enterprise consulting rates. Builders can’t access real business problems to prove their capability. The gap persists.

The AI Workflow Competition tests a different model: direct collaboration between SMEs with real challenges and builders with execution capability, supported by infrastructure partners, technical mentorship, and a structured programme that filters for quality.

If this works—if the competition produces deployable solutions that SMEs actually implement—it proves something important about the future of automation. It proves that the barrier isn’t technology or cost. The barrier is collaboration structure and execution focus.

The builders who succeed in this environment will define the next wave of SME automation. Not because they know the latest frameworks or can implement the most sophisticated architectures. Because they can ship solutions that work in messy real-world environments, deliver measurable business value, and operate reliably in the hands of non-technical teams.

The builders we need

Right now, AI consultants, automation engineers, experienced developers, startup founders, and early-career builders are entering the AI Workflow Competition. The technical backgrounds vary—AI engineers with LLM experience, full-stack developers building integrations, no-code experts mastering automation platforms, student innovators ready for real-world challenges.

What unites them isn’t a specific technology stack or years of experience. It’s the willingness to be measured by execution, not ideas. The commitment to build solutions that actually work, not just impressive demos. The understanding that business impact matters more than technical sophistication.

Only 150 builder spots are available. Registration closes 17 April 2026.

If you’re a builder who understands that shipping matters more than showcasing, that production reliability beats demo impressiveness, that business outcomes are the measure of success—this is the arena that proves it.

The execution gap won’t close through better tools or more accessible AI. It will close through builders who can deliver working solutions to real business problems.

Register now and prove you’re one of them.

About the AI Workflow Competition

The AI Workflow Competition is an e27-led programme showcased at Echelon Singapore 2026, designed to explore how AI workflow automation can solve real operational challenges faced by small and medium enterprises (SMEs). Unlike traditional hackathons or idea-based challenges, this programme focuses on execution—bringing together SMEs, builders, mentors, and ecosystem partners to create practical, deployable automation solutions. For more information, visit the website.

 

The post Why the future of AI automation belongs to builders who ship appeared first on e27.