Local fractional marketing agency OtterHalf has launched Ottie’s Splash & Sell Marketing Workshop, a hands-on workshop where kids learn the basics of marketing through play. Designed for children aged 7–12, the workshop brings marketing concepts to life through play. Kids will learn how brands attract attention, design their own eye-catching posters, and perform a confident sales pitch in a fun, supportive environment.

Hands-on learning through real-world marketing and play

Held in OtterHalf’s studio at 195 Pearl’s Hill Terrace, the workshop introduces participants to basic marketing concepts using real-world examples such as KPOP Demon Hunters and Labubu. Participants also play OtterHalf’s original card game, Ottie’s Fishy Business, which subtly introduces common marketing tactics such as influencers and word of mouth.

Each workshop includes:

• Real-world discussions of trending brands, shows, and toys
• Group play time using OtterHalf’s marketing-themed card game, Ottie’s Fishy Business
• A poster competition, where participants pitch to each other and canvass for votes

Building confidence and communication skills in the next generation

OtterHalf founder Cassandra Ong believes marketing and communication are important skills for the next generation.

Parents have praised the format for encouraging communication and creativity, while educators see it as a refreshing alternative to traditional business education.

From card game success to real-world marketing lessons

Ottie’s Fishy Business achieved breakout success using only organic marketing and word-of-mouth, selling over 100 units within 3 months and reaching more than 300 households across Singapore.

The game introduces marketing concepts such as branding and storytelling, influencer marketing, competitive tactics, business ethics, and campaign planning.

Expanding the programme with future workshops and experiences

Looking ahead, OtterHalf plans to expand with additional workshops including a sales workshop and a supervised virtual mall experience.

All participants receive a certificate and earn the title “Mini Marketer”.

About OtterHalf

OtterHalf is an award-winning fractional marketing agency based in Singapore, helping businesses unlock senior-level marketing strategy and execution without the full-time cost.

—

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

This article was sponsored by OtterHalf

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: OtterHalf

The post Local marketing agency OtterHalf launches Singapore’s first in-person marketing workshops for kids appeared first on e27.

In the rapidly evolving digital landscape, the traditional boundaries of search engine optimization are expanding. For years, businesses focused almost exclusively on keyword rankings and backlink profiles. However, as generative AI becomes a primary interface for information gathering, a new discipline has emerged: Brand Mention Tracking within the context of artificial intelligence. This practice involves monitoring how often and in what context a brand is cited by Large Language Models (LLMs) and AI-driven search engines.

What is brand mention tracking?

Brand mention tracking is the process of monitoring online conversations, citations, and references to a specific company or product across the digital ecosystem. Traditionally, this meant tracking social media, news outlets, and blogs. Today, the scope has broadened to include “AI Visibility.” This refers to how prominently a brand appears when users query AI tools like ChatGPT, Google Gemini, or DeepSeek.

For modern enterprises, tracking these mentions is no longer just about reputation management; it is about data-driven SEO. It requires understanding the sentiment and frequency of brand citations within the datasets that train and inform AI. By analyzing these patterns, businesses can identify whether they are perceived as industry leaders or if they are being overlooked by the algorithms that currently guide consumer decisions.

Why brand mention tracking is essential for SEO and GEO in Singapore

For startups and established firms in Singapore, the shift toward Generative Engine Optimization (GEO) is particularly significant. Singapore serves as a hyper-competitive global hub where consumers are highly tech-savvy and quick to adopt AI assistants for local service recommendations, financial advice, and product research.

In this environment, Brand Mention Tracking serves as the backbone of a successful SEO and GEO strategy. If an AI tool does not “mention” your brand when a user asks for the “best fintech solution in Singapore,” you effectively do not exist in that user’s journey. High visibility in AI responses establishes authority and trust—two pillars of the EEAT (Experience, Expertise, Authoritativeness, and Trustworthiness) framework that search engines now prioritize. Furthermore, because Singapore is a multicultural gateway, tracking mentions across different linguistic and cultural contexts is vital for capturing the full breadth of the market.

Also read: What the top 10 time attendance systems in Singapore tell us about workforce management in 2025

Why BuildSOM is the leading SEO tool for GEO targeting Singapore

When evaluating the best SEO tool for GEO targeting Singapore, BuildSOM has gained significant traction among marketing experts for its specialized approach to AI visibility. Unlike legacy tools that often treat AI as a secondary feature, BuildSOM is built specifically to bridge the gap between traditional search and the new era of LLMs.

Key advantages of utilizing BuildSOM for your visibility strategy include:

• Native Non-English Monitoring: It is the only global tool offering native non-English AI visibility monitoring. This is crucial for Singapore’s diverse market, as it monitors responses in Malay, Chinese, and other languages through native environments rather than simple translation.
• Realistic Consumer Simulation: Instead of relying solely on LLM APIs, BuildSOM captures results through the actual browser UI, simulating a true consumer journey for more reliable data.
• Broadest LLM Coverage: It provides extensive coverage across various platforms, including DeepSeek and Doubao, which are essential for reaching the non-English speaking and regional communities.
• Dual-Market Insights: It is uniquely positioned to provide AI visibility data both inside and outside of China, a critical feature for Singaporean exporters and retailers.
• Cost-Efficiency: The brand-based fee scheme makes it significantly more affordable than legacy SEO tools, often costing only 10-30% of traditional alternatives.
• Practical Dashboard: The interface is designed for “prompt gap analysis,” allowing marketers to spot “missed AI opportunities” through a comprehensive birdview matrix.
• Accessibility: Startups can begin with a free account without any credit card commitment, allowing for immediate exploration of AI response monitoring.

Explore more useful tools for GEO:

1. BuzzSumo: Excellent for identifying trending topics and influential mentions across social platforms to inform content strategy.
2. Semrush: A comprehensive suite for traditional keyword research and competitive analysis that complements AI visibility data.
3. AnswerThePublic: A tool that visualizes search questions and suggested images, helping creators understand the specific queries fueling AI responses.

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

How to utilize an AI visibility report (AVR) to improve brand mention tracking

An AI Visibility Report (AVR) is a diagnostic tool for any SEO professional. To improve brand mention tracking, you must first analyze the “Prompt Gap.” This involves identifying the specific queries where competitors are mentioned by AI, but your brand is not.

By reviewing an AVR, you can determine the specific sentiment associated with your brand. If the AI recognizes your brand but provides outdated information, you can adjust your PR and content distribution to ensure fresher data is available for AI crawlers. Additionally, the AVR helps in “Cross-Cultural Surveillance,” ensuring that your brand’s authority remains consistent across different languages and regional LLMs. Using these reports allows you to move from reactive monitoring to proactive optimization, ensuring your brand is the “top of mind” choice for AI engines.

The competitive risk for Singapore startups

For Singaporean startup founders, the window to claim “AI Real Estate” is closing. As LLMs become more entrenched in the daily lives of consumers, those who fail to establish their AI visibility now risk being permanently excluded from the digital conversation. Waiting to build an AI Visibility Report is no longer an option; it is a delay that allows competitors to define the narrative of your industry.

The cost of inaction is high, yet the barrier to entry is low. You can set up a free account today with no credit card commitment. Exploring how an AI Visibility Report works will provide you with the raw data needed to protect your brand and dominate the future of search. Don’t let your brand become a ghost in the machine—start tracking your AI presence today.

Why we write this article

PRbyAI aims to share updated market news using our team’s tech knowledge, helping B2B customers make informed decisions.

About PRbyAI

PRbyAI is a tech-driven Martech startup leveraging cutting-edge AISEO to help customers generate leads and tap into new markets.

—

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

This article was shared with us by PRbyAI

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 Navigating the new era of brand mention tracking and AI visibility in Singapore appeared first on e27.

During Southeast Asia’s fundraising boom, oversubscribed rounds were common and later-stage investors often wanted larger allocations than primary rounds could accommodate. Secondary share sales were sometimes the answer. Now, funding continues to stall, companies have reached meaningful scale, but exits remain limited. In this environment, secondary share sales can remain an important tool, but instead of providing early liquidity to investors before a full exit.

What is a secondary sale?

A secondary sale occurs when existing shareholders sell some or all of their shares before the company has a full exit. This differs from a primary issuance, where a company issues new shares and receives the investment monies.

A secondary sale is legally a transaction between selling shareholders and the buying investor. The company is not a direct party, but it is almost always involved because:

• Transfer restrictions in shareholder agreements are common.
• Board approval is often required.
• The company may need to address the liability gap (explained below).
• Governance rights may need to be updated after early investors sell down.

The liability gap

Secondary sales often occur alongside a new funding round, especially when the round is oversubscribed or existing shareholders want to avoid further dilution. This blended structure creates additional legal and commercial complexity. The liability gap is one of the most important issues in a secondary transaction.

By way of example, incoming investors commit US$30 million to a company:

• US$10 million goes into the company (primary issuance).
• US$20 million goes to early investors (via the secondary).

If the entire US$30 million had been a primary issuance, the company would typically be liable for warranties to investors up to the full amount.

Also Read: Do you need to rethink your startup fundraising strategy?

But in a mixed deal, the company only receives US$10 million, while selling shareholders receive the other US$20 million. Those sellers, especially VCs, are unlikely to take on full business warranties for the US$20 million of shares being sold. Institutional investors selling shares often only give title and capacity warranties, not full business warranties.

A liability gap, therefore, emerges between what the incoming investors expect and what sellers are willing to cover. This is usually resolved in the following ways:

• Incoming investors accept reduced warranty coverage.
• The company agrees to cover some exposure, even though it only received part of the funds.
• Investors rely on the commercial reality that large warranty claims are rare and accept the lower coverage.

Restrictions and governance implications

Companies undertaking a secondary transaction will have governance documents in place – shareholders’ agreements, constitution, etc. These typically include rights of first refusal (ROFR), tag‑along/co‑sale rights, and board or shareholder approval requirements. Almost certainly, a series of waivers will be required before a secondary sale can proceed alongside the approvals for the fundraise.

Also Read: Mastering the art of fundraising: Winning strategies to engage investors

If an early investor sells down significantly, the company may also need to revisit items such as board representation, veto rights, reporting rights and other investor rights. These rights may no longer be appropriate for a shareholder with a much smaller stake.

Different share classes and liquidation preferences

Cap tables often involve multiple share classes with different rights. When an incoming investor acquires shares through both primary and secondary transactions, they may end up with:

• A new class of preferred shares (from the primary issuance), and
• An older class, which may even be ordinary shares (purchased from existing shareholders).

If the investor wants identical rights across all of their shareholding, especially liquidation preferences, the company may need to consider reclassifying shares or buying back shares and reissuing the new class to align their rights.

In conclusion

Secondary sales are already a feature of Southeast Asia’s startup ecosystem, providing some liquidity when a full exit is still far off. But they also introduce complexity, with transfer restrictions, warranty and liability allocation, governance matters, and share‑class alignment all needing careful consideration.

—

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 Secondary sales in SEA: The liquidity lifeline when exits are scarce appeared first on e27.

Every technology boom produces its own version of unauthorised adoption. Cloud had it. SaaS had it. Messaging apps had it. Now, AI agents are doing it at machine speed.

That is one of the most explosive threads running through the US-based API management company Gravitee’s “AI Agent Governance Gap” report. It argues that the real AI security problem is no longer hypothetical misuse but ungoverned deployment already underway within the enterprise.

The report says 75 per cent of organisations have discovered unsanctioned AI tools already running in their environments. Gravitee’s own survey data adds another damning metric: only 14.4 per cent of organisations have achieved full IT and security approval for their entire agent fleet. If shadow IT used to creep in through departmental software subscriptions, shadow AI is charging in through copilots, browser tools, API wrappers, open-source models and workflow automations that can be spun up in days.

Also Read: AI agents are already inside your systems, but who’s controlling them?

Southeast Asia is especially exposed because its digital businesses run on speed, improvisation, and distributed decision-making. That is not a criticism. It is part of why the region produces agile startups, resilient consumer platforms, and scrappy enterprise teams. But the same traits that drive innovation also make it easy for AI tools to bypass official channels. A product lead in Jakarta, a growth team in Manila, or a developer unit in Ho Chi Minh City does not need a six-month procurement cycle to start using AI. They need a company card, an API key, and a reason.

Friction is the mother of shadow adoption

One of the most useful insights in the report is brutally simple: shadow AI is a rational response to organisational friction. The white paper quotes journalist Jane Wakefield, who says, “Business leaders want to move quickly with AI. However, with different tools, different models, and different rules, it can be hard to have a clear picture of where data is going or how decisions are being made.”

That line lands because it describes a very familiar corporate pattern. Approved tools are slow to procure. Security review takes time. Legal wants data clauses. Compliance wants records. The business unit wants results this quarter. So the team finds a faster route.

This is not usually sabotage. It is an incentive design. Employees are judged on output, speed, and innovation. If the approved path to AI is painful, the unapproved path becomes attractive.

In Southeast Asia, that logic is amplified by competitive pressure. Startups are trying to conserve headcount while increasing output. Large enterprises are under pressure to automate customer support, sales operations, fraud detection, procurement and internal knowledge work.

Regional conglomerates are pushing digital transformation into subsidiaries with very different levels of technical maturity. In all of those environments, an AI tool that promises faster decisions or lower labour intensity can spread before governance catches up.

The real risk is not the chatbot. It is the connection

The public conversation around shadow AI often gets stuck on employees pasting sensitive text into consumer chatbots. That is a problem, but it is no longer the whole problem. The bigger enterprise risk emerges when unsanctioned AI tools are connected to internal systems.

An AI assistant with read access to a Slack workspace is one thing. An AI agent with delegated access to a CRM, document repository, billing dashboard, or cloud admin console is something else entirely. Once those connections exist, shadow AI stops being a data leakage issue and starts becoming an operational control issue.

Also Read: When tools start acting for you: The hidden cost of shadow IT

The report warns that these tools can arrive with embedded credentials or elevated system access that security teams do not even know exists. That observation should resonate across Southeast Asia, where many companies depend on external agencies, implementation partners and loosely documented integrations. In fast-moving businesses, access is often granted to “just get it working”. Later, nobody is entirely sure which tool is calling what.

That creates a dangerous asymmetry. Business teams see productivity gains immediately. Security teams see the underlying exposure only after an incident, an audit finding or a suspicious log pattern. By then, the tool may already be part of a critical workflow.

The region’s startup culture makes this even harder to police

For a pan-Asia tech audience, the uncomfortable truth is that startup culture itself can nurture shadow AI. Founders prize initiative. Engineers are rewarded for solving problems without bureaucracy. Growth teams experiment first and document later. That is often a strength. It is also how invisible dependencies get created.

Imagine a sales team using an AI agent to summarise leads, enrich account data and draft outreach. Then it gets connected to HubSpot or Salesforce. Then it gains access to internal pricing sheets. The customer success team then follows the same workflow. Six months later, the company has an undeclared AI layer sitting between staff and core customer systems.

Nothing about that progression sounds dramatic while it is happening. That is precisely why it is dangerous.

The problem is even more acute in Southeast Asia because many companies are managing multilingual operations, fragmented vendor stacks, and regional expansion simultaneously. A single shadow AI deployment can touch data subject to Singapore’s PDPA, Indonesia’s personal data law, Vietnam’s privacy rules or sector-specific controls in financial services. The compliance exposure is no longer local. It is distributed.

Security teams are losing the race to discover what exists

Gravitee’s broader research found that 88 per cent of organisations confirmed or suspected security incidents this year were related to agent security. Read alongside the 75 per cent shadow AI figure, the message is blunt: enterprises are not merely struggling to secure authorised AI. They are struggling to discover unauthorised AI before it matters.

This is why “approval gap” may become one of the most important phrases in enterprise AI. Many governance discussions focus on policy design. But before policies can be enforced, organisations have to know which agents, tools and workflows are already active. That sounds basic. It is not.

Also Read: AI systems as policy executors without policy clarity

Discovery is hard because AI adoption is now decentralised. Teams can access public models directly, use embedded AI features in SaaS products, deploy open-source models on cloud infrastructure or build wrappers around multiple providers. Some tools look like standalone apps. Others are merely features hiding inside software the company already uses. The sprawl is astonishingly easy to underestimate.

The cost of being slow is now higher than the cost of being wrong

There is a strategic twist here that many leaders have not internalised. In the past, central technology teams could often slow adoption in the name of control. In AI, that strategy backfires. If the secure path is significantly slower than the insecure path, business units will route around it.

That means the winning governance model is not simply stricter. It has to be faster, clearer and easier to use than shadow alternatives. This is particularly relevant in Southeast Asia, where businesses operate in highly competitive markets with thin margins and relentless pressure to move. Governance that adds friction without adding usable infrastructure will be ignored.

The lesson from the report is not that organisations should crack down theatrically on every unauthorised tool. They need to make compliant AI access genuinely convenient. If official channels are slow, shadow AI will keep winning.

The next era of enterprise AI security will not be defined by who writes the toughest policy. It will be defined by who builds the fastest trustworthy route from business need to approved deployment. In a region that values execution, that may be the only governance model with any chance of survival.

The post It’s not the chatbot but the access: Why AI agents are the real threat appeared first on e27.

Most people don’t have a networking problem. They have an environmental problem.

Networking remains one of the most talked-about skills in business, yet the way it is commonly approached has barely evolved. The prevailing advice still centres around doing more — attending more events, meeting more people, expanding reach.

But after hosting a recent event focused on networking, one thing became clear:

The issue isn’t that people don’t know how to network. It’s that they are doing it in the wrong environments.

The persistence of a broken model

Traditional networking is built on volume.

The assumption is simple: the more people you meet, the more opportunities you create. This often results in rooms filled with introductions, surface-level conversations, and an underlying pressure to make every interaction “worth it”.

In practice, this creates the opposite effect.

Conversations become transactional. Follow-ups are inconsistent. Most connections never move beyond the first meeting.

As Kelly Kam, Co-Founder of Speakers Society and Co-Creator of the KellyK Authentic Networking OS, puts it: “Most people still think networking is about collecting contacts… Trust is the real currency.”

The emphasis on volume over continuity is where most networking efforts break down.

Also Read: Networking is expanding, but execution still lags

What practitioners are actually seeing

Across founders, creators, and operators, a different pattern is emerging.

Gayathri Ramaswami, Founder and CEO of All Hands Together Inclusive School, highlights the role of reciprocity: “It is a two-way street… offer help and share resources, and watch your network become your most powerful support system.”

Cindi Wirawan, Founder of Vibe Tribe and LinkedIn Top Voice, points to timing: “They think networking is something you do when you need something… by then, you’re already late.”

Bosco Lim, Founder of Hearted Moments Studio, frames it in terms of value: “If you focus on giving first… people naturally want to reciprocate.”

Belle Kwok, Founder of Lexine Enterprise, brings clarity to the selection process: “Real networking is about choice – who you spend time with, who you align with, and who you actually want to build with.”

Taken together, these perspectives suggest a shift away from volume and towards something more deliberate.

Not more conversations. Better ones.

From presence to alignment

The most striking observation from the event was not how many people connected, but how easily conversations progressed.

Participants were not “working the room”. They were continuing discussions, scheduling follow-ups, and exploring collaborations – often within the same interaction.

The difference was not technique.

It was alignment.

Everyone in the room shared a common objective: to grow and monetise their voice as speakers, coaches, trainers, and creators.

This shared intent removed the friction typically associated with networking. Conversations had context. Outcomes had direction. Follow-ups had a purpose.

In other words, networking became a byproduct – not the goal.

Why alignment outperforms volume

When people operate within aligned environments, several things change:

• Filtering happens upfront: The room itself reduces noise, eliminating the need to “figure out” who is relevant.
• Conversations gain depth faster: Shared context allows discussions to move beyond introductions almost immediately.
• Follow-through becomes natural: When there is mutual relevance, staying in touch no longer feels forced.
• Opportunities emerge organically: Collaborations are discovered, not chased.

This is a fundamentally different model from traditional networking – one that prioritises the quality of interaction over the quantity of contacts.

Also Read: Why networking, not online applications, now determines career success

The role of systems in modern networking

Even when alignment exists, most people struggle with consistency – remembering conversations, maintaining follow-ups, and staying relevant across multiple relationships.

This is where systems, including AI, are beginning to play a role.

Rather than replacing human interaction, they reduce the operational friction around it – supporting continuity, context, and consistency across conversations.

As AI and automation become more embedded in how we work, the advantage will not go to those who can meet the most people, but to those who can build and sustain the most relevant relationships over time.

In practice, this shifts networking from a series of isolated interactions into an ongoing relationship system.

Beyond networking: Building environments that convert

It is worth noting that the event itself was not designed as a networking platform.

It was built to bring together individuals focused on monetising their voice – people actively working towards visibility, positioning, and opportunity creation.

The networking emerged as a natural consequence.

This distinction matters.

Because it suggests that the future of networking may not lie in better tactics, but in better environments – spaces where alignment is built into the room, not left to chance.

The shift ahead

Networking is not disappearing. But it is evolving.

The emphasis is moving away from:

• How many people do you meet towards
• How relevant those people are

And from:

• Starting conversations towards
• Sustaining them

For founders, creators, and operators, the implication is clear: The most valuable networks are no longer built by indiscriminately expanding reach, but by positioning yourself within the right ecosystems.

Because when the room is right, you don’t need better networking tactics.

You need better alignment.

And when that happens, networking stops feeling like effort.

It becomes momentum.

—

The post Networking was the topic, alignment was the outcome appeared first on e27.

Artificial Intelligence (AI) is no longer a futuristic concept; it’s a present-day imperative, dominating boardroom discussions and reshaping industries. Yet, for all the excitement, many organisations stumble on their AI journey. Having advised leaders from global conglomerates to agile, owner-driven firms, I’ve witnessed firsthand the common pitfalls and the pathways to genuine success. My goal is to share these global experiences through this article, helping your organisation navigate the complexities and truly benefit from AI.

Beyond the hype: What companies get wrong (and right) with AI

So, when companies declare, “We’re doing AI,” where do they most often go wrong?

The AI missteps: Where ambition meets reality

• The “instant gratification” trap

Many executives fall into the allure of quick wins, treating AI like an “instant button” for immediate results. This often leads to hasty, and expensive, investment choices without a solid foundation. Imagine attempting to build a skyscraper without a proper blueprint – it’s a recipe for disaster. I recall one executive who privately confessed to exhausting their entire AI budget on expensive hardware before even defining the problem they were trying to solve. That’s like buying a Ferrari before you have a driver’s license, let alone a road to drive it on!

• The missing “why”: Unclear problem formulation

Excitement over the latest AI tools, like Generative AI, is understandable. However, a common misstep is failing to clearly define the actual business problem AI is meant to solve. It’s akin to having a shiny new hammer but no nail in sight! Without a clear “why,” even the most advanced AI becomes a solution in search of a problem.

• The scattered approach: Lacking a cohesive roadmap

I’ve observed organisations launching a flurry of independent AI initiatives without a cohesive strategy. This often results in teams competing for resources, and even if projects are approved, the overall organisational improvement can be negligible. It’s like a rowing team where everyone paddles in a different direction – lots of effort, but little forward momentum. While initial exploration through understanding the concepts and trying to imagine the context in the team, attempting to solve lab-scale problems is valuable, a well-defined organisational roadmap is crucial to be drawn in a reasonable time from the start of exploration. Otherwise, you’re just building a collection of really cool individual rooms, but no functional house.

• The data dilemma: Overlooking data integrity

AI thrives on data. Yet, the importance of accurate, clean, and accessible data is frequently overlooked. This, in my experience, is the single most critical bottleneck. If your data isn’t robust, your AI efforts will struggle. It’s the classic “garbage in, garbage out” scenario, but with much more expensive garbage!

• The human factor: Fear and resistance

People inherently resist any change. If it is coupled with the fear of job displacement, the resistance becomes even stronger. This situation can potentially slow down any AI initiative at the execution level, and it’s imperative to properly address this genuine concern. My message is simple: AI is inevitable. You can’t put the genie back in the bottle. Embracing AI and learning to work with it is about acquiring a new superpower, not facing a new threat.

In essence, “getting it wrong” often stems from treating AI as a magic bullet or a purely technical endeavour, rather than a strategic business transformation. It’s not just about the tech; it’s about the entire orchestra playing in harmony.

Also Read: Balancing ambition and well-being: A founder’s take on sustainable company building

The ingredients for AI success: A recipe for impact

To distil it down, successful AI initiatives typically require:

• AI literacy at the top: Board and executive levels need a clear understanding of AI’s potential and limitations.
• Contextual understanding: AI capabilities must be understood within the unique context of your specific organisation.
• Foundational investment: Allocate sufficient time for building robust foundational capabilities.
• Business value focus: Clearly define the business problem and the expected value outcomes.
• Company-wide strategy: A cohesive, well-defined roadmap ensures alignment and efficiency.
• Addressing human emotions: Empathy and clear communication are vital to mitigate fear and uncertainty.
• Data sanity: Clean, reliable data is the lifeblood of effective AI.
• Top-down commitment: AI is a strategic imperative requiring unwavering support from leadership.
• Tolerance for failure: Expect initial setbacks; they are opportunities for learning and adaptation.

From vision to reality: Making AI deliver

Moving from an AI vision to tangible business impact requires significant organisational transformation and, sometimes, tough decisions. A true cultural shift demands strong stakeholder buy-in and, frankly, top-down enforcement. Making the organisation “AI aware” and up-skilling key executives are paramount.

Here are the critical decisions that determine whether AI creates real business impact or remains theoretical:

• The executive sponsor: An executive sponsor with a complete understanding of the goal, approach, and unwavering commitment is absolutely key. He/she is the champion, the cheerleader, and the bulldozer, moving initiatives from the drawing board to tangible benefits.
• Strategic sourcing: I’ve also seen organisations stumble because they made the wrong decision between in-house skill development versus outsourcing, or they ended up with the wrong implementation partner or product. These are critical choices that can make or break a project.
• Avoiding the “lab-trap”: It’s easy for in-house teams to prove a concept in a lab environment and become complacent. However, scaling to production demands an entirely different approach, requiring robust engineering and operational expertise. A proof-of-concept is like baking a single cupcake; scaling to production is like running a bakery that churns out thousands daily.
• Robust data infrastructure: Once again, robust data infrastructure and governance are non-negotiable. AI initiatives frequently stall because their data isn’t sanitised or is simply insufficient. It’s like trying to bake a cake while basic ingredients are missing – you’re just going to end up with a mess.

Leadership, ownership, and decision-making: The pillars of success

For AI initiatives to truly deliver results, several internal conditions must be met:

• Visionary executive sponsorship: A strong executive sponsor must articulate a compelling vision, positioning AI as a transformative and strategic imperative. A dedicated AI or data leader, accountable for adoption and monetary impact, is also crucial. True AI adoption rarely happens without an executive actively “pushing” (emphasis is on “pushing”) from the top, not just passively monitoring.
• Cross-functional ownership: AI implementation is inherently cross-functional. Ownership must be distributed across diverse teams – data scientists, engineers, business analysts, domain experts, legal, and compliance. Each member needs a clear understanding of their role and how their contribution fits into the larger picture. It’s a team sport, and everyone needs to know their position and strategy.
• Data-driven culture and iteration: The organisational culture should foster data-driven decision-making, embracing rapid prototyping, testing, and iteration. This means moving away from lengthy development cycles and adopting shorter feedback loops. In the world of AI, it’s “fail fast, get up, gather yourself, use the learning and try differently”.

Also Read: AI agents could become the new OTAs — What it means for Agoda and the future of travel

Measuring what matters: Quantifying AI’s impact

When it comes to measurable results, leaders must focus on tailored metrics. I recently spoke with a CEO whose manpower costs were only five per cent of his operational costs, having recently rationalised his workforce by 30 per cent. In his context, simply discussing human productivity enhancement, while valuable, wouldn’t be the most impactful objective for his business.

So, what to measure? It depends entirely on the business problem you’re solving. It could be:

• Revenue growth: From new AI-powered products or services.
• Cost reduction: Through process automation or optimisation.
• Improved customer satisfaction: Due to personalised experiences or faster service.
• Reduced risk: Through AI-driven fraud detection or predictive maintenance.
• Faster time-to-market: For new innovations.
• Real-world examples: I’ve led teams implementing AI combined with physics (Digital Twin) that saw a 15 per cent yield increase in an oil rig. In another instance, quality and customer satisfaction improved, and production output increased by over 25 per cent in a process manufacturing plant.

The key is to link AI initiatives directly to strategic business objectives, define quantifiable metrics before you start, and compare them post-implementation.

Beyond efficiency: Focusing on human outcomes

Perhaps the most important question is how to ensure AI adoption genuinely improves human outcomes – for teams, customers, and society. Any technology developed by humans should ultimately enhance human comfort and well-being. Therefore, embedding ethical AI principles from the very beginning is imperative.

This includes considerations like:

• Fairness and equitable outcomes
• Transparency and explainability
• Sustainability
• Community well-being
• Inclusion (to moderate the digital divide)

The focus should always be on employee empowerment and augmentation, rather than automation that simply replaces jobs. How can AI make our employees better, more effective, and happier? How can it serve our customers more thoughtfully? How can it contribute positively to society? These are the questions we must continually ask.

The smartest first step: Don’t boil the ocean

For senior leaders feeling both excited and overwhelmed by AI, my recommendation is clear: Do not try to create a five-year AI master plan to start with. That would become obsolete quickly, given the pace of evolution of this technology

Instead, identify and champion one or two high-impact, low-complexity AI initiatives that solve a critical business problem and can deliver measurable results within 1 to 3 months. Think of it as a pilot project, a quick win to build momentum and confidence.

Also Read: AI at work: Moving forward with employee engagement

The steps are straightforward:

• Select a concrete, high-value business problem: What’s a genuine pain point AI could alleviate where success would be clearly visible?
• Ensure clean data for that problem: Focus on the specific data needed, not trying to clean all your data at once.
• Define clear, measurable business outcomes: What does success look like, specifically, for this pilot?
• Assemble a small, dedicated, cross-functional team: Empower them by freeing them from routine work and providing necessary training.
• Commit to success: Provide resources and remove roadblocks.
• Achieve that first tangible success: Celebrate it! Make a big deal out of it.
• Replicate and scale: Then, and only then, replicate what you’ve learned to other areas.

This iterative approach builds confidence, demonstrates value, and allows organisations to learn and adapt without getting bogged down in overly ambitious plans from day one. It’s about taking smart, actionable steps, not giant leaps into the unknown.

Ultimately, the companies that succeed with AI will not be the ones that move fastest, but the ones that build the right foundations and make it work in practice.

—

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 Kickstarting your AI journey: How to avoid the million-dollar mistakes most companies make appeared first on e27.

When my Co-Founder, Samuel Adcock, and I started The Ortus Club in Singapore in 2015, we were in our mid-twenties, pitching a fairly unsexy idea to enterprise clients who had never heard of us: that a dinner for 15 people would generate more pipeline than a conference booth in front of 5,000.

Nobody was particularly interested in hearing that from a small agency in Southeast Asia. The established event houses were in London and New York. The enterprise brands we wanted to work with — the Googles, the Visas, the Metas — were already being courted by firms with decades of track record and offices in every major market. We had a WeWork desk and a thesis.

A decade later, The Ortus Club has produced more than 2,500 invitation-only executive events across 40+ countries. Our client list includes Google, Visa, Meta, Adobe, IBM, Zendesk, and Airwallex. We operate across APAC, EMEA, and North America. And we have never run a single open-registration conference.

This is not a success story disguised as a LinkedIn post. This is what we actually learned — including the parts that were painful — about building a global B2B services company from Southeast Asia.

Specialisation was the only thing that made us credible early on

The most important decision we made was also the most uncomfortable one commercially: we said no to everything except invitation-only executive roundtables. No conferences. No exhibition stands. No sponsored speaking slots. One format, done obsessively well.

For the first couple of years, this meant turning away revenue. Prospective clients would ask us to run a panel at their conference or manage a large-format event, and we would say no. That felt reckless when we were trying to build a business. But it gave us something that turned out to be far more valuable than early revenue: a clear identity in a crowded market.

When a VP of Marketing at a Fortune 500 company asks, “Who does invitation-only executive roundtables?” — we wanted the answer to be us, immediately, without qualification. That only works if you are not also doing twelve other things.

We eventually documented our entire methodology in a free guide called The Art of Networking. It remains the most practical thing we have published, and it is the backbone of our company trainings.

Also Read: Strategic investment 101: A founder’s playbook for winning without losing control

Southeast Asia was not a disadvantage — it was a training ground

There is a specific advantage to building a global B2B business from this region that does not get discussed enough in founder narratives. Southeast Asia is the most culturally diverse business environment on the planet. Running events across Singapore, Manila, Jakarta, Kuala Lumpur, Bangkok, and Sydney in the same quarter forces a localisation discipline that companies built in more homogenous markets simply do not develop.

Our team is largely based in Manila and Singapore. The operational muscle we built, adapting tone, format, guest curation approach, and follow-up cadence across six or seven dramatically different cultures in APAC, gave us something we did not fully appreciate until we expanded into Europe and North America: we were already better at localisation than most of our competitors because we had been doing it since day one.

The cultural nuances are not trivial. The formality expected in a Tokyo executive dinner is fundamentally different from what works in Sydney. The way you position an invitation to a CTO in Singapore is not the way you position the same invitation in Jakarta. Getting this wrong does not just reduce attendance — it damages the client’s brand with exactly the people they are trying to reach. We learned this the hard way more than once in our early APAC expansion, and those lessons became the foundation for everything we built afterwards.

Scaling into EMEA and the US meant rebuilding, not copy-pasting

The move into London was our first real test of whether the model could travel outside APAC. The answer was yes — but not without significant adaptation.

London’s senior executive community operates differently from Singapore’s. There is more scepticism toward event invitations generally, longer relationship-building timelines, and a much higher premium on credibility signals in the invitation itself. Who else is attending, who is hosting, what is the venue — these details carry more weight in EMEA than in APAC, where the topic and format tend to do more of the heavy lifting.

Also Read: How founders should build for a Meta-national suture

The US market presented a different challenge entirely. American executives reward directness and clear commercial framing in ways that would feel abrupt in most of Asia. The post-event follow-up expectations are faster and more transactional. And the sheer volume of competing events in markets like New York, San Francisco, and Chicago means your invitation is fighting for calendar space against a much larger field.

The constant across every market — the one thing that has not changed in a decade — is the core thesis: get the right people in a room, design a conversation that creates genuine value for every person present, and the commercial outcomes follow. That has been as true in Zurich and San Francisco as it has been in Singapore and Sydney.

Delegate acquisition is the real challenge

If I had to identify the single most underrated capability in B2B event marketing, it is delegate acquisition — the process of actually getting senior executives to say yes to your invitation and show up.

Anyone can book a nice venue. Anyone can write a compelling agenda. The part that separates event companies that deliver genuine commercial value from those that do not is whether the right people are actually in the room. A beautifully produced roundtable with the wrong 15 people is worthless. An average venue with the right 15 people is transformational.

Our entire operational model is built around this. We do not sell sponsorship packages and hope people register. We identify the specific executives our client needs in the room, and then we do the work — the research, the outreach, the personalisation, the follow-up — to get them there. That process is manual, labour-intensive, and does not scale elegantly. It is also the reason our clients keep coming back.

The 2026 Event Marketer’s Playbook — our annual research publication based on data from 295 senior B2B marketers across 29 roundtables in 30 cities — confirmed what we have seen operationally for years: the cost-per-qualified-conversation for curated invitation-only events is significantly lower than for open-attendance formats when you weight for pipeline quality. The per-event production cost is higher, but the outcome is not comparable.

Also Read: The alliance economy: How founders and investors should position in a fragmented world

What I would tell a founder building a services business in Southeast Asia

First, specialise earlier and more aggressively than feels comfortable. The temptation to be a generalist is strongest when revenue is scarce, which is exactly when saying no matters most.

Second, treat the cultural complexity of this region as a competitive advantage, not a logistical headache. If you can operate effectively across APAC, you are better prepared for global expansion than you realise.

Third, document your methodology and publish it for free. The Art of Networking has been our single most effective resource — not because it generates leads directly, but because it establishes the core mission of what the company stands for. And that’s contagious.

And fourth, invest disproportionately in the part of your service that is hardest to replicate. For us, that is the guesting process. What goes on behind the scenes? For your business, it will be something else. But the principle is the same: the thing your competitors find most difficult to copy is the thing your clients will value most.

We are a decade in now. The thesis has not changed. The rooms have just gotten bigger — and more global.

—

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 How we scaled a B2B events business across 40+ countries appeared first on e27.

Across Southeast Asia, banks and financial institutions are entering a new phase of digital transformation. Customers increasingly expect financial services to be available instantly, whether they are checking balances, resolving account issues, applying for loans, or interacting with support teams across mobile apps and messaging platforms.

Southeast Asia’s financial sector is also expanding rapidly as digital adoption accelerates across the region. According to Google, Temasek and Bain & Company’s e-Conomy SEA 2023 report, the region’s digital economy is projected to reach US$600 billion in gross merchandise value by 2030, with financial services and digital payments playing a central role in that growth.

Artificial intelligence is emerging as one of the key technologies helping banks navigate this transition. McKinsey estimates that AI could generate up to US$1 trillion in additional value annually for the global banking industry, driven by improvements in customer engagement, fraud detection, and operational efficiency.

In particular, conversational AI is gaining traction as institutions look to automate routine customer interactions, support digital onboarding processes, and provide real-time assistance across voice, chat, and video environments.

The rise of conversational banking

Conversational AI has rapidly become a central component of modern digital banking strategies. Deloitte notes that the majority of customer interactions with banks now occur through digital channels such as mobile apps and messaging platforms, increasing the need for scalable automated support systems.

Traditionally, banks relied heavily on call centres and human agents to handle customer enquiries. While effective, these systems often struggled to keep up with growing volumes of customer interactions, particularly as mobile banking adoption surged across Southeast Asia.

AI-powered conversational systems are now helping financial institutions automate many of these routine tasks. Virtual assistants can respond to frequently asked questions, guide users through account services, and provide real-time support through messaging platforms and mobile applications.

In many cases, these systems operate in hybrid environments where AI handles initial interactions while human agents step in for complex issues. This allows banks to improve response times while ensuring customers still receive personalised assistance when needed.

For financial institutions operating in multilingual markets like Southeast Asia, conversational AI also helps scale customer engagement across languages and regions while maintaining consistent service quality.

Also Read: A new era of automation: Establishing best practices for intelligent automation and generative AI

A partnership targeting Southeast Asia’s financial sector

Across Southeast Asia, banks and financial institutions are increasingly exploring conversational AI to improve customer engagement and operational efficiency.

As demand grows for faster, more responsive digital banking experiences, technology providers and system integrators are forming partnerships to help financial institutions deploy AI-driven interaction systems at scale.

Similar collaborations are emerging across the global financial technology ecosystem. Banking software provider Temenos, for example, has partnered with Microsoft to integrate AI capabilities into digital banking platforms, enabling financial institutions to automate customer engagement and improve operational efficiency.

Another example in Southeast Asia is a collaboration between real-time engagement technology provider Agora and Vietnam-based IT services and digital transformation company FPT, aimed at accelerating conversational AI adoption among banks and financial institutions across the region. By combining Agora’s real-time engagement and conversational AI capabilities with FPT’s enterprise integration expertise, the collaboration supports digital banking interactions across voice, chat, and video channels, enabling workflows such as customer support, payment enquiries, lending interactions, insurance onboarding, and multilingual customer engagement across regional markets.

Real-world deployments across the banking sector

Enterprise adoption of conversational AI within financial services is already gaining momentum.

In Singapore, DBS Bank has deployed AI-powered virtual assistants across its digital channels to handle routine customer enquiries, helping reduce response times while allowing human agents to focus on more complex financial services. OCBC Bank has taken a similar approach with its AI-powered chatbot “Emma”, which assists customers with home loan and banking enquiries through digital platforms.

In Vietnam, Sacombank has implemented AI voice agents as part of a next-generation AI contact centre initiative. The deployment increased call handling capacity by more than 58 per cent and allows the system to manage up to 41,000 calls per day, improving service responsiveness while enhancing overall customer experience. 

Similarly, Vietcombank uses Intelligent Virtual Assistant VCB Digibot across messaging channels to answer common customer enquiries related to loans, cards, interest rates, promotions, and currency exchange information. By automating routine requests, bank staff can focus more on complex customer needs and advisory services. 

Another example comes from Home Credit Vietnam, which uses AI voice agents to automate large volumes of call centre interactions each month while maintaining consistent service quality across its customer operations. 

These deployments illustrate how conversational AI can improve operational efficiency while also helping financial institutions handle rapidly growing interaction volumes.

Also Read: Why the AI revolution depends on reinventing energy infrastructure

Balancing innovation with trust and compliance

While AI-driven automation offers clear efficiency benefits, financial institutions must also navigate increasingly complex regulatory environments.

Across Southeast Asia, banking and financial services organisations operate under strict frameworks governing data protection, electronic systems, and consumer safeguards. Any new digital infrastructure must therefore meet rigorous standards for security, privacy, and operational resilience.

Solutions built for the sector must be designed to operate within these regulatory boundaries while still delivering real-time engagement capabilities. For example, Singapore’s Monetary Authority of Singapore (MAS) has introduced technology risk management guidelines that require financial institutions to ensure robust cybersecurity, system resilience, and responsible use of emerging technologies when deploying digital services.

These frameworks highlight the need for AI-powered banking solutions to balance innovation with strong governance, ensuring that automation improves customer experience without compromising regulatory compliance.

The next phase of digital banking in Southeast Asia

Looking ahead, conversational AI is likely to play a growing role as financial institutions across Southeast Asia modernise their digital infrastructure.

Financial institutions are also accelerating the use of artificial intelligence. According to McKinsey’s State of AI report, financial services is among the industries seeing the fastest growth in AI adoption.

Across the region, this shift is becoming visible in how banks manage customer interactions at scale. In Thailand, for example, Kasikornbank has expanded the use of AI across its digital banking services to support automated customer support and personalised recommendations within its mobile banking ecosystem.

Deploying conversational AI in financial services, however, requires more than new software. Banks must integrate real-time communication infrastructure, enterprise AI platforms, and secure data systems while operating within strict regulatory frameworks. As a result, partnerships between AI platform providers, real-time engagement infrastructure companies, and enterprise technology integrators are becoming increasingly important. These collaborations help bridge the gap between emerging AI capabilities and the operational realities of large financial institutions.

For banks facing rising customer expectations and growing operational complexity, the ability to deliver secure, intelligent, and responsive real-time interactions may become a defining factor in the next phase of Southeast Asia’s banking transformation.

—

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 Inside the next phase of AI-driven banking in Southeast Asia appeared first on e27.

Artificial intelligence is stepping off the screen.

At NVIDIA GTC 2026, one theme stood out across announcements, demos, and developer activity: AI is moving beyond cloud-based software into robots, edge devices, and real-world environments.

This shift is driving the rise of embodied AI. And increasingly, these systems are being designed with one interface in mind: voice.

From models to machines

For the past few years, AI innovation has largely centred around models, with stronger reasoning and multimodal capabilities. But GTC 2026 signals a transition from models to machines.

Instead of asking what AI can generate, developers are now exploring what AI can do in real-world environments. This includes AI systems embedded directly into physical devices and environments.

This shift is enabled by the convergence of several layers:

• Edge AI computing platforms like NVIDIA Jetson
• Multimodal models capable of processing vision, audio, and text
• Real-time infrastructure for interaction and response
• Accessible hardware platforms for rapid prototyping

Together, these layers are turning AI from a passive tool into an active system. This shift is not just theoretical. It is already shaping how developers build and experiment with AI systems today.

The rise of developer-first robotics

One of the most notable ways this shift is materialising is through the emergence of developer-first robotics platforms.

These systems are not built solely for industrial deployment. Instead, they are designed to be programmable and modular, allowing developers to prototype embodied AI applications more easily.

NVIDIA’s Isaac platform continues to play a central role here, offering simulation and development tools that allow teams to train and test robotics systems before deploying them in the real world. Jetson-powered kits are also becoming a standard foundation for edge AI and robotics experimentation.

Alongside these, newer platforms are lowering the barrier to entry even further. Reachy Mini, an open-source humanoid robot developed by Pollen Robotics in collaboration with Hugging Face and integrated with Seeed Studio’s hardware ecosystem, is one such platform gaining attention.

Unlike traditional robotics systems, Reachy Mini is designed for interaction. It combines expressive movement, modular hardware, and compatibility with modern AI models, making it easier for developers to build embodied AI agents that can engage with humans.

Why Reachy Mini stands out

What makes Reachy Mini particularly relevant in the current wave of embodied AI is its focus on real-time, human-like interaction.

While many robotics platforms are still centred on automation or industrial tasks, Reachy Mini is designed for developers building interactive AI systems. This distinction has made it increasingly visible across GTC 2026 and its surrounding ecosystem events, where it was also highlighted during NVIDIA CEO Jensen Huang’s keynote.

Developers are using Reachy Mini alongside:

• NVIDIA Jetson Orin Nano for edge AI computing
• Multimodal models from platforms like Hugging Face
• Speech and voice technologies for natural interaction

This combination enables a new class of applications where robots are not just executing predefined workflows, but continuously engaging with users in real time.

Instead of fixed tasks, these systems can:

• Understand spoken input and intent
• Process context using multimodal models
• Respond instantly through voice, movement, or gestures

This reflects a shift in how robotics is designed, from task-based automation to adaptive, real-time interaction. In that sense, Reachy Mini is not just another robotics platform. It reflects a broader move toward developer-first, interaction-driven AI systems built for real-world environments.

Voice as the default interface

As AI moves into physical environments, traditional interfaces become limiting. You cannot rely on screens or keyboards in many real-world scenarios. Interaction needs to be immediate and hands-free.

This is where voice becomes critical.

At GTC, multiple demos and ecosystem collaborations highlight how voice is evolving from a feature into a core interface layer. In systems built on real-time conversational AI infrastructure, voice is not just used for commands, but for full real-time interaction.

Across emerging systems, several capabilities are becoming standard with capabilities such as:

• Far-field audio capture for hands-free interaction
• Speaker recognition for personalised responses
• Wake-word activation for always-on systems
• Real-time speech-to-speech interaction that feels conversational

In robotics setups such as Reachy Mini, this allows users to interact with machines more naturally, without needing structured prompts or predefined commands.

The result is a shift in how humans engage with AI. Instead of typing instructions or navigating interfaces, users can speak, listen, and interact in a way that mirrors human conversation.

As these systems become more reliable and widely deployed, voice is likely to become the primary way users interact with embodied AI.

Beyond robots: The expansion of voice-native devices

The implications of embodied AI extend far beyond humanoid robots.

At NVIDIA GTC 2026, there is a clear push toward voice-native edge devices powered by compact hardware and real-time AI pipelines. Instead of relying on cloud-only systems, developers are increasingly building AI that can operate directly on devices while maintaining real-time responsiveness.

One example comes from collaborations between companies like Agora and Seeed Studio, which are building voice-native edge systems that combine hardware, AI models, and real-time infrastructure.

Microphone array platforms such as Seeed Studio’s reSpeaker, powered by AI voice processors, are designed to capture voice input reliably even in noisy environments. When paired with edge AI computing and conversational AI engines, these systems can:

• Capture voice input through far-field microphones
• Process speech and reasoning in real time
• Deliver responses with ultra-low latency

What makes this architecture notable is the continuous interaction loop it enables. Audio is captured on-device, transmitted through real-time networks, processed by AI systems for understanding and response, and streamed back almost instantly.

This creates a more seamless, always-on experience compared to traditional voice assistants. As a result, developers are starting to build voice-native systems across a wide range of applications:

• Smart home devices that respond contextually to users
• Conferencing systems with real-time transcription and interaction
• AI assistants embedded directly into hardware
• Robotics interfaces that enable natural human-machine communication
• Industrial IoT systems that can be controlled and monitored through voice

The next interface for AI

If the past decade of AI was defined by screens and text, the next decade will be defined by interaction in the physical world. Voice is emerging as the interface that enables AI to operate seamlessly across environments.

What GTC 2026 makes clear is that embodied AI is no longer a distant concept. It is becoming a practical reality, shaped by advances in robotics, edge computing, and real-time interaction.

We are already seeing early signals from companies actively building in this space.

Figure AI is developing humanoid robots designed for real-world work environments, while 1X is focused on safe, human-centric robots for the home.

Tesla continues to push its Optimus robot as part of a broader vision of AI-powered automation, and Boston Dynamics is advancing mobility and autonomy in robotics through systems like Spot and Atlas.

At the same time, Hugging Face is also playing a growing role by expanding open-source models into robotics, making it easier to combine perception, language, and action.

On the interface layer, companies such as Amazon and Google are evolving voice assistants beyond smart speakers into more context-aware, multimodal systems embedded across devices.

What connects these efforts is a shared direction: AI is becoming embodied, interactive, and continuously present.

FFIn the near future, interacting with AI may feel less like prompting a system and more like interacting with systems that can listen, respond, and act in real time. For builders and startups, the question is no longer whether this shift will happen. It is how quickly they adapt.

—

The post What NVIDIA GTC 2026 reveals about the future of embodied AI appeared first on e27.

AI product teams are under constant pressure to make systems more accurate, more personalised, and more “helpful.” The simplest path is obvious: give the model more context. Ingest more documents. Retain more history. Build long-term memory. Expand what the assistant can see, and performance usually improves.

But privacy regimes and privacy expectations push in the opposite direction. Data minimisation, purpose limitation, and collection restriction are not abstract ideals. They are the principal regulators that customers rely on to keep data usage bounded and accountable.

This creates a direct design conflict: the incentives that make AI feel smarter are often the same incentives that make privacy controls weaker.

The right question isn’t “which side wins.” It’s how to build AI systems that improve without defaulting to maximal collection.

Why is this tension structural, not philosophical

In traditional software, minimisation is easier to align with product goals. You collect the fields you need for a feature, you store them for a defined purpose, and you can often explain why each piece of data exists.

AI is different because value comes from correlation and context. Models are better when they can connect fragments across time, across systems, and across interactions. Personalisation improves when the system remembers. Retrieval improves when the corpus is large. Assistance improves when the model sees the full picture.

Teams begin with a narrow scope, then expand it for quality. A support copilot starts with ticket history, then wants CRM data, then wants billing context, then wants internal notes. A productivity assistant starts with documents, then wants email, then wants a calendar, then wants chat logs. Each step can be justified as “improving user experience.”

Individually, these expansions look reasonable. Collectively, they turn an assistant into an always-on observer.

Also Read: Balancing ambition and well-being: A founder’s take on sustainable company building

Data minimisation is not anti-AI; it is pro-boundaries

Minimisation is often misunderstood as “collect less, at any cost.” In practice, it is a boundary principle. It forces organisations to answer three questions clearly.

• What data is required for this feature?
• What purpose does it serve?
• How long do we need it?

AI teams struggle with these questions because the benefits of extra data are often real, but diffuse. More history can improve outcomes in unpredictable ways. More context can reduce edge case failures. More ingestion can make answers more complete.

But that uncertainty is exactly why minimisation matters. If you cannot clearly define why you need a dataset, you are not making a product decision. You are building optionality at the expense of privacy.

How “context maximisation” quietly expands risk?

The privacy risk is not only about what you store. It is also about what you expose and how broadly it can be inferred.

When AI systems ingest broad corpora, they create new pathways for leakage. Users can receive summaries that reveal sensitive details they were never shown directly. Assistants can surface internal information through conversational queries. Models can retain fragments of sensitive text in ways that are hard to reason about operationally.

Long-term memory features introduce a different category of risk: the system remembers things users did not intend to persist, and those memories can resurface out of context. Even when memory is user-facing and configurable, it changes the default posture from “ephemeral interaction” to “persistent profile.”

There is also a governance risk. The more systems you connect, the harder it becomes to explain data flows. When a user asks, “Where did the assistant get that?” the answer needs to be more than “It had access.”

Performance metrics reward collection

This tension becomes sharper because performance is measurable and privacy degradation is often invisible until it is not.

AI teams can track accuracy, resolution time, customer satisfaction, deflection, and engagement. They can show improvements when they add more context. Those wins are immediate and quantifiable.

Also Read: AI agents could become the new OTAs — What it means for Agoda and the future of travel

Privacy risks are delayed and probabilistic. They appear as edge incidents, customer discomfort, regulatory scrutiny, or an erosion of trust that is hard to attribute to one design choice. This leads to a predictable outcome: teams optimise what they can measure.

If you want minimisation to hold, you have to make privacy constraints visible and product-relevant, not just a review step at the end.

Reframing the problem as “context precision”

The practical way forward is to shift from context maximisation to context precision.

Context precision means the system gets the right context for the task, not all context that exists. It treats data access as a targeted operation, not a broad entitlement.

This starts with task-based scoping. What does the assistant need to do right now? Draft a reply. Summarise a document. Recommend next steps. Each task has a minimum viable context. Build around that minimum first, then expand only with explicit justification.

It also requires separating retrieval from retention. Many systems conflate “the model needs access” with “we should store it.” In reality, the assistant can fetch context when needed without permanently retaining it. Not every useful piece of data needs to become part of a long-term memory layer.

Design patterns that reduce conflict

A few patterns consistently help reconcile performance with privacy.

Make context opt-in and visible. If the assistant is going to use email history or calendar content, make that a clear user decision, not an implied default. Users tolerate data use better when it is transparent and controllable.

Use short-lived, purpose-bound context windows. Instead of giving the assistant broad, continuous access, provide time-bounded slices aligned to the task. This improves relevance while limiting exposure.

Prefer selective retrieval over bulk ingestion. Build retrieval mechanisms that pull only what is needed, rather than indexing everything “just in case.” This reduces both the attack surface and the risk of accidental cross-context leakage.

Separate sensitive classes of data into stricter zones. Some data can be used for convenience features with minimal risk. Other data should require higher assurance and tighter policies. Treat “what the assistant can see” as a tiered model, not a single permission.

Treat memory as a product contract. If you introduce long-term memory, define what can be remembered, how it is edited, how it expires, and how users can inspect it. Memory without clear controls becomes a persistent privacy liability.

Build “privacy cost” into AI evaluation. If a model improves with more context, measure the tradeoff explicitly. The question becomes: what incremental performance did we gain, and what additional data exposure did we introduce? When teams are forced to articulate that exchange, minimisation stops being abstract.

Also Read: Why startups fail at offshore expansion (and how to fix it)

Purpose limitation is the hardest line to hold

Purpose limitation is where most AI systems struggle. Data collected for one purpose becomes attractive for another.

A dataset gathered to improve support responses becomes a training corpus. Logs captured for debugging become long-term analytics. Conversations intended to be ephemeral become personal memory.

The danger is not malice. It is reused for convenience.

The only reliable defence is governance that is enforceable in architecture, not just policy. If the system cannot technically access data outside a purpose boundary, the boundary holds. If it can, the boundary will eventually erode.

The most practical path is not extreme minimisation or extreme maximisation. It is precision: giving AI the context it needs for a specific task, for a defined purpose, for a bounded period, with user-visible control and auditable data flows.

—

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 Data minimisation vs AI context maximisation: The battle defining the future of smart systems appeared first on e27.