Echelon Singapore 2026 returns on 3 to 4 June at the Suntec Singapore Convention & Exhibition Centre, bringing together startups, investors, enterprises, and ecosystem builders driving innovation across Asia. Alongside the main stage sessions and networking opportunities, the startup exhibition floor offers attendees a closer look at emerging companies developing solutions across AI automation, fintech, healthcare, accessibility, and digital engagement.

This group of startups reflects the diversity of ideas shaping the region’s next phase of growth. From agentic AI tools that streamline business operations to inclusive communication technologies and mobile-first financial platforms, these companies are building practical solutions designed to solve real operational and societal challenges. Whether attendees are exploring partnerships, investment opportunities, or new technologies, these are some of the startups to watch at this year’s event.

Julia automates back-office workflows with agentic AI

Julia is an agentic AI web app that streamlines quotes, invoices, and routine back-office workflows for SMEs and growing teams. By extracting key details, reasoning over requirements, and generating accurate documents in minutes, the platform reduces manual work, errors, and follow-ups. Julia enables sales and finance teams to move faster while keeping operations efficient and organised. At Echelon Singapore 2026, the team will connect with businesses looking to optimise internal workflows and improve productivity.

Ducket.IO transforms event engagement through Web3-powered audience intelligence

Ducket.IO is an event platform that combines a Web2-like user experience with Web3 infrastructure to help organisers better understand and engage their audiences. By tokenising tickets, the platform unlocks deeper visibility into attendee behaviour across the event lifecycle, enabling stronger community building, repeat attendance, and new monetisation opportunities. At Echelon Singapore 2026, Ducket.IO will connect with organisers and partners interested in data-driven event experiences.

Also read: 10 ecosystem players shaping how startups scale at Echelon Singapore 2026

SP Entrepreneurship Centre (SPiNOFF) nurtures student-led ventures for real-world impact

SPiNOFF is Singapore Polytechnic’s entrepreneurship centre supporting students and recent graduates in building impactful ventures. Grounded in human-centred innovation, the programme equips founders with the mindset, tools, and support needed to turn ideas into real-world businesses. By bridging education and industry, SPiNOFF enables startups to test and refine their solutions in practical environments. At Echelon Singapore 2026, the team will showcase emerging ventures and connect with ecosystem partners.

Assistive Technologies enables communication without barriers through AAC innovation

Assistive Technologies develops communication tools designed for individuals with non-verbal disabilities, redefining how people connect and express themselves. Its messaging solutions for augmentative and alternative communication (AAC) break traditional limitations, enabling more inclusive and meaningful interactions. By leveraging technology for accessibility, the company supports greater independence and participation for users. At Echelon Singapore 2026, Assistive Technologies will connect with partners in healthcare, accessibility, and inclusive tech.

SuperAgent enables smarter automation through AI-powered agents

SuperAgent builds AI-driven agents designed to automate workflows and enhance productivity across business operations. By leveraging intelligent automation, the platform helps organisations streamline repetitive tasks and improve efficiency at scale. Its solutions are built to integrate seamlessly into existing systems, enabling faster adoption and impact. At Echelon Singapore 2026, SuperAgent will connect with businesses exploring AI-driven automation.

Also read: Meet the companies taking the floor at Echelon Singapore 2026

PharmKulen enhances healthcare access through digital pharmaceutical solutions

PharmKulen is a healthcare platform focused on improving access to pharmaceutical services through digital innovation. By streamlining how patients connect with pharmacies and healthcare providers, the platform enables more efficient and accessible care delivery. Its solutions aim to bridge gaps in healthcare access while improving operational efficiency for providers. At Echelon Singapore 2026, PharmKulen will engage with healthcare partners and innovators.

WeMoney Mobile (We Gro Up Co.,Ltd) empowers financial access through mobile solutions

WeMoney Mobile is a financial technology platform designed to improve access to financial services through mobile-first solutions. By providing tools that support financial management and inclusion, the platform helps users better manage their finances and make informed decisions. Its approach focuses on accessibility, convenience, and scalability across diverse user segments. At Echelon Singapore 2026, the team will connect with partners interested in fintech innovation and financial inclusion.

Join the conversations shaping Asia’s digital future

As organisations across Asia accelerate digital transformation, startups continue to play a critical role in building the tools and platforms shaping how businesses and communities operate. The startups exhibiting at Echelon Singapore 2026 showcase emerging ideas across AI, fintech, healthcare, accessibility, and digital engagement, reflecting the region’s growing focus on scalable and impact-driven innovation.

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

Register now to join the conversation on 3 to 4 June at the Suntec Singapore Convention & Exhibition Centre.

—

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

The e27 team produced this article

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

Featured Image Credit: Canva Images

The post Startups driving AI automation, fintech, and accessibility gather at Echelon Singapore 2026 appeared first on e27.

Most founders and CEOs I speak to are not short of strategy. They know where they want to take the business, which markets matter, and what success should look like. Boards are aligned. Investors understand the ambition.

And yet, months later, very little has changed.

Decisions still take too long. Teams remain busy, but progress feels slow and uneven. Old behaviours persist, even when everyone agrees they no longer serve the business. This frustration is common in growing SMEs, where leadership time and execution capacity are stretched.

The issue is rarely the quality of the strategy itself. It is what happens after it is agreed.

When strategy and measurement lose focus

One of the most striking patterns I have seen across companies of different sizes and stages is how differently organisations define strategy. For some businesses, it is a 150-page PowerPoint deck. For others, it is three slides and a short narrative. Neither approach is inherently wrong.

The problem starts when strategy loses focus.

The same happens with measurement. As companies grow, many start measuring everything. Dashboards expand, core metrics multiply, and soon no one can tell what truly matters. When there are too many measures, focus disappears. Reducing that list is not about presentation. It is about making execution possible.

Strategy also loses traction when it is disconnected from incentives. Many organisations have sensible strategic priorities and well-defined KPIs, yet their reward structures reinforce entirely different behaviours. When strategy, metrics, and incentives are not aligned, execution stalls quietly but predictably. This is not a cultural issue. It is structural.

Also Read: 5-step strategy for agri e-commerce startups to engage customers

Why clear priorities and the right “why” drive execution

Where strategy really breaks down is in the gap between intent and reality.

Strategy is typically set at a high level. Execution happens in how decisions are made, how trade-offs are handled, and which behaviours are rewarded day to day. If those elements do not change, the strategy remains theoretical.

Before leaders even think about execution mechanics, one question matters more than most: why.

Too often, strategy conversations default to financial outcomes alone. Growth targets, valuations, acquisitions. These matter, but they are not enough. Execution improves when leaders consistently explain why the strategy matters to the organisation and what impact it should have on people inside it, not just on shareholders outside it.

At an individual level, people need to understand why change is necessary, why their behaviour must shift, and why it makes sense to commit. When that connection is missing, execution becomes compliance at best.

Cascading strategy is where many SMEs lose momentum. Leadership teams assume communication is the main task. Updates are given, slides are shared, and messages are repeated. Yet behaviour remains unchanged.

Cascading fails when the strategy stays abstract. Leaders explain what the business wants to achieve, but not what must now be different. Priorities remain vague. Trade-offs are left implicit. Each function fills in the gaps in its own way, and execution fragments.

Also Read: Achieving product-market fit: The ultimate guide to growth, strategy and positioning

How leaders turn strategy into real behaviour change

Every strategy creates different reactions. Some people buy in immediately and become champions. Others resist but can be won over. The most damaging group is quieter: those who agree publicly but undermine privately. Left unaddressed, this behaviour erodes execution far more than open disagreement.

Strong execution also depends on feedback. Teams need to be able to say when something is not working without fear. This does not mean abandoning the strategy. It means adjusting execution before problems compound. Businesses that allow honest feedback move faster and learn quicker.

Ultimately, strategy is won or lost in execution.

Most SMEs already know what they want to achieve. The challenge is not vision or intelligence. It is the willingness to confront what execution actually demands.

Strategy does not fail in boardrooms or planning sessions. It fails quietly, in the decisions leaders avoid, the priorities they refuse to narrow, and the behaviours they continue to reward despite saying otherwise.

Good strategies do not create value on their own. They only matter when leaders are prepared to make them real.

—

The post The strategy trap: Why your best plan is failing to launch appeared first on e27.

The semiconductor boom is no longer just about building more fabs; it’s about keeping the tools inside them running with near-perfect reliability. As global chipmakers pour billions into new plants and equipment, the spotlight often stays on the giants: the OEMs, the mega fabs, and the trillion-dollar supply chains.

But behind every high-performing production line is a quieter support layer of specialist SMEs. These firms refurbish critical tools, reduce downtime, and solve the operational problems that can make or break shipment schedules.

Also Read: Semiconductors at risk: The invisible threats that could break global supply chains

Singapore-based Global TechSolutions (GTS) is one such company. In this Q&A, its CEO Kenneth Lee Wee Ching, shares how the company has built a regional footprint across Singapore, Malaysia, Taiwan and the US — and why reliability, auditability, and near-site agility are becoming just as strategic as the next breakthrough node.

Edited excerpts:

Semiconductors are heading toward a trillion-dollar market, and fabs are pouring money into tools. Yet the “support layer” gets little attention. In one sentence, what does GTS do that directly protects fab revenue, and why can’t big OEMs do it as effectively?

At its core, GTS restores and upgrades front-end semiconductor tools to OEM-equivalent—or better—hardware performance, helping fabs reduce downtime, bypass long new-tool lead times, and protect shipment schedules. Large OEMs are not structured for the same near-site agility, deep customisation, or selective execution model where we only take on work we can certify to OEM-level outcomes with warranty. That focus is why our work sits so close to protected fab revenue.

For founders and VCs, defensibility matters. In semiconductor equipment services, what is your moat?

Our defensibility stems from a combination of breadth of capability, execution discipline, and customer proximity. GTS is among the few regional players offering a full suite of new equipment, refurbishment, upgrades, and field engineering—supported by cleanroom-certified facilities and test platforms that simulate real fab environments.

Our footprint across Singapore, Malaysia, Taiwan, and the US enables near-site response and supports a “close-to-customer” strategy that holds up even when supply chains tighten. We pre-stage critical spares, run parallel testing, and compress time-to-qualification without compromising performance.

Equally important is know-how. We maintain proprietary jigs, fixtures, firmware, and automated test routines developed in-house. We also deliberately decline work if we cannot meet OEM-equivalent standards. That discipline preserves trust and yields for customers.

Together, these capabilities address cost, lead time, customisation, and sustainability — while reinforcing defensibility in an industry where reliability, repeatability, and auditability are non-negotiable.

What does “reliability” mean in numbers? Which metrics matter most to customers, and what improvement ranges are realistic?

We define reliability using metrics that production and finance teams already care about. These include Mean Time Between Failures (MTBF), time-to-qualification (the speed at which a tool is released back to production) and chamber-level performance indicators such as thermal uniformity, vacuum stability, and gas-flow calibration, all of which underpin line yield.

Also Read: Indonesia courts Nvidia and AWS as it eyes a bigger role in global chip supply chains

In advanced-packaging-relevant lines, customers have seen roughly a 15 per cent reduction in downtime and around a 7 per cent improvement in line-yield stability following refurbishment and targeted upgrades. We also treat documentation completeness as a KPI. Clean, ISO- and SEMI-aligned documentation with traceable test logs shortens audits and keeps production lines compliant.

Trust is the real currency in semiconductors. How did GTS win its first serious customers?

Trust in this industry is built incrementally. Qualification standards are stringent, and introducing a new partner involves lengthy approval cycles. We started with smaller scopes and incremental improvements, then expanded into more complex parts and equipment only after demonstrating consistent results.

That caution is necessary: with thousands of steps in chipmaking, minor errors can cascade into significant losses. We also operate under strict SOPs and controlled environments, including Class 100 and Class 1,000 cleanrooms that mirror fab conditions. This reassures customers that our processes behave predictably when deployed on production lines.

Over time, that translated into a track record of high performance at optimised cost—combined with faster turnaround and greater customisation than traditional OEM approaches. Customer retention and referrals followed naturally.

How do you operate through uncertainty, like export controls, shifting trade rules, audits, and supply disruptions, without breaking delivery promises?

Semiconductors are deeply intertwined with geopolitics. Export controls, tariffs, and regulatory shifts often translate directly into supply-chain disruptions. To manage this, GTS built a “global supply chain mirroring” approach years ago.

We maintain engineering presence, parts strategy, and execution capability close to where customers operate, while aligning closely with them on technology roadmaps and requirements. Where appropriate, this allows us to localise execution and rely on locally available parts rather than a single cross-border supply route.

When sudden policy changes occur, this resilience prevents disruption from becoming downtime. Even when the “cleanest path” is no longer available, our proximity and documentation discipline allow us to align on acceptable alternatives with customers and keep execution controlled, predictable, and auditable.

Are fabs shifting from break-fix to predictive reliability engineering? What must SMEs build to stay relevant?

Yes, the shift is underway, especially in high-mix lines where advanced packaging intersects with front-end steps. Maintenance is moving from reactive break-fix toward predictive diagnostics and reliability engineering.

For SMEs, relevance requires a capability stack that includes high-quality data capture, component-level cleanroom testing, predictive diagnostics tied to known failure modes, and disciplined teardown-to-QA loops. Just as necessary is documentation that integrates cleanly into fab workflows so insights translate into approvals and action.

Also Read: Thailand enters the chip race, without challenging Singapore head-on

At GTS, we’re investing in software-driven diagnostics and fault prediction. We are developing systems that learn from troubleshooting patterns and equipment signals (such as vibration, motion anomalies, and acoustic changes) to detect early warning signs. In parallel, we’re standardising knowledge-sharing across regions so improvements in one site can be replicated quickly elsewhere without compromising quality controls.

How are AI and advanced packaging changing customer demands, and what must SMEs build by 2026?

AI workloads and advanced packaging are raising expectations across the board. We’re seeing demand for tighter thermal control and film uniformity, more line-specific modifications instead of generic upgrades, faster ramp-to-production timelines, and deeper metrology and certification discipline.

To stay relevant through 2026, SMEs need modular upgrade paths, cleanroom testing capacity, predictive diagnostics, and documentation that is export-control-ready and audit-friendly. Our focus remains on improving chamber-level reliability with auditable performance, so innovation reaches production with stability—not just speed.

At the same time, software-driven capabilities will become increasingly important, enabling customers to shift from reactive fixes to earlier, proactive interventions as tolerances tighten.

From a scaling perspective, what breaks first when an SME expands across countries? What do you standardise, and what stays local?

The first thing that breaks is consistency of execution—not technical skill, but how reliably teams diagnose issues, control variation, and sign off outcomes under pressure. Small differences in training or test interpretation can create big swings in customer confidence.

To prevent this, we standardise the “spine” of delivery across all regions: ISO- and SEMI-aligned quality systems, refurbishment and test protocols, structured documentation and sign-off gates, training pathways, and parts qualification strategies. This ensures predictable quality without reinventing processes at each site.

What remains local is how we integrate into each customer’s operating reality—site-specific compliance requirements, fab conventions, and coordination with local stakeholders. The goal is a common engineering playbook with local fluency.

Looking to 2026, what’s your base case and contrarian view for the semiconductor services ecosystem? Who wins?

Our base case is that advanced packaging continues to scale and fabs increasingly prioritise predictability—uptime, faster qualification, and auditability—over pure capex expansion. Service partners that win more scope will be those that can consistently return tools to certified performance and prove it with clean documentation.

Also Read: ‘The future of semiconductor manufacturing is regional’: Global TechSolutions CEO

The contrarian view is that even when equipment is available, execution friction becomes the real bottleneck. Compliance overhead, export controls, and specialist talent shortages may matter more than hardware availability. In that environment, speed alone doesn’t win; controlled agility does—the ability to move fast while remaining auditable and safe.

Across both scenarios, the biggest pinch points will be human capital, compliance, and critical spares. Talent is particularly challenging given the industry’s complexity, which makes ecosystem-building and collaboration essential.

On winners, it’s not simply niche specialists versus scaled platforms. It’s whichever model can repeatedly prove outcomes—performance, auditability, and production stability—again and again.

The post The quiet layer keeping the chip boom alive appeared first on e27.

We are living through the rise of micro-businesses.

A decade ago, building a company meant hiring a team, finding capital, building infrastructure, and waiting months, sometimes years, to validate whether the market even wanted what you were selling.

Today, that timeline has collapsed.

A founder can launch an e-commerce store in a day. A creator can monetise an audience with a single product. A consultant can package expertise into digital programmes. A builder can launch a micro-SaaS with AI tools and no traditional technical team.

The modern business landscape has shifted from scale-first to speed-first.

And nowhere is that shift more visible than in Southeast Asia’s MSME ecosystem.

Micro, small, and medium enterprises have always formed the backbone of regional economies, but technology has fundamentally changed how they operate. Social commerce, live selling, creator-led commerce, and AI-assisted businesses have accelerated the ability for individuals to start faster than ever before.

But in this new economy, one thing has not changed: Trust remains the currency of business.

In fact, for smaller businesses, trust may be the infrastructure itself.

Recently, entrepreneur Shawn Yeo highlighted a case involving a seller whose public platform rating fell sharply after a cluster of repeated low-rated reviews from a single customer over a short period of time.

Whether the reviews were justified is not the point.

Whether the customer was genuinely dissatisfied is not the point.

The real issue is structural: Should one customer interaction, however negative, carry enough system weight to materially affect the viability of a business?

That question matters far beyond one seller.

Because as more founders build leaner, faster, and smaller businesses, the systems that govern trust are becoming just as important as the systems that govern payments, logistics, and traffic.

And increasingly, those trust systems are algorithmic.

The new economy has lowered the barrier to building, but not the cost of trust

One of the most overlooked shifts in entrepreneurship today is this: It is easier than ever to build. But it is not easier to earn trust. If anything, it is harder.

Consumers are overwhelmed with options. Markets are noisier. Competition is denser.

And because of that, trust signals have become shortcuts. Ratings. Reviews. Social proof. Comments. Public sentiment.

These signals help buyers make faster decisions. That is useful. But it also creates dependency.

For MSMEs, especially those built on social platforms, trust signals are no longer just social validation. They are operational assets.

• A lower rating can affect discoverability.
• A lower rating can affect conversion.
• A lower rating can affect partnership opportunities.
• A lower rating can affect affiliate privileges.
• A lower rating can affect cash flow.

This is especially true in social commerce ecosystems where the algorithm decides visibility. And visibility, in digital commerce, is survival.

That changes the weight of reputation entirely. For large corporations, reputation damage is painful. For micro-businesses, it can be operationally destructive. That difference matters.

Also Read: Singapore’s digital asset market grows up: Why trust and discipline now trump momentum

Customers should always have the right to complain

To be clear: Customers deserve the right to voice dissatisfaction. That should never be removed.

Feedback is part of market accountability. It is how businesses improve. It is how standards rise. I have personally left negative reviews before — not to punish, but to reflect an actual experience.

Usually, because there was poor service. Or poor response. Or no response. That is valid. That is healthy. A trust system without criticism is not a trust system. It is marketing.

But there is a line between customer feedback and structural over-amplification. A review should reflect an experience. Not become a disproportionate threat.

That distinction becomes critical when platforms use trust as part of business infrastructure. Because once trust affects access, visibility, and monetisation, review systems are no longer passive.

They become economic mechanisms. And economic mechanisms require better design.

Platforms are no longer marketplaces — they are trust engines

This is where the conversation becomes more nuanced.

Platforms today do far more than facilitate transactions.

• They shape perception.
• They determine visibility.
• They influence conversion.
• They govern access.
• That makes them trust engines.

And trust engines carry responsibility.

The challenge is that human emotion moves faster than context.

A customer has one bad experience. They react emotionally. They leave a harsh review. That is human.

But when systems fail to contextualise patterns — frequency, repetition, anomalies — that emotion can become disproportionately amplified.

And that is not always fair to either side. Not because customers are wrong. But because systems may be too simplistic.

Trust systems often assume equal weight across actions.

But human behaviour is rarely equal. Repeated review patterns. Emotional clustering. Behavioural inconsistency. These are signals. And signals can be understood better.

Which brings us to AI.

AI’s next big role may not be productivity — it may be fairness

Most founders talk about AI in terms of growth.

• How to automate content.
• How to reduce costs.
• How to scale customer service.
• How to build products faster.

All valid.

But one of the most underrated applications of AI is trust architecture.

Also Read: From fraud fighters to zero-trust builders: SEA’s cyber stars

AI is uniquely positioned to improve how trust systems operate because it can process patterns humans often miss. Not to replace human judgment. But to strengthen it.

Imagine a system that could detect:

• Whether multiple reviews come from one unusually concentrated pattern,
• whether review sentiment is behaviourally inconsistent,
• whether customer feedback reflects product quality or emotional escalation,
• and whether anomalies should trigger manual review before affecting seller privileges.

That is not censorship. That is context. And context creates fairness.

We already trust AI to detect fraud. We trust AI to identify spam. We trust AI to detect unusual financial activity. Trust systems should evolve, too. Especially in an economy increasingly powered by micro-businesses.

Because if AI can help people build businesses faster, it should also help protect the integrity of how those businesses are judged.

Reputation has always been fragile, but community changes the equation

As founders, we know reputation is fragile. But we also know something else: People forget. Public criticism, while painful, is rarely permanent.

There is an old PR saying: All publicity is good publicity. Not always true. But visibility does create familiarity. And familiarity creates memory.

I have experienced this firsthand. I run ads for workshops, programmes, and educational products.

And like many founders who market publicly, I get comments from people who have never attended my classes or purchased my offers. “Scam.” “Fake guru.” Criticism about how I speak. How I look. How I present.

People forming opinions without ever experiencing the actual product. It happens.

And while that is part of being visible, it reinforces something important:

• Public opinion is often shaped by proximity, not truth.
• The people closest to your work know its value.
• The people who are furthest often make the loudest assumptions.
• This is why community matters.

A strong community becomes your defence layer.

If enough people trust you, enough people speak for you. And that changes everything.

Trust is no longer platform-dependent. It becomes people-dependent. That is far more resilient.

Founders must build owned trust, not rented trust

This is the founder’s lesson. Platforms can distribute your business. But they should never fully define your business.

Traffic can be rented. Trust should be owned. That means building:

• Your email list,
• your CRM,
• your community,
• your direct customer relationships,
• your repeat buyer systems.

Also Read: Building trust in turbulent times: The new security paradigm for crypto exchanges

Too many founders optimise for traffic. Not enough optimised for trust continuity.

And in today’s market, trust continuity is the real moat. Especially for MSMEs. Especially for solo founders. Especially for AI-powered micro-businesses.

Because the future of entrepreneurship is leaner. Smaller teams. Faster launches. Higher automation. Lower operational cost.

But also: Higher reputational sensitivity.

That is the tradeoff.

The rise of micro-businesses means trust systems must evolve

As AI continues lowering the barrier to entry, we will see more micro-businesses emerge than ever before. One person can now build what used to require teams. I know this firsthand.

AI has accelerated my ability to build, execute, automate, and deploy ideas faster than traditional structures ever allowed. That is the opportunity of this era.

But speed without resilient trust systems creates fragility. And fragility is dangerous in founder ecosystems.

Customers deserve a voice. Businesses deserve fairness. Platforms deserve accountability.

And AI may be one of the strongest tools available to help create a better balance between all three.

Because in the age of micro-businesses, trust is no longer just a branding asset. It is an operational infrastructure.

And if we are building the future of commerce on digital trust systems, then those systems need to become smarter, fairer, and more context-aware.

Because trust takes years to build. And for micro-businesses, one flawed system can damage them overnight.

—

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 post Trust takes years to build but one flawed system can damage a micro business overnight appeared first on e27.

AI agents are increasingly presented as the next leap in industrial transformation — systems that can move beyond analysing data to making decisions and taking action autonomously.

In theory, they promise a future where manufacturing becomes faster, smarter and more adaptive.

But in food manufacturing, there is a harder truth we need to confront first: AI cannot fix processes we do not yet fully understand.

Manufacturing is not a stable system

This is especially evident in food systems, where manufacturing is far from stable.

Unlike highly controlled digital environments, food production deals with biological raw materials that are inherently variable. Moisture content shifts with storage conditions. Protein functionality changes depending on source and prior processing history.

Small differences in formulation or temperature can lead to significant changes in final product quality.

Take extrusion, for example — a process commonly used to produce puffed snacks and plant-based protein products.

A successful outcome depends on balancing moisture, temperature profile, screw configuration and ingredient behaviour with precision. When conditions align, the product expands and forms as intended. When they do not, the result may collapse, become dense, or fail to form the desired structure.

These are not rare anomalies.

They are part of the everyday reality of manufacturing.

The promise of AI — and what it assumes

In my own work at SIT, including pilot-scale trials at FoodPlant, I am often asked whether AI can be used to predict outcomes or recommend processing conditions in extrusion.

It is an understandable question. If enough production data is collected, it seems reasonable to expect that AI should be able to identify patterns and optimise performance.

Also Read: Beyond the buzz: How AI and sustainability are reshaping design, manufacturing, and construction in APAC

In principle, this promise is compelling.

It suggests a shift from trial-and-error towards more predictive, data-driven manufacturing.

But this vision rests on a critical assumption: that the data available fully captures how the system behaves.

In food manufacturing, that assumption rarely holds.

Where the gap lies

AI systems can only learn from what is measured.

Yet some of the most influential variables in food processing — such as how materials behave under heat, pressure or shear — are not always directly observed or consistently recorded. Many process interactions remain tacit, built through experience rather than explicit data capture.

Even when data exists, relationships between variables are often non-linear and context-dependent.

The same processing condition can produce different outcomes depending on formulation, material history, or environmental conditions.

What AI receives, therefore, is often only a partial and unstable representation of reality.

When AI performs poorly in such settings, the conclusion is often that the technology is not mature.

In many cases, the issue lies elsewhere.

We are asking algorithms to optimise systems that remain insufficiently characterised.

More data is not the same as a better understanding

There is growing emphasis on shared datasets, digital toolboxes and industrial AI platforms.

These are important developments — but more data alone does not resolve the underlying challenge.

If variables are defined differently, measured inconsistently across facilities, or recorded without a common structure, combining datasets does not improve understanding.

Also Read: Costing comparison of top 7 popular ERP software for food manufacturing in Singapore

It amplifies inconsistency.

A meaningful dataset — much like a well-designed dashboard — reflects a clear understanding of what variables matter and how they relate to outcomes.

Without that structure, aggregating more data does not lead to better insight.

It simply scales the same limitations.

Why this matters now

These questions extend beyond manufacturing efficiency.

For Singapore, they are becoming increasingly relevant as food resilience rises on the national agenda. Recent geopolitical tensions and disruptions to global supply chains have once again highlighted how vulnerable food systems can be under external shocks.

Singapore imports more than 90 per cent of its food.

In such a context, resilience cannot be defined only by where supply comes from.

It must also include our ability to convert available inputs into stable, nutritious and scalable food products locally.

That capability is a resilience multiplier.

What needs to be built

AI can play an important role in this future.

It can accelerate learning, improve consistency and help detect patterns that are not immediately obvious.

But AI is not the foundation.

Before autonomous systems can make reliable decisions, manufacturing systems must first become more observable, more structured and better understood.

This means:

• Better characterisation of material behaviour
• Clearer definition of operating windows
• More consistent ways of capturing process–material interactions

Also Read: Anomaly Bio powers the future of ingredient manufacturing with US$2.6M in pre-seed funding

Only then can AI move beyond pattern recognition into dependable decision support.

Beyond automation: Building real capability

The future of intelligent manufacturing will not be built by algorithms alone.

It will be built on a deeper mastery of process.

Until we close that gap, AI will not transform manufacturing. It will simply make visible how much of it we still do not fully understand.

—

The post AI won’t fix manufacturing, until we fix our understanding appeared first on e27.