India’s startup ecosystem is entering a more disciplined phase, one where capital efficiency, sustainable growth, and talent retention are taking precedence over unchecked expansion. In this environment, Employee Stock Ownership Plans (ESOPs) are no longer viewed as optional perks; they are becoming a critical lever in people strategy.

For founders and HR leaders alike, the question is no longer whether to offer equity, but how to structure it effectively. Despite their growing importance, ESOP pools are often designed reactively, shaped by investor expectations or immediate hiring needs rather than long-term workforce planning. This can lead to misalignment between business goals and employee incentives. Having said that, a more deliberate approach is needed.

ESOPs as a strategic people lever

At a fundamental level, an ESOP pool represents a portion of company ownership reserved for employees. But from a people and culture standpoint, it serves a deeper purpose. Well-designed ESOPs:

• Strengthen alignment between employee performance and business outcomes
• Enhance retention, particularly in critical and leadership roles
• Enable startups to compete for talent despite cash compensation constraints
• Foster a sense of ownership and long-term commitment

In talent-scarce sectors, ESOPs can significantly influence offer acceptance and employee loyalty, especially when employees clearly understand their potential value.

Also Read: From perk to power: Rethinking ESOPs in the modern talent economy

Moving beyond the “standard percentage” mindset

A common mistake organisations make is relying on broad benchmarks when determining ESOP pool size. While many Indian startups allocate between five per cent and 25 per cent, this range offers limited guidance without context. The more relevant considerations include:

• Workforce expansion plans over the next two to three years
• Seniority mix and critical roles to be hired
• Market competitiveness for key talent segments
• Investor expectations and future funding rounds

For HR leaders, this is an opportunity to play a more strategic role, linking equity allocation directly to workforce planning rather than treating it as a finance-driven decision.

Structuring ESOPs: governance matters

An effective ESOP programme is not just about allocation; it requires robust governance and operational clarity.

• Clear ownership and administration: Organisations should define who is responsible for ESOP management, typically a combination of leadership, HR, and finance. This includes grant approvals, compliance, and ongoing communication.
• Vesting design as a retention tool: Vesting schedules are one of the most powerful retention mechanisms within an ESOP framework. Standard structures—such as a four-year vesting period with a one-year cliff—encourage continuity while rewarding long-term contribution. However, companies may need to tailor vesting terms for senior hires or critical roles.
• Thoughtful grant strategy: Equity distribution should be intentional —

• Early-stage employees may receive higher allocations due to higher risk
• Performance-based grants can reinforce meritocracy
• Reserving equity for future leadership hiring is essential for scalability

A static, one-time allocation approach often limits flexibility as the organisation grows.

Managing dilution while driving value

Dilution remains a key concern for founders when creating or expanding ESOP pools. However, it should be viewed through a value-creation lens. Strategic dilution used to attract and retain high-impact talent can significantly enhance enterprise value over time. From a people perspective, the focus should be on ensuring that equity allocation drives:

• Business growth
• Leadership stability
• Long-term employee engagement

The trade-off is not ownership versus dilution; it is short-term control versus long-term value creation.

Also Read: The best new year resolutions for startup founders: Offering ESOPs that actually work

Choosing the right equity instruments

While stock options remain the most widely used ESOP structure in India, organisations are increasingly exploring alternatives such as:

• Restricted Stock Units (RSUs)
• Employee Stock Purchase Plans (ESPPs)
• Phantom stock or cash-settled plans

Each instrument differs in terms of taxation, complexity, and employee perception. HR and leadership teams must align the choice of instrument with:

• Company stage and liquidity outlook
• Employee demographics and financial awareness
• Administrative and compliance capabilities

Bridging the employee understanding gap

One of the most overlooked aspects of ESOP programmes is employee communication. While equity is often positioned as a high-value benefit, many employees lack a clear understanding of vesting timelines, exercise processes, tax implications and realistic value scenarios. This gap can reduce the perceived value of ESOPs, even when the underlying structure is strong. Organisations that invest in ESOP education, through workshops, dashboards, or transparent communication, tend to see higher engagement and retention outcomes.

Risks of poorly designed ESOP programmes

Without careful planning, ESOPs can create unintended challenges:

• Over-allocation leading to excessive dilution
• Under-allocation reduces competitiveness in hiring
• Lack of transparency impacting employee trust
• Compliance and regulatory risks
• Administrative complexity and cost

For HR leaders, this underscores the need to treat ESOPs as an ongoing programme rather than a one-time initiative.

Also Read: How to do ESOP right for your startup

Building a culture of ownership

As India’s startup ecosystem matures, ESOPs are becoming more meaningful due to increasing liquidity events such as IPOs, buybacks, and secondary transactions. However, the true impact of ESOPs extends beyond financial outcomes. When implemented effectively, they contribute to stronger accountability, long-term decision-making and a culture where employees think and act like owners. This cultural shift is often what differentiates high-performing organisations from the rest.

Final thoughts

ESOP pools are not merely financial structures; they are integral to how organisations attract, retain, and engage talent. For founders and HR leaders, the priority should be to:

• Align ESOP design with business and workforce strategy
• Build transparent and well-governed frameworks
• Continuously evolve programmes as the organisation scales

Ultimately, the success of an ESOP programme is not defined by how much equity is allocated, but by how effectively it aligns people with the company’s long-term vision. Because sustainable growth is rarely built by founders alone, it is built by teams that feel invested in the outcome. 

—

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

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

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

The post Rethinking ESOP pools in India: Building ownership without losing control appeared first on e27.

There was a time when building a product was the hardest part of entrepreneurship.

Today, that is changing.

With AI, founders can generate code, design landing pages, create marketing assets, automate workflows, and launch products faster than ever before. What once took months can now happen in days.

But this shift introduces a new challenge.

If building becomes easier, founders risk creating things that nobody actually wants.

The new founder skill is no longer execution alone. It is learning how to validate demand before investing certainty.

For years, startup advice revolved around one central idea: Build the product, launch it, then figure out how to monetise it later.

That approach made sense when building itself was expensive. When engineering resources were scarce, simply getting a product into the market was an achievement.

But in today’s environment, where AI dramatically reduces the cost and speed of execution, the question has changed.

It is no longer, “Can we build this?”

It is, “Should we build this at all?”

This distinction matters.

Many founders still spend months researching, refining, and polishing their ideas before introducing them to the market. They work quietly behind the scenes, convinced that perfection increases the chances of success.

Then launch day arrives.

And nobody buys.

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

I have seen this pattern repeatedly among aspiring entrepreneurs. They pour countless hours into creating programmes, products, and services without ever testing whether genuine demand exists.

Interest and demand are not the same thing.

Someone might follow you because they find you entertaining. They may like your posts, comment enthusiastically, or even share your content with others.

None of those behaviours guarantees they will become paying customers.

Revenue reveals something that engagement alone cannot.

It reveals commitment.

Monetisation is often viewed as the finish line. In reality, it can be one of the earliest and most valuable forms of validation available to founders.

Revenue is information.

It tells us whether the problem is significant enough for people to pay to solve it. It tells us whether our positioning resonates. It tells us whether the timing is right.

Most importantly, it tells us whether we should continue investing our time, energy, and resources into building.

This was a lesson I learned firsthand.

Years ago, I started a media and technology venture that began as a school project. The focus was on creating something valuable and useful. Monetisation was never part of the original strategy.

People enjoyed the content.

They consumed it consistently.

However, because the audience had been conditioned to receive everything for free, introducing paid offerings later became extremely difficult.

The challenge wasn’t generating attention.

The challenge was converting attention into commercial intent.

That experience fundamentally changed how I approach new ventures today.

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

When I conceptualised Seraphina AI, I already had a version that I used internally. It helped me streamline workflows and supported my day-to-day operations.

What I didn’t have was a consumer product.

Instead of immediately building one, I asked a different question:

Would other people value this enough to pay for it?

Rather than spending months creating features based on assumptions, I started with a waitlist.

I shared the idea.

I sent newsletters.

I nurtured conversations around the problem the product was designed to solve.

Eventually, I opened pre-orders.

Only after people committed financially did I decide to invest fully in developing the consumer version of the product.

Those early customers joined in the first half of the year.

The product itself launched approximately nine months later.

Validation came before development.

Today, this philosophy shapes how I launch almost everything.

When exploring a new programme or initiative, I rarely begin by building the entire experience upfront.

Instead, I start with a waitlist.

If there is enough interest, I invite people to place a small deposit.

That deposit is not simply about generating revenue.

It is about measuring conviction.

If people are unwilling to commit a modest amount towards solving a problem, it raises important questions about whether the market truly exists.

This approach helps founders avoid one of the most expensive mistakes in entrepreneurship: building based on assumptions rather than evidence.

In an AI-powered world, ideas are abundant.

Execution is increasingly accessible.

The real constraint is no longer technical capability.

It is a judgment.

The founders who thrive in this environment will not necessarily be the ones who build the fastest.

They will be the ones who validate the smartest.

The ones who understand the difference between curiosity and commitment.

The ones who recognise that not every idea deserves to become a product.

The ones who are willing to test demand before investing in certainty.

Because when building becomes easier, discernment becomes more valuable.

I could build countless products, programmes, and systems.

Many founders can.

But if nobody is willing to use them, what is the point?

The future belongs not to founders who build everything they can.

It belongs to those who know exactly what is worth building in the first place.

—

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

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

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

The post The new founder skill is knowing what not to build appeared first on e27.

I was chatting with my mother last week, and she mentioned the weavers of Bengal.

Not as history. As family memory, the way an older generation talks about things their grandparents lived through. Dhaka muslin had once been the finest textile in the world, exported across Europe, Asia, and the Arab world for centuries. Then the Industrial Revolution arrived. Manchester mills, British tariffs against Indian cotton, and a few decades later, the weavers of Bengal — generations of inherited craft, an entire economic ecosystem — were destitute. The skill did not save them. The market for the skill simply went away.

I have been thinking about that conversation ever since. Because I am also a professor at a business school, and the question I get asked most often, by students and by parents of students, is some version of: what should they study, what should they do, how should they prepare for the workforce of tomorrow?

And I do not have an honest answer that is also a comfortable one.

The thing we cannot keep saying

For two years, the comfortable position in education circles has been that AI is a productivity tool. That it will augmentknowledge workers, not replace them. That the disruption will be gradual, manageable, similar to other technology cycles.

That position is becoming harder to hold honestly.

In May 2025, Dario Amodei, the CEO of Anthropic — one of the companies actually building this technology — told Axios that AI could eliminate roughly 50 per cent of entry-level white-collar jobs within one to five years, and push unemployment to between 10 per cent and 20 per cent. He named tech, finance, law, and consulting specifically. The line that has stayed with me: “We, as the producers of this technology, have a duty and an obligation to be honest about what is coming. Most of them are unaware that this is about to happen.”

A year later, the data is moving in that direction. Big Tech hiring of new graduates has dropped roughly 50 per cent from pre-pandemic levels, according to venture firm SignalFire. Wall Street banks have announced cuts concentrated in entry-level analyst seats. Tech entry-level hiring fell 30–50 per cent across 2025. The first rung of the white-collar ladder is the one being sawed off.

Also Read: The real AI threat isn’t your job, it’s your mind

This is not the metaverse. This is not crypto. Those were narratives in search of use cases. What is happening now is the opposite — capability arriving faster than the use cases, faster than the labour market, faster than education systems can adapt. Every senior leader I speak with this year is seeing it inside their own organisation.

And the next wave is physical

The instinct so far has been to tell young people: go into the trades. Become a plumber or an electrician. The body is safe even if the desk job is not.

I do not think we get to say that for much longer, either.

Self-driving vehicles, until recently a punchline, are now running commercial robotaxi services in multiple cities across the US and China. Humanoid robotics that two years ago could barely walk are now folding laundry and stocking shelves in pilots. The combination — large models meeting physical actuators — is what people in the field are starting to call physical AI. It is at roughly the stage knowledge AI was at in 2022. Look at how far that has come in three years.

I am not predicting that plumbers will disappear by 2030. I am saying I am no longer willing to tell a sixteen-year-old that physical work is a permanent moat. The honest answer is we don’t know. And the pace at which that answer keeps moving makes any specific prediction we make today suspect by next year.

What we cannot predict, and what that means

Here is the other half of the honesty.

The most lucrative careers of the last twenty years are the ones nobody in 2005 could have advised a child to prepare for. The full-time YouTuber. The Twitch streamer. The prompt engineer. The TikTok creator earns more than a partner at a top consultancy. The DevOps engineer. The growth marketer. The mobile app indie developer. None of these was on a syllabus. None had a college pathway. The most we could have done in 2005 was say: the internet seems important; learn to use it, follow your interests, be ready to invent the rest.

This will be true again. Almost certainly more so. There will be wealth, professions, and entire categories of human work that we cannot picture from here and that will become obvious in retrospect. The graduates of 2026 are going to invent jobs we do not yet have words for.

This is the strangely hopeful part of the answer. The thing we cannot do is hand them a map. The thing we can do is make sure they are equipped to draw one.

Also Read: The future is full of humans working with humans, AI systems and other technologies

What I tell graduates now

I have stopped trying to point to specific professions as safe harbours. Instead, I share three things, in roughly this order.

Become fluent with AI before it becomes furniture

Not as a search engine. As a thinking partner, a builder, a critic, a research team in your pocket. The graduates who treat AI as a tool to dodge will be displaced by the graduates who treat it as a force multiplier. The latter group is small today. It will be the entry condition tomorrow.

Build judgment around something you genuinely care about

AI is flattening the cost of producing anything; what becomes scarce is taste, judgment, and the ability to decide what is worth producing. That cannot be taught from a syllabus. It is built by going deep on something — a craft, a domain, a question — that you would care about even if nobody paid you for it. The depth becomes the platform from which you can leverage AI. Breadth without depth produces nothing memorable.

Expect to reinvent yourself, and treat it as normal

My generation built careers around the idea that you would do one thing well for thirty years. The next generation will need to be comfortable doing several things across thirty years, with two-to-three-year reinvention cycles. This is uncomfortable to us. It is not, it turns out, uncomfortable to them. The teenagers I meet are already pattern-matching to this faster than their parents are.

What my mother actually said

After we talked about the weavers, my mother said something I keep returning to. She said the weavers’ children eventually found new ways to live. Not the same way. Not as wealthy, not for a long time. But Bengal did not end with the looms. Something else came after.

That is the most honest thing I can say to a young person right now. The looms you were trained for are changing under your feet. We do not know exactly what comes next. But something will. And the people who do best in any disruption are the ones who stop arguing with the change and start positioning for what is on the other side of it.

The youth I meet are already doing this. Quietly, mostly without us. They do not need us to predict their future. They need us to be honest about ours.

—

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

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

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

The post The weavers of Bengal, my mother, and what to tell tomorrow’s graduates appeared first on e27.

As AI reshapes the marketing function, Southeast Asian startup founders face a deceptively simple question: what does good actually look like now?

AI is restructuring the marketing function faster than most startups have had time to notice. The skills that made a strong marketing hire in 2022 are being automated. The skills that actually matter now are different, and most hiring managers don’t yet have a clear framework for identifying them.

It’s not a question of whether AI will replace marketers. It has largely already replaced specific tasks. The more useful question for founders and operators is: given that, what should your marketing team actually look like?

The execution layer is gone

For lean startup marketing teams, which describe most of the Asia region, AI has effectively eliminated the cost of execution. Content production, campaign setup, basic reporting, and social scheduling: these are now table stakes that AI handles faster and cheaper than a junior hire.

That sounds like good news. In some ways it is. But it creates a structural problem. Many early marketing hires in startups were valued precisely for their ability to execute at volume. If that’s the primary value proposition, the role is under pressure.

A recent conversation with fintech marketing leaders across the region made this tension explicit. Teams are at wildly different stages of AI adoption, from basic prompting to fully agentic workflows, and the gap between early adopters and the rest is widening fast. The consensus: the most valuable marketing hire right now is someone who can adapt to change, operate across multiple functions, and direct AI systems rather than just use them.

Also Read: AI as an audience: Welcome to the citation economy

The profile that keeps coming up: T-shaped specialists who can act as orchestrators. Depth in one discipline, whether that’s demand generation, brand, or content strategy, combined with enough breadth to work across the AI toolchain. Pure generalists, interestingly, may be losing ground. The winning profile is depth plus adaptability, not breadth alone.

Three questions worth asking before your next hire

• Can they tell when AI output is wrong?

Anyone can generate copy, build a campaign brief, or pull a competitive analysis with AI now. The rarer skill is editorial judgment: knowing immediately when the tone is off, the claim is shaky, or the output doesn’t reflect your brand. For APAC startups operating across multiple markets, this is especially critical. AI tools trained predominantly on Western data consistently underrepresent Asian consumer behaviour, local nuance, and regional context. A marketer who can catch that gap and correct for it is genuinely valuable. One who can’t will ship content that quietly erodes trust.

• Are they waiting to be trained, or training themselves?

Only 25 per cent of workers receive formal AI training from their employers, even as skills in AI-exposed roles are evolving 66 per cent faster than other jobs. The marketers pulling ahead aren’t waiting for a curriculum. They’re running experiments, building workflows, and developing their own framework. For founders evaluating candidates, this is a useful signal. Ask what they’ve built or tested with AI in the last three months. The answer tells you a lot.

• Do they understand the trust problem?

This one is particularly relevant in fintech and financial services, but it applies across sectors. AI-generated content at scale risks producing what some are calling “AI slop”: homogenised, generic output that erodes brand differentiation and credibility. In categories where trust is the product, that’s an existential risk, not a content quality issue. The marketer who understands this, who treats AI as a tool for amplification rather than a replacement for judgment, is the one who protects your brand as you scale.

The build vs buy question

One unresolved tension for startup founders right now: whether to build AI marketing capabilities in-house or buy them through agencies and tools. The honest answer is that most startups are doing both, somewhat chaotically, without a clear framework for when each makes sense.

Also Read: The future is full of humans working with humans, AI systems and other technologies

A few rough principles worth considering. Use AI tools for execution that’s repeatable and low-stakes: content variations, SEO drafts, campaign copy. Retain human judgment for anything that touches brand voice, customer trust, or strategic positioning. And be cautious about cutting agency relationships entirely in favour of AI-generated output, the consensus among marketing leaders is that AI isn’t yet ready to own branding at scale. The cost savings can be real; the brand risk is also real.

What this means for how you structure the function

The CMO or marketing lead role is shifting toward orchestration, setting creative and strategic direction while AI handles activation.

AI fluency across the function is now a baseline requirement, not a specialist skill. That doesn’t mean everyone needs to be a prompt engineer. It means everyone needs to understand enough to work with AI intelligently, to direct it, evaluate its output, and know when to override it.

APAC’s talent scarcity makes this more acute. Skills shortages already affect 77 per cent of employers in the region, with sales and marketing among the hardest roles to fill. The pool of candidates who combine domain expertise, AI fluency, and genuine regional judgment is small. Founders who know what they’re looking for and can articulate it clearly in a job description have a meaningful advantage.

The talent reset is already underway. The startups that adapt their hiring frameworks now will be better positioned than those still hiring for the job that existed three years ago.

—

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

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

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

The post What AI means for your next marketing hire appeared first on e27.

One of the fastest ways to weaken an AI programme is to declare it responsible before the organisation has agreed on what that word means in practice.

This is a common mistake because responsible AI sounds mature, board-ready, and difficult to argue against. It travels well in policy documents, governance forums, investor language, and internal announcements. It signals seriousness. It suggests the organisation has thought ahead. It gives the impression that the hard questions are already under control.

Often, they are not.

In many companies, responsible AI is still being treated as a label applied after the real decisions have already been made. The model is selected, the use case is funded, the vendor is approved, the pilot is underway, and then the organisation asks how to make the initiative responsible. By that point, the most important definitional work has usually been deferred. Nobody has been forced to settle what kind of system this actually is, what kind of judgment it is influencing, what kind of harm matters most, what level of error is acceptable, what counts as meaningful human oversight, or which decisions should never be delegated to probabilistic systems at all. 

Most responsible AI programmes are stronger on language than on meaning

The surface signs of seriousness are now familiar. Principles are published. Review committees are formed. Risk templates are created. Training is rolled out. Human in the loop language appears in design documents. Fairness, transparency, explainability, and accountability are all referenced in the right places.

None of this is useless. Much of it is necessary. But none of it matters enough if the core terms remain vague.

What exactly counts as a high impact use case?  What counts as decision support rather than decision making? What counts as a customer affecting output? What counts as automated action? What counts as a material model change? What counts as explainable enough for the real context in which the system will be used? What counts as acceptable performance when the harm is not evenly distributed? What counts as sufficient review when the humans involved do not fully understand the model but are still expected to sign off on its behaviour?

These are not drafting issues. There are operating issues.

The real weakness is the definition debt

Every organisation understands technical debt. Fewer understand the definition of debt.

Definition debt accumulates when an institution moves faster on deployment than on conceptual clarity. It uses broad terms that sound robust but remain internally unstable. It talks about safety, fairness, explainability, oversight, harmful use, customer impact, model drift, and accountability as though these were settled ideas, while different teams are quietly operating with different meanings.

Also Read: Responsible AI won’t scale on good intentions alone

This creates the worst kind of governance problem because it often looks like alignment from a distance.

Legal may think human oversight means a named approver exists in the process. The product may think it means a user can technically ignore the model output. Engineering may think it means the model is not directly triggering an automated downstream action. Operations may think it means an analyst glances at the result before moving on. Audit may think it means there is an evidential record after the fact. Everyone uses the same phrase. Nobody is governing the same reality.

That is the definition of debt in action. The language of control exists, but the operational meaning remains fractured. Over time, this debt becomes expensive. 

Responsible AI fails first as a framing problem

Much of the current debate still assumes that responsible AI is mainly a model problem. How do we reduce bias? How do we improve explainability? How do we strengthen monitoring? How do we govern vendors? How do we prevent misuse?

Those are important questions, but they often arrive too late.

The first failure is usually one of framing. The organisation does not define the system in a way that matches the consequences it is about to create.

A model assisting with internal drafting is one thing. A model shaping customer communications, fraud handling, cyber response, financial recommendations, hiring decisions, investigation summaries, claims triage, or exception management is something else entirely. Yet many institutions still group these under the same technology umbrella and then try to manage them through generic policy language.

That is not governance. That is category collapse.

A serious responsible AI programme starts by distinguishing what kind of influence the system is being granted. Is it informing, recommending, ranking, screening, approving, acting, or persuading? Is it being used in a reversible context or an accumulative one? Is the output advisory in theory but determinative in practice? Is the system affecting a user directly, or affecting the employee who affects the user? Is the harm visible immediately, or does it compound quietly through repeated use?

A more mature approach begins by accepting that the big words in responsible AI are not self-executing.

Fairness for what decision, against what baseline, across which groups, measured over what period, with what acceptable trade-offs. Safety for what use case, against which harms, under what misuse assumptions, with what residual risk tolerance? Oversight by whom, with what expertise, with what authority to intervene, and with what evidence available at the moment intervention is needed. Explainability for which audience, for what decision, and with what purpose. Accountability is assigned to which actor when the output was produced by one team, approved by another, deployed by a third, and acted on by a fourth.

Also Read: 5 dimensions of responsible AI: Enhancing societal needs with blockchain

These are definitional questions disguised as governance questions.

That matters because responsible AI has become crowded with high-level commitments and light on decision-grade clarity. Too much of the discussion still assumes that shared vocabulary means shared understanding. It does not.

Real governance starts when the organisation is willing to pin terms down hard enough that they shape investment, architecture, approval rights, monitoring design, incident response, and executive accountability.

Until then, the programme is mostly speaking in values while operating in approximation.

Process matters, but only when it is tied to consequence

To say responsible AI is a process is not to defend bureaucracy. It is to argue that responsibility must be continuously produced, not merely declared.

A serious process does not begin and end at model approval. It starts with use case framing, continues through design, testing, deployment, monitoring, escalation, retraining, change management, incident learning, and sometimes withdrawal. It recognises that the model will be used differently from how it was originally described, that humans will adapt around it, that workflows will stretch it into adjacent roles, and that the meaning of harm may change once the system interacts with real customers, regulators, operations, and frontline pressure.

That is why a checkbox cannot work. A checkbox assumes the relevant question has been settled at a single moment. Responsible AI assumes the opposite. It assumes the organisation must keep asking whether the system is still behaving within the boundaries that were originally judged acceptable, whether those boundaries were defined well enough in the first place, and whether the real use of the system has drifted beyond what was approved.

This is not red tape. It is the minimum discipline required when deploying systems whose outputs can look more stable than their consequences.

—

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

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

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

The post Responsible AI is a process, not a checkbox appeared first on e27.