According to McKinsey, Southeast Asia is touted as the world’s AI arena, with stronger AI adoption momentum tracking ahead of the global average. This rapidly maturing regional ecosystem demands regulation and robust guardrails. From Singapore’s Model AI Governance Framework to Vietnam’s emerging regulatory approaches, it is becoming clearer that trust in an organisation’s data is crucial to the success of AI projects. 

Yet, building that trust is fundamentally a human variable. At the current speed of AI adoption across the region, productivity gains and long-term impact depend heavily on a talent reset. With Southeast Asia’s digital economy projected to exceed US$1 trillion by 2030, integrating human capability into AI design from the start should be a commercial imperative. What this means is rethinking how we hire, deploy, and retain the professionals tasked with overseeing these systems.

The limits of binary trust in AI and the oversight paradox

According to a recent report by the Singapore Economic Development Board (EDB), Southeast Asia is emerging as a key growth market for AI. Organisations are increasingly prioritising AI adoption to drive productivity, manage rising labour costs, and address structural workforce constraints. 

However, the report also highlights a critical challenge: while many companies are accelerating AI deployment, far fewer have developed the governance models, workforce capabilities and trust frameworks needed to scale AI responsibly and effectively.

Realising AI’s true value depends on having a workforce capable of governing its use. Too often, trust in AI is treated as a binary decision: either humans manually review everything, or they review nothing at all. In practice, both extremes fail—either destroying productivity or eroding systemic trust.

This tension creates the “human oversight paradox.” Solving it requires moving beyond binary workflows to embed selective, risk-based oversight into AI workflows. This demands a new breed of talent equipped with the specific skills to interpret, challenge, and guide AI outputs.

Also Read: A 65% probability explains the next likely move for Bitcoin as leverage clears

Redesigning oversight for scale

To scale enterprise AI, human review must evolve into scalable human oversight. Effective AI governance shifts from universal review to a selective, risk-based model, where people act as decision-governors, focusing only on outcomes that carry real impact. This is not a reduction in governance but a redesign that makes oversight scalable and practical for enterprise‑grade AI.

Effective oversight must be deliberate and proportional. Low-risk, repeatable tasks like invoice matching, form classification and operational forecasting can and should be largely automated. High-risk decisions with real human impact, such as financial approvals, healthcare determination, fraud detection, or regulatory reporting, require structured human judgment and clear accountability.

This risk-based approach aligns with how mature industries operate. Aviation, healthcare, and energy sectors do not apply uniform oversight, but instead calibrate intervention based on risk exposure. AI systems should be governed using the same logic.

The hidden risk of skills atrophy

A less visible but critical risk is emerging as AI adoption accelerates across Southeast Asia.

If humans are only engaged during rare edge cases, they gradually lose the situational awareness needed to intervene effectively when systems fail. This dynamic has been observed in semi-autonomous driving systems, where prolonged disengagement reduces response quality when it matters most. 

In AI systems, this manifests as skills atrophy. Humans remain technically “in the loop,” but are no longer meaningfully engaged in decision-making. In Southeast Asia’s already constrained talent market, skills atrophy becomes a governance risk as there may be too few experienced practitioners left with the judgment and situational awareness required to oversee AI systems effectively.

Also Read: The talent reset: Why AI is changing what makes people valuable

To avoid it, humans must remain active decision-governors, shaping thresholds, testing edge cases, and refining escalation pathways. At the same time, AI does not eliminate the need for expertise; it raises the bar. Organisations must actively design roles that keep humans engaged as decision-makers instead of just fallback operators.

With the right foundations in place, oversight becomes operational rather than aspirational, which requires a shift in hiring and workforce strategies: prioritising adaptability, critical thinking, AI literacy, and technical skills. 

Visibility as the foundation of trust

None of this works without visibility and control across the AI pipeline. As AI systems scale, governance quickly breaks down when organisations lack consistent data lineage, policy enforcement, and auditability across their environments.

This remains a significant challenge for many organisations across Southeast Asia, where data is often spread across fragmented ecosystems spanning on-premises infrastructure, multi-cloud environments, and SaaS platforms. Research from Cloudera’s Data Readiness Index highlights that data fragmentation and integration challenges remain key barriers to scaling AI effectively.

A more sustainable approach is to bring AI to the data, rather than moving data across disconnected systems. This enables organisations to maintain consistent governance, lineage, security, and oversight regardless of where the data resides, while giving teams greater confidence in how AI systems are trained and deployed.

Engineering trust at scale

Southeast Asia does not need to trade off speed for control, or between innovation and governance. The opportunity is to engineer trust by combining robust systems with a workforce capable of managing them. 

Solving the human oversight paradox ultimately depends on people: how they are trained, how they are deployed, and how they are empowered to work alongside AI. With the right balance of technology and talent, organisations can move beyond reactive oversight to operationalised trust; scaling AI responsibly while maintaining performance and accountability.

—

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 Bridging the AI trust gap: Overcoming the human oversight challenge in Southeast Asia appeared first on e27.

Try to use any mainstream flagship tech product today. Most products work and feel pretty similar, with very minute differences in the UI, UX, and outputs they have. They all have similar UI elements, with the same animations and visual language optimised for your interaction. This all feels like a hyper-optimised monoculture, which is somewhat impersonal.

The companies that have defined the tech space in the modern era all used to have a unique, distinct personality, perspective, and culture, which shaped behaviour and their products, creating a competitive difference between their services that led to factions of users supporting certain cultures or perspectives. However, in the interest of serving shareholder expectations, all of these unique elements of pivotal tech organisations converged into a singular monolith chasing infinite optimisation of workflows, costs, and services.

This is more directly visible in the AI race as they have all arrived at the same place: an AI-first strategy, focused on outspending the competition to seem more competitive, and cutting headcount to preserve their own margins. All of this to invest and lead in a field that has no conflicting reports of ROI as verified by independent institutions.

This convergence is a cyclic phenomenon as explored by DiMaggio and Powell in 1983, who call it institutional isomorphism. This refers to the process by which organisations in the same field, facing the same pressures, gradually come to resemble one another regardless of the efficacy of their monolithic methods. They identified three mechanisms driving it.

Coercive isomorphism: External regulation or market pressure forces conformity;

Normative isomorphism: Shared professional education and networks produce shared assumptions about how things should be done

Mimetic isomorphism: Organisations facing uncertainty simply copy whoever looks most successful.

Through a combination of these processes, they argue that organisations end up chasing a goal to seem legitimate or secure by their internal and external shareholders rather than a goal of being efficient or innovative. Becoming structurally similar to other organisations creates a sense of safety and security both internally and externally, which counteracts the uncertainty and anxiety of being truly innovative, and is, in my opinion, the playbook of the consumer tech industry right now.

What they used to be

It is worth remembering that each organisation in big tech at one point had unique identities, not in a brand and marketing sense but in terms of their internal culture and approach to their visual identity, innovation frameworks, and product creation. This helped create more “human” products that may not have had the utility they have today, but were loved by consumers because they could feel the creativity, passion, and culture of the people who created them. It acted as a competitive advantage that tied consumers to companies if their perspectives and behaviours aligned with the organisation’s culture.

Also Read: Why Singapore’s deep tech founders need more than good science — and how National GRIP is filling that gap

Clifford Geertz, one of the great anthropologists of the twentieth century, argued that culture is not a set of rules or structures but a web of significance: a shared system of meanings that people create through symbols, rituals, and language, and within which all action takes place. He insisted that you cannot understand why people do what they do without first understanding what things mean to them in their particular context. Behaviour, stripped of its meaning, is just movement. Culture is what makes it legible.

By that measure, the early tech giants were genuinely distinct cultures, with their own symbols, rituals, and webs of meaning.

Google’s “Don’t be evil” is remembered now as naive, or hypocritical, or both, but the origin of the phrase is more interesting than its ending. In 2000, before a meeting with the Washington Post about monetising search, an engineer named Amit Patel walked into the conference room and wrote these words on the board because he was worried that Google might tell a media company their articles would rank higher if they paid for it, and it resonated with and was adopted by the rest of the organisation. The phrase was a symbol in the Geertzian sense: a condensation of meaning for an entire set of values about what the company was and was not allowed to become, and for a while, it actually organised decisions.

Facebook had a different kind of culture: “Move fast and break things” encoded a worldview that valued speed and disruption, and criticised caution, polish, and consensus. By understanding that worldview, all stakeholders understood what kind of person thrived in this culture and who failed. The IPO letter Zuckerberg wrote seemed to describe a social mission: We don’t build services to make money; we make money to build better services. Whether or not there was any truth to it, it created distinct practices of innovation, culture, and product development, enshrined through ritual artefacts of the culture like the hackathons, the hacker ethos, and office layouts.

Apple embodies a culture of obsession with the “Think Different” campaign. Their obsession with a specific weight of trackpad, a specific shade of white, and the deliberate rejection of the fan all showcased their perspective and culture on what technology should be in relation to human life: subordinate, beautiful, invisible, and in service of the person rather than the engineer. The premium Apple commanded had nothing to do with benchmark scores. It came from the coherence of that meaning system.

These cultures were not perfect, and some were actively harmful. But they created distinct symbolic worlds, which meant employees inside each company knew how to act in the interests of the organisations, and consumers on the outside could read the difference, and chose to align with them in some cases.

How isomorphism erased them

Then, one by one, the symbolic worlds were flattened.

René Girard, French anthropologist and literary theorist, argued that human desire is not original, but mimetic. We do not independently decide what to want: we want what we see others wanting. This produces what he called the mimetic double bind: the more closely rivals imitate each other, the more intensely they compete, and the more indistinguishable they become in the process.

In the early years of the internet, the tech giants were differentiated enough that the mimetic dynamic was manageable. Google wanted to organise information. Amazon wanted to be the everything store. Facebook wanted to connect people. Apple wanted to make beautiful objects. These were genuinely different projects, and the competition between them was productive as each one was forced to excel on its own terms.

Also Read: Faster tech, slower brains: The biological blind spot of the AI race

But as the companies matured, the pressure for institutional legitimacy (DiMaggio and Powell) combined with the pull of mimetic desire (Girard) to produce convergence. Wall Street rewarded certain behaviours like aggressive growth, margin expansion, and platform lock-in, and punished others, creating the coercive isomorphic pressure. The circulation of talent between companies, through elite MBA programmes and shared Silicon Valley professional networks, created normative isomorphism: a shared set of assumptions about how a serious technology company should operate, what metrics it should optimise, what a good quarterly report looks like.

The uncertainty of the technology landscape, in terms of figuring out which bet will pay off and which platform will win, being left behind. The new meta was focused on being similar to each other to be safe, or to let the start-up ecosystem innovate and then acquire and dismantle them for parts.

Google dropped “Don’t be evil” quietly in 2015, Facebook became Meta, and Amazon’s working culture became a liability in the press. The webs of meaning unravelled and were replaced by professional management practices, which have led to the current day, where you could swap the AI announcements from any of these companies, and nothing would feel different.

The Geertzian analysis is that these companies lost the ability to produce thick culture, which is the kind that gives behaviour its specific, locally intelligible meaning, and replaced it with thin culture: the universal focus on shareholder value, which means the same thing everywhere, and just monolithic cultures.

Differentiation still exists

The appetite for identity-led companies hasn’t disappeared; if anything, it seems to be growing with the coming of Gen Z, but it’s just being served at the margins.

Patagonia is the clearest non-tech example. Sustainability at Patagonia is not a marketing initiative. It is a core operating constraint and a value embedded deeply enough in the organisation’s culture that it shapes what products get made, how they get made, and how they get sold. When the founder transferred ownership to a climate trust, it was the logical conclusion of a meaning system that had been informing consumer choices and trust for decades. Consumers pay a premium because they can read the culture, and they agree with it.

In tech, modern examples are rarer. Nothing leads with a unique design philosophy and visual language, which has led to focused consumer attention and retention. Their use of transparent backs, a glyph interface, and a refusal to conform to the industry trends, alongside their open and honest communication in the smartphone and consumer audio space, has garnered them a sizable but fiercely loyal set of consumers.

Similarly, Teenage Engineering makes expensive, slightly impractical audio hardware for people who care about the relationship between aesthetics and function. Both command loyalty disproportionate to their market share. Both are, in Geertz’s terms, producing thick cultures rather than thin ones: you can read what they believe from the objects they make.

But neither has proven the model scales in the way Apple has.

Apple remains the only major tech company that has sustained something close to a genuine cultural identity, and its strongest recent act of identity was the M-series chip transition, followed by the MacBook Pro line. The move to ARM-based silicon was not an incremental update, but a platform-level rearchitecting of chip manufacturing and OS design on a platform that the rest of the industry had written off for high-performance machines.

They reignited the race in ARM computing and forced Windows on ARM, Chromebooks, Snapdragon, and Qualcomm, all to begin innovating and competing in a way that hadn’t happened in years. This innovation was one that came from a consistent, decades-long point of view about the relationship between hardware, software, and the place of the machine in the user’s life. That coherence in culture and their distinct identity is what makes MacBook users such a loud, loyal, and highest spending minority in the PC market.

Where the rest actually are

Other big tech organisations that have left behind their identities aren’t faring as well as Apple has.

Also Read: Why Apple’s MacBook Neo is subsidising the next generation of engineers

Google Search, the product that made the internet navigable and the one that Google built its identity on, is now widely regarded as being absolutely useless due to its advertising policies and the hyper SEO-optimisation dominating search results. Users append “reddit” to queries to reach human-written results. The AI-generated summaries at the top of the page hallucinate with regularity. The irony is that the company, which built its entire identity around the integrity of information, has a search product that people have learned not to trust.

Microsoft Windows has similarly become one of the largest pain points for its consumers, accumulating lots of harsh criticisms for its bloat and intrusive AI features while engineering resources flow toward Azure and Copilot. It started with the anger around the recall feature from a privacy standpoint, to a near-universal hatred for the Copilot integrations. The consumer product is tolerated rather than chosen, and at this point, even that is contestable, as recently they have, for the first time in a long time, begun to lose market share to not only macOS, but also to Linux.

What does this mean? 

There is a particular kind of loss that is difficult to grieve because it arrives through entirely rational decisions. There was never a singular moment at any of these organisations where they decided to let go of their cultural and political identities. Each step along the way was seen as rational and defensible and as a reasonable response to growth.

The tragedy of isomorphism is that it doesn’t require villains. There is no singular decision or time that can be blamed for the hubris of these giants, but only institutional pressure and a need to appease their anxieties.  It only constitutes organisations making locally “sensible“ decisions inside a system that produces globally homogenising outcomes. 

What makes this truly difficult to reverse is that the tools available to large organisations in the form of restructuring programmes, cultural initiatives, leadership reviews, and others are themselves products of the professional management apparatus that isomorphism generates. You cannot rebuild a house with the sledgehammer that brought it down. The conditions that produce genuine cultural identity: conditions like urgency, collective beliefs, agility, and flexibility for small teams, and even the constrained, bootstrapped anxiety, all are slowly dismantled by isomorphism.

I do not want this article to at any point come off as an iliad of despair about the current state of technology. This is purely and simply just my observation stemming from an anthropological background, hoping to shed some light on what is happening, why it’s happening, and to incite discussions on how to tackle it. 

These organisations are the same that built the internet as we know it. They did it, in part, because they had strange points of view that organised behaviour and made certain products inevitable. That strangeness was not incidental to their success, but integral to it, which unfortunately has been squandered. The only question that remains is to see whether these Goliaths will be able to crawl back up to their initial dreams, or will there be a string of Davids ready to eat their lunch. 

In my next article, I will examine what happens when cultural homogeneity becomes a financial strategy, though tracing the AI bubble, the capital chasing it, and the companies building something real in the space the monoculture has left open.

—

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 Big tech’s innovation illusion — Part 1 appeared first on e27.

I’ve sat through dozens of Series B pitches, and there is a specific, palpable moment where the air leaves the room. The founder has just finished a brilliant presentation where the slides are clean, the vision is grand, and the growth curves are flawless. But then, an investor asks an unpolished question about a messy detail—a product delay, a spike in churn, or a key hire that didn’t work out.

In that moment, the founder has a choice: stay inside the “compelling story” they’ve spent months perfecting, or step out of it and talk about the messy reality of their business. When they stay inside the story, they flounder. They fail not because they lack vision, but because they have spent two years building a culture that values the appearance of success more than the accuracy of information. At Stage 2, this information decay is the single greatest threat to your valuation.

In the founding stage, a CEO is rewarded for optimism because they must sell a dream to keep the team moving. However, as you scale, this optimism often becomes a filter that creates severe operational risks. First, there is the issue of information decay, where bad news is softened as it moves up through layers of management until the CEO receives a “polished” version of reality and makes strategic decisions based on inaccurate data.

This leads to delayed course correction; if the culture rewards “green KPIs,” teams will hide failing projects longer than they should, burning capital that should have been reallocated months ago. Finally, this culminates in the “due diligence haircut.” Professional investors look for data-driven storytelling, and when they find a discrepancy between your pitch and your raw logs, they don’t just question that metric—they question your entire ability to manage the firm.

To increase your valuation and decrease execution risk, you must move from managing the story to engineering the feedback loop, starting with the acceleration of the “bad news” signal. A startup’s survival depends on the speed at which a failure reaches the decision-maker; if it takes a month to find out a marketing channel is failing, you have wasted a month of runway.

You must explicitly reward employees who flag failures early, making “speed of reporting” a more important metric than the “success of the initiative” itself. Research on psychological safety shows that organisations that normalise error-reporting mitigate long-term damage, a trait investors value because it ensures the company remains capital-efficient.

Also Read: Funded: SEA does not need more impact capital, it needs fewer weak capital seekers

Furthermore, you must prioritise stress-testing over consensus, as groupthink is the primary cause of strategic failure in scaling startups. If everyone in the room is nodding, it is a signal that nobody is thinking critically. For every major decision, you should appoint a “Red Team” whose only job is to find the flaws in the plan.

If the strategy cannot survive an internal attack, it will certainly not survive the market. Project Aristotle at Google proved that the highest-performing teams are those that allow for rigorous dissent, creating a scrutiny-tested business model that investors view as a de-risked asset.

Scaling is inherently a tangle of trade-offs, and attempting to force your business into a perfect, linear narrative suggests you don’t actually understand your own complexity. Instead, you should manage the complexity openly by leading with the trade-offs. When discussing your roadmap with stakeholders, explain exactly what you are sacrificing to achieve your goals, which demonstrates a mastery of the operational reality.

Columbia Business School research demonstrates that ignoring “bad news” signals—like operational friction or customer complaints—leads to a lower Net Present Value (NPV) for the firm. Showing you are aware of these signals builds institutional trust that a “perfect” story never could.

Also Read: The talent question every founder needs to ask before they try to scale

Ultimately, trust is a predictability asset that relies on your “Say/Do” ratio. If you tell investors you are a “product-led” company, but your engineering team is losing headcount and focus, that inconsistency becomes a glaring red flag. You must ensure your internal resource allocation matches your external messaging because if your actions and your words don’t align, you are creating organisational friction that slows down every transaction.

The “Say/Do” ratio is a core driver of firm value; when what you say and what you do are identical, you remove the “risk premium” that investors otherwise apply to your valuation. Investors at Series B are not looking for a visionary who is disconnected from their own operations; they are looking for a reliable engine. Stop trying to make the pitch sound better and start making the information move faster. In the high-stakes world of scaling, the truth isn’t just a moral choice—it’s a financial one.

Preparing for this level of scrutiny requires a radical internal audit before you ever step into the pitch room. You must look at the last six months of your operation and identify exactly when a major failure occurred and how many days passed before that information reached the executive suite; if that loop is slow, your feedback velocity is broken. You need to verify if your leadership team can articulate three credible reasons why your current strategy might fail, ensuring that your path is stress-tested rather than just a product of consensus.

Finally, audit your calendar and your capital; if your actual resource allocation doesn’t mirror the “compelling story” in your deck, you are essentially pricing in a credibility tax that will surface during due diligence. In the end, if a Series B investor sat in on your internal management meetings today, they should hear the same company described in your pitch—anything less is just performance.

—

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 The Series B collision: Why your execution is falling behind your pitch appeared first on e27.

In March 2026, a US federal court ruled in Amazon v. Perplexity that a user’s permission to an AI agent does not constitute the platform’s authorisation.

The ruling was narrow in its legal scope but enormous in its commercial implications: it established, in a jurisdiction that sets global precedent, that the chain of accountability in agentic commerce is not solved by obtaining a user’s consent. The platform, the merchant, and the rail all retain independent liability exposure.

The Agentic Economy Report by blockchain firm Morph, published in Q2 2026, frames this as one of the defining risks of the current technology cycle and makes a pointed prediction: a Fortune 100 company will publicly attribute a major cyber breach to an AI agent before the end of 2028.

Also Read: The US$0.20 payment that could rewire Asia’s financial rails

When that disclosure happens, it will, in the report’s words, “reset the liability map for every merchant and every issuer that depends on one.”

The accountability gap is structural, not incidental

The scale of the attack surface is not theoretical. Citi GPS has tracked deepfake scam growth at more than 2,000 per cent over three years. The public MCP ecosystem — the infrastructure layer through which AI agents discover and invoke external tools and APIs — now exceeds 10,000 servers, each one a potential entry point for a malicious actor or a poorly scoped agent instruction. AP2’s cryptographic mandates were designed precisely because authorisation and accountability remain unsolved at the protocol level.

The problem is architectural. As Dr Changhao Jiang, CTO at Cobo, states in the report: “Prompts are not permissions. The industry’s greatest risk is a failure of architecture: granting agents the power to act without the guardrails to stop them. To bridge the accountability gap, we must decouple an agent’s reasoning from its power to spend through the pact. By replacing ‘assumed trust’ with infrastructure-level enforcement, we ensure that while execution is autonomous, liability is absolute.”

This distinction — between an agent’s reasoning and its power to transact — is the central design challenge that the agentic payment stack has not yet solved at scale. The Mandate layer of the stack (Layer 2 in the Morph framework) attempts to address it through AP2’s Cart and Intent Mandates, which are cryptographic and hardware-backed. But the protocol is voluntary, implementation is uneven, and the legal framework for adjudicating disputes among agents, users, merchants, and issuers has not been tested at commercial scale in most jurisdictions.

Asia’s specific exposure

Southeast Asia and broader pan-Asia face compounded exposure on this question for three reasons. First, the region’s regulatory frameworks for AI liability are nascent compared to those that are taking shape in the EU and the US. Singapore’s Model AI Governance Framework is a voluntary standard; it does not create binding liability rules for agentic transactions. Most of the region’s emerging economies have no comparable framework at all.

Also Read: The invisible shopper rewriting Asia’s e-commerce playbook 

Second, Asia is disproportionately exposed to the deepfake and social-engineering threat vectors that feed into agent-based fraud. The Citi GPS 2,000 per cent figure aggregates global data. Still, security researchers have consistently found that Asia-Pacific is the fastest-growing target region for AI-generated fraud, driven by the region’s high mobile penetration, cross-border commerce volumes, and varying levels of digital literacy across income groups.

Third, the region’s super-app and embedded-finance architecture — where a single platform may function simultaneously as a social network, marketplace, logistics provider, and financial institution — creates uniquely complex liability chains. When an AI agent transacts within a super-app ecosystem, determining which layer should bear the loss for a disputed instruction is a question those platforms have yet to answer publicly.

The card networks’ defence and its limits

Visa’s Trusted Agent Protocol (TAP), listed in the Morph report’s standards comparison table, represents the card networks’ primary response to the accountability problem. TAP layers network-level agent identity and fraud-signalling onto card rails, essentially attempting to keep agent traffic inside Visa’s visibility and accountability perimeter. Mastercard has tied its agentic commerce strategy to its 40 per cent tokenised base and global issuer rollout.

The approach has institutional logic. Card networks carry deep fraud-management infrastructure, chargeback mechanisms, and regulatory relationships that open-protocol stablecoin rails do not yet replicate. For regulated, ticket-sized purchases — a flight, a hotel, a large electronics order — the card model retains meaningful advantages even in an agentic world.

But the economics break down at the volume layer. The Morph report’s Prediction 2 holds that most agent-initiated payments, by transaction count, will settle outside traditional card rails — not because card networks lose the high-value category, but because the count of agent transactions is dominated by sub-dollar machine-to-machine calls that the card model was never designed to handle. The liability framework that travels with the card does not automatically extend to x402-settled stablecoin micropayments. That gap is currently uninsured.

The disclosure that changes everything

Jordan Patapoff, VP of Ecosystem at Quicknode, captures the broader stakes in a quote cited by the Morph report: “Every technology wave has a moment when its infrastructure gets defined: TCP/IP, HTTP, OAuth. The agent economy is in that moment right now. The protocols adopted in the next eighteen months are the ones a generation of agents will run on.”

Also Read: Agentic commerce: How autonomous AI is quietly rewriting the payments stack

The liability question is part of that infrastructure definition. Whoever writes the standard for agent accountability — whether it is a protocol consortium, a card network, a central bank, or a regulator — will shape the commercial terms of agentic commerce for the next decade. For Asia’s fintech sector, the risk of arriving late to that standard-setting conversation is not merely competitive. It is the risk of inheriting liability frameworks written by and for markets elsewhere, applied to a region with fundamentally different commerce architectures, fraud profiles, and consumer protection regimes.

The Fortune 100 breach prediction may or may not resolve before 2028. The accountability gap it will expose is already open.

The post Why Asia faces the sharpest agentic fraud exposure appeared first on e27.

Bitcoin gained 2.62 per cent to reach US$63,048.16 over a 24-hour period. This price action closely tracks a 2.51 per cent rise in the total crypto market capitalisation. The entire digital asset market simply rebounds from multi-week lows. The Fear and Greed Index currently sits at an extreme fear reading of 15.

This metric confirms that broad market sentiment dictates the price action rather than any catalyst specific to Bitcoin. The broader market still remains down over 12 per cent for the week. We witness a relief rally operating within a larger downtrend. Participants must watch for sustained growth in market capitalisation above US$2.2 trillion to confirm a genuine shift away from bearish momentum. We must look past these temporary fluctuations and focus on the underlying network fundamentals.

Derivatives activity provides the mechanical explanation for this sudden upward push. Bitcoin open interest rose five per cent in the last 24 hours. This metric indicates fresh capital entering leveraged positions. Concurrently, liquidations totaled US$108.03 million. This figure represents a 19.41 per cent increase from the prior day. These numbers point directly to a squeeze of highly leveraged short positions during the upward move.

The price rise was significantly amplified by forced buying as the market liquidated these shorts. Traders should monitor the average funding rate for a flip from negative to positive. Such a shift would signal growing bullish leverage and confirm the strength of this derivatives-fueled bounce. I always treat these leveraged squeezes as speculative gambling where the odds temporarily favour the bulls. The underlying trend requires much more than a short squeeze to reverse.

Also Read: Why Asia faces the sharpest agentic fraud exposure

Institutional flow data presents a more fragile picture of the current market structure. ETF assets under management experienced slight outflows. The total dropped from US$105.32 billion last week to US$102.05 billion currently. This capital withdrawal contradicts the retail frenzy we see in the derivatives market. The immediate technical path hinges entirely on holding the US$62,000 support level. A successful defence of this floor could propel the price toward the US$65,000 resistance zone.

The market needs a daily close above US$64,500 to signal stronger bullish conviction. Without this conviction, the asset risks falling back into the US$60,000 to US$64,000 consolidation range. The trend appears to be stabilising right now, but it remains highly vulnerable amid a multi-week decline. I monitor these ETF flows closely because they reveal the true appetite of traditional finance amid macroeconomic uncertainty. Smart money moves cautiously before major economic announcements, and this behaviour perfectly illustrates that approach.

We cannot analyse these crypto movements in a vacuum because traditional macroeconomic forces dictate global liquidity. The tape repriced everything on Friday when the May jobs report came in hot and wages firmed. This data landed on a market already nervous about inflation.

The last Consumer Price Index print stayed uncomfortably high in annual terms. Producer Price Index readings remain warm, and the current tariff regime continues feeding into prices. A strong labour market and sticky inflation lead to only one conclusion. The Federal Reserve possesses no room to cut rates and has a real reason to maintain a hard stance. The market performed the mathematics in real time. Market participants pulled forward rate-hike expectations, the 10-year yield jumped toward 4.71 per cent, and the US$ broke higher. Gold and equities subsequently took the hit. This environment highlights the inherent flaws in centralised monetary policy. Policymakers react to past data instead of anticipating future realities, creating endless cycles of boom and bust.

Also Read: B2B founders keep skipping brand, and it is costing them more than they realise

This inflation reckoning arrives right before the first Federal Open Market Committee meeting for the new leadership. Markets trade the May data through the lens of Kevin Warsh. He serves as the 17th Chair of the Federal Reserve and took the oath on May 22. Jerome Powell remains a voting Governor. Warsh will preside over his first meeting from June 16 to 17.

The market already decided what it expects from this transition. With a hot labour market, sticky inflation, and tariffs still in the system, a new Chair who built his reputation as an inflation hawk has every incentive to come out hard. He needs to establish credibility from day one. This logic drives the current repricing of rate hikes. A hot Consumer Price Index print hands Warsh the cover to sound hawkish and keeps the USD bid. A soft reading provides the only thing that can take the edge off this move. We will see exactly what kind of leader he truly is very soon.

This macroeconomic tightening also accelerates the push toward decentralised alternatives. As central banks tighten their grip to fight inflation, they simultaneously accelerate the development of Central Bank Digital Currencies. I view these retail digital currencies as ultimate surveillance tools and mechanisms of control. They represent the exact opposite of the financial freedom that Bitcoin provides. When traditional institutions restrict liquidity and monitor every transaction, the value proposition of a permissionless network becomes undeniable. The current inflationary environment forces policymakers into a corner. They must choose between crushing the economy with high rates or allowing inflation to erode the currency.

This dilemma drives visionary individuals and institutions toward assets that operate outside their direct control. The resilience of the Bitcoin network during these periods of extreme monetary tightening proves its viability as a sovereign store of value. People increasingly recognise that true ownership requires absolute independence from government interference and centralised banking systems.

Also Read: Funded: SEA founders need a capital sequence, not another funding scramble

The underlying architecture of Bitcoin demonstrates remarkable structural integrity despite this overwhelming macroeconomic pressure. The psychological floor of this market reveals itself in the order book dynamics. Bid density increases significantly by 42 per cent as the price approaches the US$60,000 threshold. This metric correlates perfectly with the Glassnode Production Cost Metric and the Miner Shutdown Price. Staying above US$60,000 is the mission now.

—

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.

Image Credit: Art Rachen on Unsplash

The post Why US$60K is the most important number in crypto right now appeared first on e27.

Grab Holdings has formally become the majority shareholder of Superbank, marking a strategic escalation in the Singapore-headquartered group’s push to control financial services infrastructure in Southeast Asia’s largest economy.

The ownership milestone was reached after related entities, including A5-DB Holdings and GXS, acquired additional shares in May 2026, pushing Grab’s effective stake above the controlling 50 per cent threshold.

Also Read: Superbank, Genesis launch US$40M financing solution for Indonesian startups

The move cements Grab’s role not merely as a distribution partner but as a controlling owner of a licensed bank operating in Indonesia, where mobile-first financial services are rapidly reshaping consumer behaviour.

A clearer route to scale lending

Superbank has been one of the most prominent success stories among Indonesia’s digital lenders. The bank reported a 55 per cent year-on-year increase in its loan portfolio as of April 2026, a surge industry observers attribute largely to tighter integration with the Grab and OVO consumer ecosystems. Those platforms offer abundant data and customer touchpoints, from ride-hailing and e-commerce to payments, that can be used to underwrite loans and cross-sell financial products.

Profitability has followed growth: Superbank’s profit before tax leapt 1,529 per cent to 142 billion rupiah (about US$7.81 million) for the four months ending 30 April 2026. While the absolute profit figure remains modest relative to legacy banks, the scale of the improvement signals that digital distribution and low-cost customer acquisition can rapidly compress time-to-profitability when a bank is embedded within a large consumer platform.

This commercial logic appears central to Grab’s willingness to convert a commercial partnership into outright control. Owning a bank removes certain regulatory frictions around product development and gives Grab greater latitude to integrate credit, deposit and payment services across its apps.

Consortium backing and strategic partners

Superbank’s ownership reflects a consortium approach that mixes regional tech companies with local media and telco know-how. Alongside Grab, Singtel, KakaoBank, and Indonesia’s Emtek Group, these backers have steered product development and distribution since the bank rebranded from Bank Fama International to Superbank in 2023.

Also Read: Digital banks win transactions, not loyalty: A missed opportunity in Indonesia

That consortium model has been influential in Superbank’s rapid product rollout. Singtel and KakaoBank bring regional digital-banking experience, while Emtek offers local distribution channels and brand recognition. For Grab, the arrangement combines foreign capital and regional expertise with on-the-ground local partners, a pragmatic route into a market where domestic understanding and regulatory navigation remain crucial.

A crowded, competitive market

Grab’s majority stake comes at a moment when Indonesia’s digital banking sector is noticeably crowded. Regulators have licensed some 17 digital banks, and policy changes have encouraged foreign participation, allowing non-Indonesian investors to own up to 99 per cent of local lenders. That regulatory openness has invited cross-border competition, with established internet giants, telcos and financial groups all vying for scale.

For Grab, securing majority control of Superbank is both an offensive and defensive play. Offensively, it positions the company to accelerate product innovation, from point-of-sale financing to savings and insurance, while using its consumer touch points to drive scale. Defensively, it pre-empts rivals from buying the same infrastructure or forming competing alliances that could lock Grab out of lucrative financial flows generated by its apps.

The Southeast Asian angle

Indonesia is a bellwether for digital finance across Southeast Asia. With hundreds of millions of mobile-first consumers, many still underbanked or underserved by traditional lenders, the opportunity for platform-led banks remains substantial. Grab’s acquisition therefore has implications beyond Indonesia — it signals an intensifying phase of consolidation in Southeast Asia, where platform companies are moving from partnerships to ownership of financial infrastructure.

Other markets in the region will watch closely. If Superbank’s model — rapid user acquisition via platform integration, machine-learning-based credit underwriting, and low marginal cost distribution — continues to deliver outsized growth and profits, it could accelerate similar moves elsewhere. Regulators in countries such as the Philippines, Vietnam and Thailand are also revising digital banking rules, and Grab’s latest step will likely shape competitor strategies and regulatory conversations across the region.

Questions and risks

Despite the strategic logic, owning and operating a bank brings new sets of risks. Credit quality can deteriorate rapidly if underwriting standards loosen during aggressive origination pushes, competition could compress interest margins, and regulators may tighten oversight as digital banks grow systemic importance. Grab will need to demonstrate robust risk management, capital adequacy and operational resilience as Superbank scales.

Also Read: How digital banking is driving financial inclusion in SEA

There is also the broader question of ecosystem concentration. Critics argue that platform companies owning banking infrastructure can create single points of control over many aspects of consumers’ economic lives. Regulators balancing financial inclusion goals against concentration risks may respond with stricter scrutiny, a dynamic that could complicate rapid expansion plans.

What’s next

For now, the acquisition gives Grab a stronger hand in shaping the future of embedded finance in Indonesia. The company can expand credit and savings product distribution through its app, OVO, and partner networks, while experimenting with product bundles that tie payments, lending and marketplace services together.

How successfully Grab translates control into sustained, responsible growth at Superbank will influence whether the move becomes a template for further consolidation across Southeast Asia, or a cautionary tale of the challenges that come with running a bank in one of the world’s most dynamic digital-finance markets.

The post Superbank under Grab: what the takeover means for Indonesia’s crowded digital banking scene appeared first on e27.

Modern entrepreneurship dictates that we must be software-first, where there is low capital expenditure, rapid prototyping, and infinite scalability. This approach is designed to attract venture capital, which views hardware, inventory, and physical assets as toxic liabilities that inhibit explosive growth.

The result is a market saturated with businesses that are US$50,000 to start and US$50 million to sell, built on the fragile foundation of easily replicable code.

The contrarian truth, which I have seen validated across multiple industries, is this: High initial capital expenditure is a strategic advantage. Founders who prioritise an asset-first approach (embracing complexity, physical friction, and substantial up-front investment) are building businesses with superior long-term moats, stronger unit economics, and vastly higher liquidation value.

The liability of the easy start

The bad of the Software-First approach is the liability it invites. When the barrier to entry is low, competition is instantaneous and fierce. A successful SaaS application, having raised US$2 million, must spend the next five years fighting off dozens of highly efficient, low-cost competitors that can replicate the code and the business model in under a year. The capital is spent on fighting for market share, not on product differentiation.

Conversely, consider the asset-first approach:

A founder decides to enter the specialised commercial equipment rental market, focusing on niche, highly regulated machinery (e.g., cryogenic freezers for bio-labs or specialised aerial drones for industrial inspections).

• The bad (initially): They must secure US$2 million in debt or equity immediately to purchase the equipment. The process is slow, involves negotiation, legal work, and insurance. The market says this is inefficient.
• The good (long-term): The US$2 million in expenditure is now a non-replicable barrier to entry for every competitor.

The competitor cannot start their business tomorrow with a credit card and a laptop; they must raise the same US$2 million, navigate the same procurement hurdles, and wait the same six months for delivery. This friction creates a long-term moat that code simply cannot replicate.

Also Read: Funded: The quieter capital path founders keep missing

The unit economics advantage

The asset-first model, while demanding initial capital, offers significantly better control over long-term unit economics.

In a pure software business, the gross margin is high (often 80 per cent), but the customer acquisition cost (CAC) and customer retention costs are perpetually volatile and must be paid monthly. Competitors can always bid up ad costs or undercut subscription fees, eroding that high margin.

In the asset-first model, once the initial capital is spent, the business controls a tangible, revenue-generating asset.

• Fixed cost stability: The cost of the asset is fixed. The monthly revenue (rent, processing fee, etc.) is directly tied to a physical object, allowing for highly stable, predictable cash flow that is protected from digital price wars.
• Liquidation protection: If the business fails, the founder retains the core asset (the equipment, the real estate, or the specialised inventory), which retains tangible liquidation value. A failed software startup leaves behind a pile of worthless code and depleted cash. A failed specialised equipment rental company retains the equipment, which can be sold to recover the initial investment.

The future: The asset-first premium

The future of durable business formation will see a strategic pivot away from the pure-software model toward asset-first businesses leveraging digital tools.

Also Read: Burning billions: AI’s capital frenzy and its global implications

The smartest founders are not avoiding assets; they are seeking out industries where the initial capital outlay is necessary to create a structural, long-term choke point. They are building businesses that are intrinsically connected to the physical world, using technology only to optimise the deployment of that asset, not to be the core product.

This involves:

• Choosing complexity: Deliberately selecting regulated niches (waste processing, specialised healthcare logistics, commercial agriculture) where high start-up costs repel the vast majority of founders and VCs looking for quick flips.
• Capital as a weapon: Viewing every large capital expenditure not as a liability, but as a defensive barrier erected against future competition.
• Prioritising downside protection: Structuring the business so that failure still returns a significant portion of the initial investment, a luxury pure-software founders rarely enjoy.

We must stop worshipping the ease of the lean start and recognise that true, enduring success often requires embracing the complexity and high cost that creates a definitive, structural moat. The US$5 million factory may have been hard to build, but it was a better investment than the US$500,000 code repository that can be cloned and undercut by the next team of efficient developers.

Are you building a business that requires a competitor to raise 10 times your initial capital to compete, or are you building a business that can be started with a credit card and a weekend of coding? Is your priority low initial cost, or long-term, non-replicable profitability?

—

The post The capital cost strategy: Why high initial investment is your strongest protection appeared first on e27.

Most conversations I have with other founders about AI land quickly on the same question: what can we automate? It is a reasonable instinct. But for founders and leadership teams building AI-native businesses, the more consequential question is a different one: if AI handles more of the execution, what exactly are humans for?

The answer is more specific than “creativity” or “vision,” two words invoked so often they have lost their precision. The human skills that matter most in an AI-first business are strategic thinking and systems thinking. Together, they form the architecture layer that determines whether an AI-native business compounds in value, or simply runs faster toward the wrong outcomes.

Why these two skills, specifically

There is a structural reason why these skills remain irreplaceable, beyond familiar arguments about nuance and context.

AI is backwards-looking by design. Every model, every output, every recommendation is built on patterns extracted from past data. You can feed it new context and real-time signals, and it will process them intelligently. But its underlying reasoning is always a function of what has already happened.

Strategic thinking is fundamentally forward-looking. It requires forming a perspective on where a market is going before the evidence is conclusive, on what customers will want before they can articulate it, on which bets to make when the data is incomplete by definition. The best strategic decisions are made precisely in the space where historical patterns are the least reliable guide. That is the space AI cannot occupy.

Systems thinking adds a different dimension: the ability to see how parts of a business interact, how a change in one area creates second and third-order effects elsewhere, and how the overall system produces outcomes individual components cannot explain on their own. Without someone thinking at the system level, AI initiatives tend to solve isolated problems while creating new friction at the handoffs between them.

Both skills are intuition-heavy, built on real-world experience, judgment grounded in a specific business context, and a tolerance for ambiguity that cannot be trained into a model on historical data.

Also Read: What to actually prioritise when your board wants AI and everything feels urgent

How to choose

The practical question for any leadership team is what goes to AI, what stays with humans, and what requires both. A simple Signal vs. Execution layer model structures every capability across three levels.

• Execution is where defined tasks get done: content generation, data retrieval, report formatting, customer query handling, workflow automation. AI operates excellently here. The inputs are bounded, the success criteria are measurable, and the patterns are well-established. It is now reasonable to default to AI at this layer and free your team from the cognitive overhead it was consuming.
• Orchestration is where workflows are coordinated across functions, systems, and agents. AI can support this layer, but someone has to have defined the logic, the rules, and the exception-handling criteria that make coordination work. Systems thinking is what makes orchestration coherent. A human needs to own this layer even as AI tools increasingly execute within it.
• Direction is where the business decides what it is building, why, and in what sequence. AI can inform direction with data and scenario analysis. It cannot own it. Direction requires the forward-looking intuition that AI structurally lacks, and the systems-level awareness to understand how today’s choices shape tomorrow’s constraints.

A practical example: we recently rebuilt the GTM stack for one of our SaaS businesses to be fully AI-native at the Execution and (partially) Orchestration layers. Prospecting, lead qualification, outreach sequencing, and follow-up logic all run through AI-driven workflows. The team now spends 30 to 45 minutes per week on that function instead of roughly 20 hours, with a 1.8x higher response rate and 1.6x higher close rate. The Direction layer did not change. Who to target, what positioning to lead with, and which segments to prioritise remained entirely human. What changed is that AI executes that judgment at a scale and consistency no manual process could match.

The errors I see most businesses make are automating Direction (outsourcing strategic and systems choices to tools that optimise for past patterns) or leaving Execution to humans (wasting human capacity on tasks AI handles better). The framework is a forcing function to avoid both.

Also Read: Think with AI: The new skill for social entrepreneurs

What to pay attention to in the next 12 months

Competitive pressure will push many teams to deploy AI broadly and quickly. Some of that pressure is real. A lot of it will produce systems that move fast in the wrong directions.

Three things worth keeping close.

• Audit your direction layer: How many strategic choices are currently being made by default, through inertia, or by deferring to tools and benchmarks? If your team spends most of its cognitive energy at the Execution layer, Direction is likely underinvested.
• Build systems thinking as an organisational capability: The businesses that compound from AI are those where someone is consistently asking: how does this initiative interact with everything else we are building? This thinking needs to be embedded in how you design and iterate on AI deployments, not just exercised at the top.
• Resist the pull toward AI as a strategy: Using AI is not a strategy. It is a capability. Durable competitive advantage comes from a clear view of where you are going and a coherent system for getting there, with AI as a powerful layer within that system.

The point

AI will keep getting better at execution, and it will increasingly support orchestration. The layer it cannot occupy is Direction, and the reason is structural: it is built on the past, and Direction requires a view on the future.

Strategic thinking and systems thinking are what make the difference between an AI-native business that compounds and one that scales its existing assumptions faster.

These are the skills worth protecting, developing, and keeping close to the center of how you make decisions.

—

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 The two human skills that make AI-native businesses actually work appeared first on e27.

In modern meeting rooms today, there is a scene that feels both strange and normal at the same time. Someone opens a laptop, types a few sentences, and within seconds, a business proposal appears, along with market analysis, source code, presentation designs, and even marketing strategies that once required a small team working for days. On the other side of the room, someone sits quietly for a moment before speaking. They are not slower. They are thinking: “Is this the right decision for the people who will be affected by it?”

Moments like that are beginning to change how we understand human value.

For decades, humans were valued for their ability to process information. The faster someone could calculate, memorise, write reports, or organise data, the more valuable they became inside organisations. But AI is beginning to shift that foundation. Machines can now write faster, analyse more broadly, read thousands of documents without fatigue, and even generate ideas that appear creative. Many professions are starting to experience a quiet discomfort: if AI can perform most intellectual tasks, then where exactly does human value still exist?

That question is no longer philosophical. It is becoming personal.

A programmer watches AI generate hundreds of lines of code in seconds. A designer sees AI create illustrations from a single prompt. Analysts watch dashboards and insights appear automatically. Even writers quietly wonder whether readers can still distinguish between human writing and machine-generated text.

Yet the longer we live alongside AI, the more something interesting becomes visible: speed was never the true core of human value.

AI is incredibly powerful at answering. But humans still live in a much greyer territory: deciding which questions are worth asking.

And that difference matters.

AI can help companies optimise profits. But humans decide whether those profits are achieved in ways that damage or strengthen society. AI can help underwriting systems assess risk within seconds. But humans understand what it feels like to be afraid because of illness, unemployment, or the desire to protect one’s family. AI can create highly efficient business strategies. But humans decide whether organisations still retain a sense of humanity.

That is where I have started to believe human value is not simply “the things AI cannot yet do.” That definition is too fragile. AI capabilities will continue to evolve. If human value is defined only by the remaining gaps in machine capability, then our identity will continue shrinking every year.

What feels more fundamental is this: humans give meaning to decisions.

Also Read: What to actually prioritise when your board wants AI and everything feels urgent

AI generates possibilities. Humans choose the consequences.

And choosing consequences means bringing morality, empathy, fear, hope, life experience, personal wounds, culture, and even love into the process of work. Those things are difficult to measure in spreadsheets. Yet they are often what separates systems that are merely efficient from systems worth preserving.

The problem is that the narrative around “uniquely human skills” sometimes sounds too comforting. Many conferences claim the future will belong to empathy, creativity, leadership, and communication. It sounds beautiful. But honestly, AI is already entering those spaces too. AI can now speak warmly, write poetry, act as a “companion,” and generate fairly creative ideas.

So are we truly redefining human value? Or are we simply rebranding the tasks that automation has not reached yet?

I do not think the answer is clear.

And perhaps that uncertainty is important to acknowledge.

Because there is a strong possibility that a large portion of human work will change dramatically. Not only repetitive jobs. Creative and strategic work is being reshaped too. Many people are quietly experiencing a professional identity crisis. They once felt valuable because of specific expertise. Now that expertise can be replicated by AI at low cost and extraordinary speed.

At that point, humans are forced to confront a question deeper than “How do I remain competitive?”

The question becomes: “Who am I when my abilities are no longer rare?”

And that is not a question AI can answer for us.

I see this shift directly in technology itself. In the past, engineers were valued primarily because they could write complex code. Today, AI can generate much of the boilerplate work. But the engineers who remain truly valuable are becoming those who understand business context, operational risk, stakeholder conflict, long-term implications, and decision-making under uncertainty.

In other words, human value is shifting from production toward judgment.

Also Read: Think with AI: The new skill for social entrepreneurs

It is no longer about who can build something the fastest. It is about who can wisely decide what should be built in the first place.

Strangely enough, those abilities are often born not only from formal education, but from life itself. From failure. From loss. From making bad decisions. From leading people during difficult situations. From understanding that behind every dashboard metric are real human lives.

Perhaps that is why future organisations will change how they hire people.

Not only by measuring technical skills, but by evaluating the ability to think across contexts. The ability to navigate ambiguity. The ability to maintain moral direction while systems become increasingly automated. The ability to build trust in a world flooded with synthetic content and algorithmic decisions.

Ironically, the more advanced AI becomes, the more expensive human trust becomes.

Because in an era where almost everything can be generated by machines, people begin searching for something that feels real.

Not merely correct according to data, but emotionally sincere.

Not merely efficient, but humane.

Not merely intelligent, but morally accountable.

At the same time, I do not want to romanticise humanity too much. Humans are also filled with bias, ego, manipulation, greed, and error. Many of the world’s worst decisions were made by humans, not AI. So not everything “human” is automatically good.

Which is why the future may not become a battle between humans and AI.

It may instead become an ongoing negotiation between machine capability and human wisdom.

And we still do not fully know what that final shape will look like.

Perhaps many jobs will disappear, while entirely new roles emerge that we cannot yet imagine. Perhaps organisations will become far smaller yet more productive because they are supported by AI agents. Perhaps one person will be able to build a global company with only a handful of people and a network of AI systems. Perhaps degrees, titles, and traditional corporate structures will slowly lose meaning.

Or perhaps humans will simply grow exhausted from living inside worlds that feel overly automated, and begin valuing slower, more authentic, imperfect human interaction again.

I do not know.

But one thing feels increasingly clear.

Also Read: How AI is changing what an SME team actually looks like

Human value may not come from being faster than AI.

Nor from being smarter than machines.

Perhaps human value emerges from our ability to retain consciousness, responsibility, and meaning in a world where almost everything can be automated.

And perhaps the most important question is not:

“Will AI replace humans?”

But rather:

“When almost everything can be done by machines, what are the things we still want to preserve as deeply human?”

—

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 In a world optimised by AI, what is left for humans? appeared first on e27.

For a while, differential privacy was spoken about as though it might do for data privacy what encryption did for data in transit. A hard technical answer to a messy institutional problem. The theory was elegant, the guarantees were rigorous, and the early signal from major adopters was powerful. 

That gap matters because it tells us something larger about technology strategy. Differential privacy did not stall because the mathematics was weak. It stalled because organisations kept treating it as a universal privacy answer when it is really a precision instrument for a narrower class of problems. It is very good at answering one hard question. How do you release useful aggregate information while limiting what can be learned about any one person? That is not the same as solving privacy in the round. 

It solved a narrower problem than the market wanted

Most organisations do not actually need a mathematically formal guarantee for every privacy question they face. They need a working combination of access controls, minimisation, retention discipline, contractual restrictions, governance, and operational trust. Differential privacy sits inside that world. It does not replace it. The mistake was to imagine that a strong formal guarantee at the output layer could make the wider privacy problem feel settled.

In practice, most institutions still need to govern collection, purpose, access, sharing, deletion, model use, and accountability separately. That is why differential privacy often ends up as a specialised control rather than the centre of the privacy operating model.

This is also why the technology feels simultaneously important and oddly non dominant. It addresses a real problem, but not the whole one. Strategists often back technologies that appear to simplify governance. Differential privacy usually does the opposite. It sharpens one guarantee while leaving the surrounding organisational obligations very much alive. That makes it more honest than many privacy narratives, but also harder to sell as a universal answer.

Also Read: How to build customer trust with improved data privacy

The trade off is too visible to ignore

Differential privacy makes privacy expensive in a way organisations can actually see.

The noise added in the wrong way can either weaken privacy or make the data less useful. Leader need help understanding the trade offs inherent in differential privacy. Its earlier discussion of open challenges goes further and says broader use will require better processes both for measuring utility and for helping users work with differentially private outputs. That is the part many executives dislike. Differential privacy does not let them pretend privacy is free. It forces an argument about how much accuracy, granularity, or downstream usefulness they are willing to give up.

In other words, differential privacy did not fail commercially because it was too academic. It failed to become ubiquitous because it was too honest. It puts the privacy utility bargain in plain sight. In large institutions, that usually means politics. Product teams want fidelity. Analysts want details. Revenue teams want more segmentation. Policy teams want stronger protection. Noisy outputs are not only a technical choice. They become a budget, power, and accountability conversation. 

Epsilon is mathematically neat and managerially awkward

Differential privacy relies on parameters that matter deeply and are still difficult to explain outside specialist circles. There is still no consensus answer on what epsilon should mean in practice or how it should be set. 

That is not a minor educational problem. It is a strategy problem. A control rarely becomes ubiquitous when the key parameter cannot be translated cleanly into board-level language, regulator language, procurement language, and customer language all at once. Differential privacy is strong where the institution can tolerate technical nuance and invest in interpretation. It struggles when leaders want a simpler sentence than the truth allows.

The engineering burden is still heavier than the story suggests

Differential privacy is easy to describe and hard to implement well. That should not surprise anyone who has actually watched privacy technologies move from paper to production. The hard part is rarely just the mechanism. It is the surrounding system. Contribution limits, privacy accounting over time, query controls, data schemas, public versus private feature choices, and monitoring all need to line up. This is not plug-and-play privacy. It is a design-heavy privacy.

Also Read: Data privacy for startups: Simple steps to protect sensitive documents

It asks for governance maturity that many firms do not yet have

Differential privacy is not just a maths layer. It is a governance discipline masquerading as a technical feature.

Explaining the protections to end users and other stakeholders is difficult because the guarantees are not absolute and need contextualisation. This is one of the most under-appreciated barriers to ubiquity. Differential privacy demands that an organisation know what data is being used, what counts as a contribution, what is being released, who decides the privacy budget, how utility is evaluated, and who signs off when those choices carry consequences. Many firms still are not good at that level of definitional discipline.

That is why differential privacy often lands best in institutions that already think like stewards rather than extractors. Official statistics agencies, mature research environments, and large platforms with dedicated privacy infrastructure can absorb the overhead. A typical enterprise trying to move fast with fragmented data ownership usually cannot. The challenge is not only whether it can add noise correctly. It is whether it can define responsibility clearly enough to use the technology honestly.

It becomes politically hardest where it matters most

The utility loss from differential privacy can fall harder on underrepresented groups, both in private data summaries and in differentially private machine learning. In plain terms, the smaller or less represented the subgroup, the more likely the noise is to hurt the usefulness. That makes deployment especially delicate in precisely the settings where fairness, public accountability, or high-stakes decisions matter most.

This is one reason differential privacy remains easier to justify in some telemetry and aggregate analytics settings than in high-consequence operational systems. A technology does not become ubiquitous just because it is principled. It becomes ubiquitous when the trade-offs are politically boring. Differential privacy is not there yet. In many important contexts, it still makes the distribution of cost too visible.

Also Read: How to unlock possibilities through data privacy enhancing technologies

So what is really going on

The more strategic reading is this. Differential privacy was sold as a privacy solution, but it behaves more like a discipline of institutional restraint.

It forces organisations to answer questions they would often prefer to blur. What are we actually trying to learn? How much precision do we truly need? Who gets to decide the privacy loss? Which users bear more of the utility cost? What other controls still matter because differential privacy does not solve them? Those are healthy questions. They are also exactly the sort of questions that stop a technology from becoming frictionless and universal.

So the right conclusion is not that differential privacy is disappointing. It is that the market misunderstood what kind of success it was likely to have. Differential privacy was never going to become ubiquitous in the simplistic sense of appearing everywhere sensitive data appears. It is becoming something else. A serious control for specific settings where aggregate insight matters, formal guarantees matter, and the institution is mature enough to live with visible trade-offs. 

—

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 Differential privacy was supposed to solve it: Why it is not ubiquitous yet appeared first on e27.