Generative artificial intelligence (GenAI) is set to affect the working lives of nearly 80 million people across ASEAN, but the technology has not yet produced the large-scale job losses often assumed in public debate, according to a new study by the International Labour Organization.

The report, Generative AI and labour markets in ASEAN: Significant exposure, limited disruption, uneven preparedness, estimates that 22.9 per cent of total employment in the region sits in occupations with more than minimal potential exposure to GenAI. Yet only 3.3 per cent of ASEAN’s workforce, or about 11.7 million workers, falls into the highest-exposure category.

Also Read: Beyond the hype: What generative AI is actually changing in startups

The distinction matters. Exposure does not automatically mean replacement. In many jobs, GenAI is more likely to alter tasks, compress workflows, or change skill requirements than eliminate roles outright. Around 67 per cent of ASEAN employment remains in occupations with no identified exposure to the technology, reflecting the region’s still-heavy dependence on agriculture, manufacturing, services, and informal work.

“The potential for labour market transformation is significant, but widespread disruption is not yet visible,” the report notes.

Singapore leads, but the exposure gap is regional

Among the nine ASEAN economies with available data, Singapore has the highest share of workers in occupations with more than minimal GenAI exposure, at 42.2 per cent of total employment. The Philippines follows at 28.1 per cent, reflecting its large services, business process outsourcing, and IT-enabled services base.

Indonesia stands at 21.7 per cent, Vietnam at 20.8 per cent, and Thailand at 20.6 per cent. The findings point to a familiar split in Southeast Asia: economies with larger formal services sectors and deeper digital adoption face earlier exposure, while those with bigger agricultural and informal workforces may see slower direct effects but also risk falling behind in productivity gains.

For Singapore, the finding is unsurprising. The city-state has spent years building AI governance frameworks, research capabilities, enterprise adoption schemes, and public-sector AI deployment. Its exposure is high because its workforce is more concentrated in professional, technical, administrative, and managerial roles — the very categories where GenAI tools can most easily automate or augment cognitive tasks.

The Philippines presents a different issue. Its BPO sector has long been one of the country’s main employment engines and a major export earner. GenAI’s ability to handle customer support, summarisation, transcription, and basic content generation puts pressure on lower-value work, even if higher-complexity services may remain resilient. This is not an immediate cliff edge, but it is a warning for a sector built on labour-cost arbitrage.

Startups face an adoption market, not just a disruption story

For Southeast Asia’s technology sector, the ILO report is less a story about robots taking jobs than about uneven enterprise adoption. GenAI use remains concentrated in technology-intensive occupations, while uptake is still comparatively limited in office and administrative roles despite their high exposure.

That gap creates a commercial opening for startups building AI workflow tools, vertical software, compliance systems, customer service automation, education technology, and human resources platforms. It also creates a harder question: whether small businesses can absorb these tools without widening the productivity divide between digitally mature firms and everyone else.

Southeast Asia’s internet economy remains large enough to support this shift. Google, Temasek, and Bain & Company estimated the region’s internet economy at US$263 billion in gross merchandise value in 2024, with revenue reaching US$89 billion. But the benefits of AI adoption are unlikely to spread evenly across a region where micro, small, and medium enterprises still account for the bulk of firms and employment.

Also Read: Is generative AI the game-changer for productivity?

The competitive landscape is already crowded. Global platforms such as OpenAI, Microsoft, Google, Anthropic, and Meta are embedding GenAI into workplace software used by regional companies. At the same time, Southeast Asian and Asia-focused players in customer engagement, voice AI, workflow automation, and sector-specific software are trying to localise products for language, regulation, and enterprise budgets. Companies such as WIZ.AI, Kata.ai, Yellow.ai, and regional system integrators are competing for the same automation budgets that banks, insurers, retailers, and contact centres are now reassessing.

Gender exposure is a policy problem

The ILO study also identifies a significant gender gap. Women in ASEAN are more than twice as likely as men to work in occupations with high GenAI exposure, largely because they are concentrated in clerical, administrative, and professional roles.

This finding complicates the common assumption that AI disruption will primarily hit male-dominated technical or industrial jobs. In the near term, it may instead affect office-based roles where women are heavily represented, including administrative support, routine documentation, customer operations, and clerical functions.

Young workers aged 15 to 24 and adult workers show broadly similar exposure levels, according to the report. That suggests the challenge is not limited to new labour-market entrants. Reskilling policies will need to cover mid-career workers as well, especially those in roles where GenAI changes the value of routine knowledge work.

Christian Viegelahn, ILO economist and lead author of the report, said the outcome will depend less on the technology itself than on institutional choices.

“Harnessing the benefits of GenAI requires more than access to technology,” he said. “Productivity gains depend on investments in human capital and social protection. Ultimately, future labour market outcomes will depend less on exposure alone than on the policy choices to build the preparedness and resilience of workers, enterprises and institutions.”

Preparedness may decide who benefits

The ILO report argues that ASEAN’s priority should be human-centred governance, broader access to skills training, support for MSMEs, and stronger knowledge exchange across member states. That agenda is not new, but GenAI makes it more urgent.

The risk for Southeast Asia is not simply job destruction. It is a two-speed labour market in which Singapore and digitally advanced firms use AI to raise productivity, while smaller companies and lower-income workers face task displacement without the tools, training, or social protection to adjust.

Also Read: Generative AI in daily life: A practical guide

For founders and investors, the next phase of the AI market in ASEAN will depend on whether products can move beyond pilots and productivity claims into measurable enterprise adoption. For policymakers, the test is whether AI strategies translate into worker-level preparedness rather than headline infrastructure announcements.

The ILO’s message is sober: GenAI is already relevant to tens of millions of ASEAN workers, but disruption is not predetermined. The region still has time to shape the outcome. Whether it does so will depend on execution, not rhetoric.

The post GenAI will affect 80M ASEAN workers, but mass job losses remain absent: ILO appeared first on e27.

The global cryptocurrency market experienced a profound structural shift over the past 24 hours, staging a major relief rally that directly challenged recent bearish momentum. Bitcoin led the charge, surging 4.10 per cent to reach a spot price of US$64,884.04 and outperforming the broader digital asset market, which posted a robust 3.71 per cent increase.

This sudden influx of buying pressure pushed the aggregate crypto market capitalisation up by 3.43 per cent, bringing the ecosystem’s total valuation to an impressive US$2.22T. Unlike isolated, native crypto events that occasionally spark volatility, this collective upward movement stemmed directly from external macroeconomic forces, signalling a tightening bond between digital assets and traditional financial markets.

The broader investment landscape witnessed a highly synchronised cross-asset response, with a remarkable 91 per cent correlation between cryptocurrency movements and the S&P 500 index and an 81 per cent correlation with Gold. These historically high statistical alignments indicate that digital assets are currently trading as a high-beta vehicle, deeply sensitive to global interest-rate expectations and broader dollar liquidity conditions.

The primary catalyst behind this aggressive market expansion was the highly anticipated release of the June United States Consumer Price Index data on July 14. In a surprise twist that caught many market participants off guard, the inflation print fell 0.4 per cent on a monthly basis due to lower energy costs, a metric that came in significantly cooler than the initial -0.1 per cent forecast.

This unexpected contraction cooled annual inflation down to a steady 3.5 per cent, delivering a massive wave of macro relief to participants who had previously feared aggressive interest rate hikes from the Federal Reserve. Because high interest rates typically drain liquidity from highly speculative, risk-on asset classes, this sudden disinflationary evidence sparked immediate expectations of future central bank rate cuts.

Traditional tech stocks and digital assets surged in tandem as capital rapidly rotated back into growth-oriented plays. For Bitcoin, this macro development reinforces its ongoing role as a sensitive atmospheric gauge of global monetary policy, meaning that any fundamental shift in the broader interest-rate outlook can trigger massive overnight capital reallocations.

Also Read: Why Bitcoin’s move to US$63K has nothing to do with crypto and everything to do with Iran

While the fundamental shift in macroeconomic sentiment laid the groundwork for the rally, the price action accelerated into a violent move due to a massive leveraged short squeeze in the derivatives markets. Traders who had positioned themselves aggressively for further downside were caught completely off guard by the positive inflation data, triggering a fierce feedback loop of forced buying.

Over the 24-hour window, the market saw a staggering US$104.12 million in Bitcoin positions wiped out by liquidations, with short sellers bearing the brunt, accounting for US$99.41 million of that total. This rapid cascading failure of short positions forced algorithmic buying engines to purchase spot and futures contracts at prevailing market rates to close out bankrupt accounts, adding immense artificial rocket fuel to the organic demand.

To complicate matters for bears, the average funding rate across major exchanges surged by an astronomical 158.42 per cent during this brief period, indicating an immediate and aggressive influx of bullish leverage as market participants scrambled to chase the breakout.

Simultaneously, the digital asset ecosystem enjoyed a healthy dose of sector leadership and speculative flow distribution that extended far beyond Bitcoin alone. Ethereum spearheaded this internal rotation by posting a notable 5.8 per cent weekly gain, significantly outperforming Bitcoin’s 2.02 per cent weekly return. This capital divergence was heavily amplified by social media chatter that framed Ethereum as a form of sound money uniquely positioned to thrive in a lower-rate economic environment, quickly establishing the Layer 1 narrative as the top-trending sector in the industry.

This speculative appetite was further validated by a massive 107 per cent surge in overall derivatives volume, alongside a steady rise in open interest, indicating that fresh institutional and retail capital was actively flowing into leveraged altcoin positions. This distinct shift in internal market dynamics indicates that the 24-hour rally was not merely a passive, index-wide response to stock market trends but rather a calculated rotation into major alternative assets, which could signal a sustained period of altcoin momentum if the Ethereum-to-Bitcoin ratio continues to strengthen.

From a strict technical and structural standpoint, the near-term market outlook remains distinctively bullish but faces immediate hurdles that will test the true conviction of spot buyers. Bitcoin successfully broke above its critical 7-day Simple Moving Average of US$63,476 and is currently working to solidify the 38.2 per cent Fibonacci retracement level near US$63,619 as a new baseline of technical support.

If the asset can decisively hold its ground above this pivotal US$63,619 line, the immediate path of least resistance points directly toward the 23.6 per cent Fibonacci retracement level located at US$65,006. Analysts must remain cautious, as 24-hour spot trading volume decreased by 21.33 per cent during this breakout, indicating a slight divergence between price appreciation and absolute spot market participation.

A failure to attract consistent spot buying volume at these elevated levels could lead to a rapid unwind of recent leveraged gains, potentially triggering a swift technical pullback toward the 50 per cent Fibonacci support level anchored at US$62,498.

Also Read: Why US$1.4 billion in Bitcoin longs could drag Bitcoin down to US$53,500?

Looking at the digital asset market as a collective whole, the aggregate valuation is currently testing a monumental technical resistance ceiling at US$2.25T, a level that represents the recent swing high for the total crypto market cap.

The immediate future of this macro-driven momentum now hinges entirely on the upcoming Producer Price Index data scheduled for release on July 15. If the incoming wholesale inflation figures confirm the disinflationary trajectory established by the Consumer Price Index print, the market will likely gain the fundamental backing needed to clear the US$2.25T barrier.

A successful technical breakout above this overhead supply zone would officially open the doors for a broader market expansion targeting the US$2.31T to US$2.38T extension zone. If the wholesale inflation data springs an unpleasant surprise on investors, the market may face a stern technical rejection at the current ceiling, resulting in a healthy period of consolidation or a temporary retreat down to the well-established US$2.14T to US$2.20T support band.

This rapid market recovery proves that while internal crypto mechanics like short liquidations and sector rotations dictate the immediate velocity of price moves, global macroeconomic liquidity remains the ultimate puppet master of valuation. The immediate trading bias for the market leans toward continued bullish momentum, but this optimistic outlook demands absolute validation beyond a single day of frantic short covering.

To transform this sharp relief rally into a legitimate, long-term market recovery, Bitcoin must comfortably sustain its position above the US$63,619 technical floor while simultaneously attracting consistent, positive institutional exchange-traded fund inflows in the coming days.

Investors must closely monitor both the immediate technical pivot points and the incoming wholesale inflation data, as the tension between overhead technical resistance and shifting global interest rate expectations will determine whether this impressive rally marks the beginning of a prolonged expansion or simply a temporary pause in a broader macroeconomic correction.

—

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 WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post Is the US$63,619 Fibonacci level strong enough to prevent a total unwind back down to US$62,498? appeared first on e27.

Nvidia’s reported move to halve the number of Asian customers authorised to buy its AI chips is more than a compliance story. It is a blunt reminder that Southeast Asia’s position in the global AI supply chain is neither neutral nor secure. The region is increasingly being treated not as a frontier for innovation alone, but as a possible circuit board in a much larger geopolitical struggle.

According to the Financial Times, Nvidia has tightened due diligence across Singapore, Malaysia, and Japan, removing more than half of its previous customers from an internal white list after tougher checks failed to clear many of them.

Also Read: Why Asia sits at the centre of the global AI chip disruption?

The obvious explanation is export control pressure from Washington, especially as the US tries to stop advanced chips from leaking to China through third countries. The less comfortable truth is that Southeast Asia has become a test case for how much trust global technology giants are willing to extend to local buyers.

This is not just about chips; it is about trust

For years, the region’s tech ecosystem has benefited from a simple assumption: if you can pay, you can play. That era is ending. In its place comes a more suspicious age in which companies are not merely customers, but potential compliance risks to be vetted, interviewed and visited in person. Data centres are inspected, contracts are checked, end users are questioned. The logic is not glamorous, but it is powerful.

For Southeast Asian companies, especially the neo-cloud providers that depend on Nvidia hardware to sell AI infrastructure, this is a serious blow. These businesses have marketed themselves as agile alternatives to the hyperscalers, offering access to scarce compute capacity in markets hungry for AI experimentation. Many have thrived on the promise that the region could become a genuine hub for distributed AI infrastructure, not just a consumer of imported technology.

Now they are being asked a more awkward question: are you a legitimate AI business, or a convenient waypoint in a sanctions-bypassing supply chain?

That question is important because, in Southeast Asia, perception can quickly become policy. Once a market is associated with trans-shipment concerns, the bar for participation rises sharply. Legitimate firms get dragged into the same scrutiny as bad actors. The result is a kind of collateral distrust. It is not an outright ban, but it can feel like one if you are the company suddenly trying to explain why your servers, customers and contracts look exactly as opaque as everyone feared they might.

Singapore and Malaysia are in the spotlight for different reasons

Singapore’s tech sector will feel this differently from Malaysia’s, but neither gets to escape the consequences. Singapore has spent years positioning itself as the region’s clean, well-regulated digital hub: the place where serious cloud players, AI labs and semiconductor investors can do business with confidence. If Nvidia’s checks are now focusing heavily on compliance in Singapore, that is not a compliment. It is a sign that even the most institutionally trusted markets are being pulled deeper into the enforcement perimeter.

Malaysia, meanwhile, sits closer to the hard edge of the issue. Its data centre boom has been one of the region’s most exciting investment narratives, with land, power and regional connectivity attracting a wave of global attention. But any boom built on the assumption of frictionless access to leading-edge chips is vulnerable when geopolitics decides to become a gatekeeper.

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

The irony is hard to miss. Southeast Asia is simultaneously being asked to build more digital infrastructure and to prove that this infrastructure will not be used in ways Washington dislikes. That is a tall order for a region that has historically preferred strategic ambiguity. Ambiguity is useful for diplomacy. It is less useful when your supplier wants names, use cases, contracts and the moral character references of your end users.

The AI race is becoming a compliance race

There is another uncomfortable lesson here: in AI, access to compute is now as strategic as access to capital. Nvidia’s chips are not just components; they are the toll gates of the modern AI economy. Whoever controls access controls the pace of development. And when those gates narrow, the impact is uneven.

Large enterprises and hyperscalers may absorb the shock. Smaller companies cannot. Startups building AI products, niche cloud providers and regional infrastructure players often depend on predictable supply and fast procurement. A white list turns supply from a business decision into a political and procedural one. That slows expansion, raises costs and makes planning harder. In short, it turns growth into a paperwork sport.

This matters because Southeast Asia is still trying to prove that it can produce AI companies, not merely host AI servers. If access to top-tier chips becomes more selective, the region’s emerging players may find themselves competing not just on product quality, but on the sophistication of their compliance teams. The irony is exquisite, and somewhat depressing: the future of AI might depend on who can produce the most convincing audit trail.

Washington’s shadow is widening

The deeper issue is that US policy is no longer simply about banning exports to China. It is about shaping the behaviour of third countries and private companies far beyond America’s borders. That is what makes this move so consequential for Southeast Asia. The region is not the target, but it is increasingly part of the mechanism.

The Commerce Department’s May guidance, aimed at advanced AI chips reaching overseas subsidiaries of Chinese companies, signals a broader enforcement mindset: if there is a route around the wall, the wall will be extended. Nvidia’s reported inspections and end-user interviews are the corporate translation of that policy logic. The company is not acting in a vacuum; it is trying to stay ahead of the regulator by turning compliance into a product feature.

This leaves Southeast Asian firms in a difficult position. They are expected to behave like sophisticated global operators, but many are still maturing operationally. Some may indeed have weak controls or murky customer links. Others may simply lack the legal, governance and documentation infrastructure demanded by American vendors in an era of intense scrutiny. Either way, the burden falls on the local ecosystem to prove innocence in advance.

What happens next will shape the region’s AI market

The immediate market effect will likely be consolidation. Firms that can clear compliance hurdles will gain advantage; those that cannot may lose access to Nvidia hardware or face delays that wreck business plans. Some will rebrand, restructure or cut ties with questionable clients. Others will disappear into the long list of regional companies that once looked promising until geopolitics discovered them.

But there is also a longer-term possibility: this shock could force Southeast Asia to professionalise faster. Better governance, cleaner customer due diligence and clearer ownership structures are not glamorous, but they are the price of admission to the high-end AI economy. The region cannot build a serious AI industry on hand-waving and optimism alone. The chip wars have ended that fantasy.

Also Read: Building smart: A tech founder’s guide to the semiconductor supply chain revolution

Still, there is a risk of overcorrection. If compliance becomes so heavy-handed that only the largest and safest buyers can participate, the region could end up with a concentrated AI market that serves incumbents and excludes the very startups most likely to drive innovation. That would be a classic Southeast Asian tragedy: enormous potential, strangled by asymmetric rules written elsewhere.

Nvidia’s white list may be a technical adjustment, but it carries a strategic message. Southeast Asia is no longer operating in a benign global market. It is operating in a filtered one. And in this new world, access to AI compute is not just a commercial advantage. It is a politically contested privilege.

The post The Nvidia clampdown is a warning for Southeast Asia’s AI boom appeared first on e27.

For the longest time, I thought burnout was simply part of entrepreneurship. You work harder. You sleep less. You push through. If you’re building something meaningful, surely exhaustion is just part of the price of admission.

Like many founders, I wore long hours almost like a badge of honour. I wanted to build big companies, create impact and prove that I could make something significant. The bigger the business became, the more responsibility I carried. At the time, that felt like success.

It took me years to realise that I wasn’t burning out because I loved building. I was burning out because I had become the operating system.

Every decision flowed through me. Every approval required my attention. Every miscommunication became my responsibility. Even when I delegated work, I still carried the mental load of remembering, checking, clarifying and correcting. The work itself wasn’t always exhausting. Carrying everything was.

Burnout often begins emotionally, but it becomes operational

When we talk about burnout, the conversation usually revolves around mental health, resilience or work-life balance. Those conversations matter, but they’re only part of the picture. As founders, we often overlook another source of exhaustion: operational complexity.

The more a company grows, the more decisions need to be made. More meetings. More approvals. More context switching. More people are interpreting instructions differently. More time is spent ensuring that what was intended is actually what gets executed. Eventually, your brain becomes the glue holding everything together. That kind of cognitive load is incredibly expensive, not because the tasks are individually difficult, but because they never stop.

Also Read: Employee burnout is real and why it needs to be taken seriously

One of the biggest stresses wasn’t the work, it was losing control of the message

One of the hardest lessons I learned wasn’t about revenue or fundraising. It was communication. I would explain something clearly, only to discover later that what was delivered wasn’t what I had intended. Somewhere between my thoughts and execution, the message changed. Yet the responsibility still landed on my desk.

The bigger the organisation became, the more this happened. That isn’t a people problem. It’s an operations problem. Every additional layer introduces friction, more interpretation, more room for information to change as it moves from one person to another.

Founders often assume they’re overwhelmed because they have too much work. Sometimes they’re overwhelmed because they’re carrying too much operational complexity.

AI didn’t remove my workload, it changed what I needed to carry

People often ask whether AI has reduced my workload. The answer is yes, but probably not in the way they imagine. AI didn’t magically eliminate my responsibilities. It reduced the number of things my brain needed to constantly remember.

Instead of repeatedly explaining the same ideas, I could build systems that preserved context. Instead of relying entirely on memory, I could rely on documented knowledge. Instead of spending hours reviewing repetitive work, I could focus on decisions that genuinely required human judgment.

The difference wasn’t simply productivity. It was mental bandwidth. That’s an important distinction.

Also Read: How to combat burnout and boost your productivity

My definition of scale has changed

When I was younger, I thought building a successful company meant having more people, larger teams and bigger organisational charts. Today, I don’t see the scale that way anymore.

I still want to build ambitious companies. I still want to create meaningful technology. I still enjoy moving quickly. But I’ve realised that success isn’t measured by how many people report to you. It’s measured by how much value you can create without unnecessary complexity.

The best founders aren’t necessarily the ones who can carry the most. They’re the ones who design systems that don’t require them to carry everything.

Every mistake shaped how I build today

Looking back, I don’t regret the mistakes. I don’t regret the burnout. I don’t regret wanting to build something bigger than myself. Those experiences shaped the founder I am today. They also shaped the way I teach entrepreneurs, not because I’ve figured everything out, but because I know how expensive certain lessons can be.

Every shortcut I share, every framework I teach and every AI workflow I build is really an attempt to help someone else avoid mistakes that took me years to understand. Failure is part of entrepreneurship. Burnout doesn’t have to be.

The next generation of founders won’t just build better products. They’ll build better operating systems.

The conversation around AI often focuses on replacing work. I think that’s the wrong question. The more interesting question is this: what if AI allows founders to stop becoming the operating system of their own companies?

Because perhaps the future of entrepreneurship isn’t about working less. It’s about ensuring that the work only humans can do is where our energy is spent. Burnout will always have a human side. But increasingly, it also has an operational one. And perhaps that’s where the next generation of founders should begin redesigning their businesses.

—

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 WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post Burnout isn’t just personal, it’s becoming an operations problem appeared first on e27.

As AI commoditises everything a company makes, the last defensible moat is not what you sell — it is the human experience you design around it. Most companies are investing in exactly the wrong thing.

In the mid-1990s, a Nike marketer told a room of executives: “We don’t sell shoes. We sell the feeling of being an athlete.” Three decades on, it reads like strategy. Walk into almost any product review today — specifications, roadmaps, feature releases. What the company makes. Rarely does the customer become.

Yesterday, a marketer’s Instagram Reel stopped my scroll. Twenty-nine likes. Just this: “People don’t pay for skincare. They pay to feel confident walking into a room. They don’t pay for coaching. They pay for certainty of achieving a goal.” You knew this already. So why does your board deck open with product metrics — not with who your customer is trying to become?

The milkshake nobody understood

In the late 1990s, a fast-food chain hired consultants to fix flat milkshake sales. Surveys. Focus groups. Flavour tests. Nothing moved. Then a researcher did something different: he watched. The most reliable customer was a lone commuter before 8 am, long drive ahead — not buying sweetness, but hiring something to defeat boredom. A banana was gone in two bites; a doughnut left sticky fingers on the wheel. The milkshake lasted twenty minutes. The competitor was not Burger King. It was the commute itself. The chain had spent months asking the wrong question.

What that researcher practised was radical empathy — not asking customers what they wanted, but inhabiting their experience long enough to see what they could not say. Clayton Christensen built Jobs to Be Done around this. His arithmetic was unsparing: 75 to 85 per cent of new products fail — not from poor execution, but because companies never understood the job the customer needed done. A concurrent McKinsey survey found nine in ten global executives dissatisfied with their innovation results. Nine in ten — after all the data, all the research, all the frameworks. The data exists. The empathy does not.

“Frame your business around the products you sell, and you get supplanted when technology changes. Frame it around the job you do, and new technologies become tools to do it better.”— Clayton Christensen, Competing Against Luck, 2016

What the East knew first

You might think this is what CRM systems are for. What recommendation engines do. What personalisation at scale delivers. The Japanese figured this out centuries before the algorithm — and arrived at something entirely different.

The word is omotenashi. Western management translates it as “hospitality.” That is not right. Hospitality responds. Omotenashi anticipates. Service gives you what you ask for. Omotenashi ensures you never have to ask. At Isetan in Tokyo, umbrella bags appear at the entrance before you notice you need one. No complaint triggered this. No model predicted it. Someone simply asked: What will this person feel when they walk in? That question — not the algorithm — is human experience design.

Also Read: When AI leaves the screen, cybersecurity becomes product responsibility

One framework came from Harvard. The other predates the printing press. Same conclusion — which most boardrooms still treat as optional: the organisation that wins understands what the customer has not yet found the words to say. Your CRM cannot do that. Only radical empathy can.

When the story collapses

Peloton is the case study nobody wants to be. In December 2020, its stock touched US$162 — a US$2,500 bicycle turned into a cultural identity: not hardware, but the sensation of being a serious athlete, accountable to a tribe. By January 2022: US$24. Most analysts blamed the reopened gyms. Wrong. The product had not changed. The instructors had not left. What collapsed was the story customers told about themselves when they used it. Peloton had never designed that story — they had stumbled into it. When the context shifted, there was nothing to hold it in place.

Apple made the opposite bet, deliberately. Jobs redesigned the Apple Store around one question: not what do people come here to buy, but what do they come here to become? The result was human experience design in its purest form — not a product environment, but an encounter with a more capable self. That encounter cannot be copied or shipped in a software update. It lives in the designed space between a brand and a human being — which is, not coincidentally, why Apple’s retail revenue per square foot still leads every category.

The speed at which AI commoditises what companies make will always outpace the speed at which companies learn to understand what people feel. Radical empathy is not a corrective. It is the only strategy left.

The trap of intelligent personalisation

Here is what most AI transformation roadmaps assume: that personalisation at scale is omotenashi. It is not. Omotenashi is radical empathy — unhurried observation of one specific person in one unrepeatable moment. AI personalisation is pattern-matching: the customer receives what people like them statistically want. That is not empathy. That is a fast guess with good data. Customers can feel the difference between being understood and being predicted.

Zurich Insurance ran the experiment. Between 2023 and 2025, more than a quarter of its workforce completed empathy training — 46,000 hours. Its Net Promoter Score rose seven points in eighteen months: not from a product launch or a price cut, but from understanding what a customer was actually feeling. The ROI of radical empathy was not soft. It was the only lever that moved.

Also Read: Seasonal product cycles: Why some features only work at certain times

Accenture’s 2025 Life Trends study found consumers in 22 markets accumulating a “cost of hesitation” — rising distrust of digital content, hunger for something real. When everything can be generated, authenticity becomes the scarce good. Your competitors have the same models. They are running the same optimisations. What they cannot replicate is the human experience you choose to design.

The thing you have not built

Starbucks did not lose a decade because the coffee got worse. It lost the third place — that feeling that the room belonged to you — the moment efficiency became the priority. The product survived. The experience did not. Most leadership teams, hearing this, nod. Then return to optimising throughput.

The most defensible asset a company can build is not a product. It is the story customers tell about themselves when they choose you. That story cannot be generated. It cannot be A/B tested into existence. It has to be designed — through radical empathy, one human experience at a time. Most organisations have more customer data than at any point in history. They understand their customers less than they did a decade ago. That is not a paradox. It is what happens when measurement becomes the goal, and the thing being measured gets forgotten.

Your company has a Chief Data Officer. Probably a Chief AI Officer. Perhaps a Chief Experience Officer. When did any of them last spend an unscripted hour inside a customer’s actual day — not an interview, not a dashboard, just watching what their life costs them? If that question requires thought, you already know what is missing.

—

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 WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post Your customers are not buying your product, they are buying a better version of themselves appeared first on e27.