The Front Door Nobody Uses Yet

In Southeast Asia, AI is not the new front door to product discovery, and the reason is that search never was. Only 24 percent of Southeast Asian consumers use generative AI to support shopping decisions at all, while 71 percent discover products on marketplaces and 88 percent finish the transaction there. This region skipped the search-shaped internet. That means the AI disruption everyone is planning for will not arrive in the shape the plan assumes.

Every deck about 2026 opens the same way. The front door has moved. Search is being replaced by answers. Optimise for the machine or disappear.

Then you look at what people in Kuala Lumpur, Jakarta and Ho Chi Minh City actually do, and the front door is a Shopee tab that never closed.

What do Southeast Asians actually use to find products?

Marketplaces, by a wide margin. In the impact.com, Cube and dentsu survey of 2,400 consumers across six Southeast Asian markets, 71 percent named online marketplaces as a discovery channel against 57 percent for search engines, and 88 percent finalise the transaction on a marketplace. Generative AI tools such as ChatGPT, Gemini and Claude are used to support shopping decisions by 24 percent, one in four (impact.com, Cube and dentsu, 24 July 2026).

That 24 percent is the number worth sitting with, because of what it sits next to. Discovery and checkout happen in the same place. There is no gap between the browse and the buy for a new intermediary to slide into. In the Western model, search sends you somewhere else, and that handoff is exactly where an AI answer can intercept you. Here the handoff barely exists.

So the interesting question is not how fast AI catches up to search in Southeast Asia. It is whether a search-replacement product has anything to replace.

A lot of the regional marketing conversation is imported. The frameworks arrive from the US and UK, where two decades of behaviour ran through a query box, and they arrive already assuming that box is what AI is dismantling. Here, plenty of people went from no internet to a phone with Shopee, TikTok and WhatsApp on it and never developed the habit in the first place. You cannot lose a habit you never had.

Are creators really beating AI here?

On the surface, yes, and the study says so. Sixty-seven percent of respondents have bought a product specifically because a creator recommended it. Four in ten say generative AI has not changed how much they rely on creators. On trust, 38 percent trust creators and AI equally, 36 percent trust creators more, and 25 percent favour AI (impact.com, Cube and dentsu, 24 July 2026).

Now the part that stings, buried in the same dataset. On a four-point influence scale for purchase drivers, family and friends score 2.42, online reviews score 2.36, and creators score 1.98. Creators lose to strangers on a review page.

Read that twice if you have a creator line item in your Q4 budget. The person you paid, briefed, and waited three weeks for is less persuasive than an unpaid sentence from someone with no name and no face. And it is not close.

The comfortable reading of this study is that creators beat AI, so keep booking creators. The numbers do not support that. What they support is much narrower and much more useful: creators still generate purchases, 67 percent of them, while ranking below anonymous reviews on influence. Those two facts only fit together one way. Creators are not winning on persuasion. They are winning on placement. They sit closer to the transaction than anything else in the marketing stack.

Where does the money in this category actually sit?

Follow it. Influencer and affiliate marketing in Southeast Asia is worth about USD 70 billion in 2026, which is 32 percent of the region’s ecommerce sales. Break that open and standalone influencer work is USD 18 billion, affiliate is USD 39 billion, and the overlap of creator content carrying affiliate or shoppable links is USD 13 billion. That overlap is the fastest-growing segment in the category (impact.com, Cube and dentsu, 24 July 2026). For scale, Southeast Asian ecommerce is USD 219 billion in 2026 and projected at roughly USD 410 billion by 2031.

The smallest slice is growing fastest, and it is the only slice where content and checkout occupy the same rectangle of screen.

That is the whole argument in one line. Not reach. Not affinity. Distance to the buy button.

Go back to that trust split for a second, the 38 percent who trust creators and AI equally against 36 percent who lean creator and 25 percent who lean AI. A near three-way split is not a preference. It is a shrug. People are not choosing between a creator and a chatbot as sources of truth, they are using whichever one is open when the urge to buy something arrives.

It also explains the motivation split in the same study. Trust and validation drive 49 percent of affiliate-driven purchases, utility and convenience drive 28 percent. Trust still leads, but convenience is a quarter of the reason people buy through a creator link, and convenience is the one variable a brand can engineer. You cannot manufacture being trusted by Friday. You can shorten the path.

Why is a single Southeast Asia plan already wrong?

Because the regional average hides a spread that would change your media mix. Generative AI use for product discovery runs at 34 percent in Vietnam, 31 percent in Indonesia, and 14 percent in Singapore (impact.com, Cube and dentsu, 24 July 2026).

Singapore, the market with the best connectivity, the highest incomes and the most mature retail, is the laggard. Vietnam is nearly two and a half times ahead of it. Whatever mental model you have that says AI adoption follows wealth and infrastructure, throw it out. The pattern here looks more like mobile-first leapfrogging, where the markets with the least legacy behaviour to unlearn move first.

Malaysia sits inside this and has its own texture. Vodus surveyed 2,556 Malaysian adults between 29 December 2025 and 2 January 2026 and found 67 percent had used an AI tool in the past three months. Of those recent users, 87 percent used chatbots or assistants, and 44 percent used AI embedded inside search engines (Vodus, January 2026). Usage tracks income sharply: 48 percent in households below RM3,500, around 80 percent in households at RM8,000 and above.

So Malaysians are heavy AI users generally. They are simply not using it as a shopping front door. General AI adoption and commerce behaviour have come apart, and if you read one as a proxy for the other you will build for a customer who does not exist yet.

There is a second Malaysian signal that matters more than the adoption rate. Microsoft’s 2026 Work Trend Index, based on 2,000 Malaysian knowledge workers, found 92 percent of Malaysian AI users treat AI output as a starting point rather than the final answer, and 24 percent qualify as Frontier Professionals against 16 percent globally (Microsoft, 23 June 2026). Sixty-nine percent say they produce work they could not have created a year ago.

An AI answer here is a draft. People are fluent with the tools and unwilling to let them finish anything, at work or at checkout. That is not a trust deficit waiting to be closed by a better model. It is a habit, and habits like that are why the 24 percent shopping figure sits so far below the 67 percent general usage figure.

Meanwhile the money in Malaysia keeps going where attention already is. Digital adex hit RM661 million in Q2 2025, up 22 percent year on year, and social media’s share of it climbed from 44 percent in Q2 2024 to nearly 50 percent in Q2 2025 (Media Specialists Association Digital ADEX Report, published August 2026). Over 75 percent of all Malaysian advertising dollars are now spent in digital, according to Ampersand Advisory co-founder and CEO Sandeep Mark Joseph (Marketing Interactive, 2026). Half of digital, in the market with the most social-heavy media diet in the region, and creator influence still scores below a review page. The spend is in the right room and pointed at the wrong wall.

Is the enterprise side building for a customer that exists?

Not yet, no. Deloitte found 29 percent of Asia Pacific consumer businesses are already adopting agentic AI and expect that to reach 76 percent within two years. Nine in ten retail executives expect AI to be used more than traditional search engines by 2026. Half expect today’s multi-step shopping journey to collapse by 2027, and agents could influence or handle as much as 25 percent of global ecommerce sales by 2030 (Deloitte, 14 April 2026).

Line those up against 24 percent consumer usage and the gap is the story. Nine in ten executives are planning around the death of a channel that, in this region, 57 percent of people use for discovery and almost nobody uses for checkout. They are replacing a front door their customers were never standing at.

The same Deloitte data has a quieter number that explains why this gap is survivable: only around 30 percent of these businesses say at least 40 percent of their AI initiatives reach production. More than half have live AI implementations across IT, cybersecurity, marketing, sales and customer support, and around one third are using AI to change their business models. So the capability is real and the shipped output is thin. Most of this is still slideware, and the mismatch between executive belief and consumer behaviour is being absorbed by the fact that not much of it reaches a customer.

I have sat in versions of this meeting. Somebody presents the nine in ten figure, the room nods, and the plan that comes out is a content programme designed to be cited by a chatbot, for a market where 88 percent of transactions close inside an app that does not read chatbots. Nobody in the room is wrong about the direction. They are wrong about the interface.

What happens by the end of 2027?

This is where I stop being smug about the 24 percent. By the end of 2027, one in five Southeast Asian online shoppers will complete a purchase natively through a generative AI platform (impact.com, Cube and dentsu, 24 July 2026).

Note the verb. Complete. Not browse, not ask, not compare. Twenty percent of shoppers transacting inside the AI layer, roughly 15 months after that study published.

Which is entirely consistent with everything above, and that is the part most people will miss. AI is not going to win here by becoming a better search engine. It will win the same way marketplaces won, by putting the recommendation and the payment in the same place. The 24 percent is low because AI has mostly been a place to ask questions. The 20 percent by 2027 is a forecast about AI becoming a place to check out.

Southeast Asia does not reward the discovery layer. It rewards whatever owns the last two centimetres before payment. Marketplaces figured that out. Affiliate-linked creator content is figuring it out, which is why that USD 13 billion sliver grows fastest. Generative AI will matter here the moment it stops being an answer engine and starts being a till.

So what is actually worth doing about it?

Stop buying reach and start buying adjacency.

If creator influence scores 1.98 against 2.36 for anonymous reviews, then the case for paying a premium for follower count has collapsed. You are paying for the weakest driver in the dataset. What you should be paying for is a creator whose content carries a working link into the marketplace where 88 percent of transactions close, and who generates the review-shaped social proof that actually outperforms them. The creator is a distribution channel to your review page and your product listing, not a substitute for either.

Second, treat reviews as a media buy rather than an afterthought. Anonymous reviews out-influence paid creators in this dataset, and the same study shows trust and validation driving 49 percent of affiliate purchases against 28 percent for utility and convenience. Review volume, recency and specificity on the marketplace listing is the highest-scoring asset in the region, and usually the one asset with no owner, no budget line and no brief.

Third, drop the single regional plan. Vietnam at 34 percent and Singapore at 14 percent are not the same brief with different currencies. The Vietnamese consumer will meet an AI checkout before your Singapore consumer finishes trusting one.

Fourth, prepare your product data for the till, not the answer. Every listing detail that a marketplace needs is the same detail an AI purchase flow will need. Price, availability, variants, specs, return terms, in text, current, machine-readable. Brands that already win marketplace search are better prepared for agentic checkout than brands that have spent 2026 writing content designed to be quoted.

The Western framing tells you to fight for the top of the answer. In Southeast Asia the answer was never worth much. The money has always been at the bottom, in the last click, and it is about to be there again in a new interface.

Follower count is a legacy asset. Proximity to the transaction is not.

Frequently asked questions

How do Southeast Asian consumers discover products in 2026?

Mostly on marketplaces. In the impact.com, Cube and dentsu survey of 2,400 consumers across six markets, 71 percent named online marketplaces as a discovery channel against 57 percent for search engines, and 88 percent finalise the transaction on a marketplace.

Do Southeast Asians use AI for shopping?

Only 24 percent use generative AI tools to support shopping decisions, and the spread by country is wide: 34 percent in Vietnam, 31 percent in Indonesia, and 14 percent in Singapore.

Are creator recommendations more influential than online reviews in Southeast Asia?

No. On a four-point purchase-driver influence scale, creators score 1.98, online reviews score 2.36, and family and friends score 2.42. Creators still drive purchases, with 67 percent of respondents buying because a creator recommended something, but they rank below anonymous reviews on influence.

When will AI actually matter for Southeast Asian commerce?

By the end of 2027, one in five Southeast Asian online shoppers is expected to complete a purchase natively inside a generative AI platform, according to impact.com, Cube and dentsu. Deloitte separately reports 29 percent of Asia Pacific consumer businesses already adopting agentic AI, rising to an expected 76 percent within two years.

The penalty in your feed is not for using AI. It is for looking like AI. Pangram scanned 1,002,627 social posts in mid-2026 and found LinkedIn supplied one third of the sample but 62% of everything flagged as machine-written, with over 40% of long-form LinkedIn posts fully AI-generated. Disclosure labels will not save you from that, because a label describes how something was made, not whether anyone stood behind it.

Start with the number that should end the argument about whether this is happening.

Pangram collected 1,002,627 posts across LinkedIn, Medium, Substack, X and Reddit between 24 April and 9 July 2026, excluded anything under 50 words, and ran them through a model with a stated false-positive rate of 0.01%. Across everything scanned, 13.8% came back as AI. That sounds survivable. Then the breakdown lands. Posts over 250 words were flagged as fully AI-generated 25.72% of the time. On LinkedIn, long-form posts crossed 40% fully machine-written. And LinkedIn, which made up roughly a third of the items scanned, accounted for 62% of all AI-flagged content on the platforms studied (Pangram, AI in Your Feed, via About Chromebooks).

Originality.AI put a harsher number on the same platform. In July 2026 it classified 4,061 of 5,000 public LinkedIn posts of at least 100 words as Likely AI, a share of 81.2%. Its own earlier pass, covering 3,368 long-form posts from 99 influential profiles between January and November 2025, had come in at 53.7% (Originality.AI LinkedIn analysis). Detectors disagree, and you should hold both figures loosely. The direction is not in dispute.

Why is LinkedIn the platform everyone points at?

Because LinkedIn built the incentive and is now shipping the complaint form.

The platform has been rolling out a native “Seems like AI slop” option in its three-dot report menu (Pangram, AI in Your Feed). Read that as an admission. A platform does not add a report category for a texture unless the texture has become a supply problem it cannot moderate any other way. LinkedIn spent years rewarding consistent long-form posting, generative tools arrived and made consistent long-form posting nearly free, and the feed filled with exactly what the incentive asked for.

Compare across platforms and the pattern gets clearer. It tracks the reward, not the technology.

Source for the whole table: Pangram, AI in Your Feed, collection period 24 April to 9 July 2026.

The Reddit rows are the interesting ones. Reddit was 36.7% of all items scanned and came back with a combined AI share of 4.4%, the lowest of the five platforms (Pangram). Controlling for length, a top-level Reddit post was 5.25 times more likely to be AI-generated than a reply (Pangram). Replies came in 98.1% human. So the same people, on the same platform, on the same day, write by machine when they are performing and by hand when they are answering someone. The variable is not the tool. It is whether anyone is going to push back on you.

Does disclosure actually fix anything?

This is where I part company with the consensus.

The Meltwater and YouGov study, reported 22 April 2026 across nearly 10,000 consumers in seven markets, found 86% say AI-generated content should be disclosed (Meltwater and YouGov, via Bernama). That result has been treated as a policy instruction. Label it, and trust returns.

It does not follow, and the same study explains why. Thirty-two percent of those consumers would trust a brand less if its content is AI-generated, against 15% who would trust it more (Meltwater and YouGov). So a disclosure label is not neutral metadata. For roughly a third of your audience it is a penalty tag. You are being asked to volunteer for a discount.

Worse, the discount lands unevenly. Labelling is self-reported, and the accounts producing forty posts a week from a template are the least likely to volunteer. The people who label are the people who care, which means they are usually the ones whose work needed the label least. The volume accounts stay unmarked and keep the ambient trust that everyone else just handed over. Any compliance system that runs on the honesty of the least honest participants ends up taxing the careful.

There is a second problem, which is that a label answers the wrong question. It tells a reader what tool touched the draft. It does not tell them whether a person read the final version and agreed to have their name on it. Those are different claims and only one of them is worth anything. A post can be 100% human-typed and still be a rewrite of a listicle its author did not check. A post can be drafted by a model, then argued with, cut in half, corrected on a number, and shipped by someone who will answer for it in the comments. The label treats those two as the same object. Readers do not.

Academic work is starting to poke at this. A two-experiment study in Electronic Markets, published online 24 March 2026 with samples of 325 and 371 participants, examines what AI content labelling actually does to engagement rather than assuming the effect (Electronic Markets, AI content labeling and user engagement on social media). A separate study in Scientific Reports on 17 February 2026 ran 680 US participants across five conditions, one control and four different AI interventions in social interaction (Scientific Reports). The honest summary of this literature in 2026 is that the effects are conditional and the “just add a label” reflex is running ahead of the evidence.

Can people even tell the difference?

Only partly, and their confidence is the problem.

Fifty-eight percent of consumers in the Meltwater and YouGov sample believe they can identify AI-generated content (Meltwater and YouGov). Set that next to Pangram’s finding that over 40% of long-form LinkedIn posts are fully machine-written. If detection by eye worked as well as people think it does, LinkedIn would not have needed to build a report button, because the feed would already have sorted itself out through unfollows.

So there are two failure modes running at once. Some machine-written posts sail through. And some careful human writing gets read as machine-written because it happens to be tidy, structured and evenly paced. That second one is the part nobody warns you about. The tell people react to is not provenance. It is a texture: no specific number, no named person, no decision that could have gone another way, nothing that would embarrass anyone if it turned out to be wrong.

That texture is what the 62% figure is actually measuring. It also explains the rest of the Meltwater data. Fifty-one percent of consumers remain unconvinced about AI against 39% who feel excitement, 73% flag misinformation as a key issue, and online discussion about AI-generated content grew 53% (Meltwater and YouGov). Your audience is not conducting forensics. They are reacting to whether anything in the post could only have come from you.

What are creative professionals actually refusing?

Not the tools. The accountability transfer.

The Uppbeat Creator Report 2026, fielded February to March 2026 across 1,792 creative professionals with more than 2,600 open-text responses, found that only 6.7% consider fully AI-generated content acceptable, while nearly 60% are willing to integrate AI tools into their workflows (Uppbeat Creator Report 2026). Those two numbers sit fifty points apart and they are not in conflict. The same survey noted that respondents aged 16 to 24 were the least likely group to use AI tools at all, which is worth sitting with if you assumed this was a generational adoption curve.

Creative Boom’s State of the Creative Industry 2026, published 29 June 2026 across 882 creative professionals, found 86% using AI tools in their work while only 10% thought AI’s overall effect on the industry was positive, with 58% mixed and 28% negative (Creative Boom). Near-universal use, near-absent enthusiasm.

The line these people are drawing is not at the tool. It is at the last decision. Draft with a model, fine. Hand over the final judgement about what is true, what is cut and whose name goes on it, not fine. Six point seven percent is the size of the group willing to release that, and it is small because releasing it means there is nobody left to answer the question “why did you say that.”

Does this look different in Malaysia?

It looks more advanced, which makes the audience harder to fool.

The Vodus AI Consumption Study Malaysia 2026, fielded 29 December 2025 to 2 January 2026 across 2,556 Malaysian adults stratified to the population census, found 67% had used an AI tool in the past three months. Among recent users, 87% used chatbots or assistants, 44% used AI embedded in search engines, and 40% used AI image or video generators or editors (Vodus).

Forty percent generating images and video. That is the figure that changes the maths for anyone posting in this market. The person scrolling past your carousel has probably typed a prompt into the same class of tool this month, watched it produce something almost right, and learned the shape of almost right. You are not presenting to an audience that cannot recognise the output. You are presenting to co-workers.

The adoption is also concentrated where your buyers are. Roughly 85% of Malaysians aged 25 to 34 used AI recently, against about one in three of those 55 and above. Usage runs at around 80% in households at RM8,000 income and above compared with 48% below RM3,500, and Central Malaysia sits at 78% against 61% in both the North and the South (Vodus). Urban, higher income, prime working age. That is the LinkedIn audience in this country, described exactly.

And Malaysians are not naive users. Thirty-nine percent say their choice between AI and traditional tools depends on the task (Vodus). Microsoft’s 2026 Work Trend Index, published 23 June 2026 from 2,000 Malaysian knowledge workers, found 92% of Malaysian AI users treat AI output as a starting point rather than the final answer (Microsoft Work Trend Index Malaysia). Ninety-two percent. The local norm is already that the machine drafts and the human decides. A disclosure label tells that audience something they assumed. It says nothing about the part they care about.

The same Microsoft study has two figures that describe the behaviour I am arguing for better than anything in the detection research. Among the Malaysian workers it classifies as Frontier Professionals, 42% deliberately do some work without AI to keep their own skills sharp, against 33% of everyone else, and 57% pause to decide what should be done by a human versus an AI, against 39% of others (Microsoft Work Trend Index Malaysia). Malaysia has 24% of its workforce in that group, versus 16% globally. The most capable users are the ones building a deliberate stop into the process. Not because they distrust the output, but because somebody has to be able to defend it later.

So what actually survives a feed like this?

Whatever could not have been produced without a specific person.

If a quarter of everything over 250 words is fully machine-written, then generic competence has no scarcity value left. It costs nothing to produce and nothing to ignore. What is scarce is the stuff a model cannot fabricate on your behalf: a number you pulled and can source, a decision you made that a reasonable person would have made differently, a client situation with a date on it, a position you would still hold if someone senior disagreed in the comments.

There is a practical test for this that takes about ten seconds. Read your own post back and ask whether any competitor could have published it word for word without lying. If the answer is yes, you did not write a post, you produced a unit of feed volume, and no label was going to change how it landed. Most of the marketing content flagged in that Pangram scan would pass that test in the worst way.

Notice that none of that is a style choice. You cannot get there by deleting em dashes or asking for a more casual tone. Those are texture fixes for a substance problem, and the substance problem is that a large share of published marketing content contains no claim anyone is accountable for. Strip a post of its adjectives and see what is left. If nothing is left, the disclosure label is the least of its problems.

It also reframes what the platform is doing. A “Seems like AI slop” button is not a detection tool, because the person clicking it has no idea what tool made the post and only 58% of consumers think they can tell anyway. It is a taste signal. What gets reported is the texture of nobody being home, and readers are considerably better at spotting that than at spotting provenance. Which means the risk to a brand page is not being caught using a model. It is being read as an account that ships without reading.

This is also the only defence that survives better models. Detection will keep getting harder in both directions. Labels will keep being optional for the people who most need them. But a specific claim attached to a specific name stays checkable no matter how good the prose generator gets, because the thing being checked is not the writing. It is whether you were right.

Sixty-two percent of the flagged feed came from one platform. Not because that platform had worse tools. Because it had the strongest reward for posting without anyone in particular behind the post.

Frequently asked questions

How much of LinkedIn is AI-generated in 2026?

Pangram scanned 1,002,627 posts between 24 April and 9 July 2026 and found over 40% of long-form LinkedIn posts fully AI-generated. LinkedIn made up about one third of the sample but 62% of all AI-flagged content.

Should brands label AI-generated content?

86% of consumers say AI content should be disclosed, but 32% would trust a brand less if its content is AI-generated against 15% who would trust it more. A label describes how something was made, not whether a named person is accountable for it.

Can people tell if content was written by AI?

58% believe they can, according to Meltwater and YouGov’s 2026 study of nearly 10,000 consumers. Detection rates in practice are lower, which is why platforms are adding report options rather than relying on readers.

Do creative professionals reject AI tools?

No. Nearly 60% are willing to use AI in their workflows while only 6.7% accept fully AI-generated output, per the Uppbeat Creator Report 2026. The objection is to handing over the final decision, not to the tool.

AI cut the hours behind creative work and 70 percent of agencies did not change what they charge. That is not holding the line. It is a quiet discount. The client keeps paying the old rate for the part that got cheap, and pays nothing for the part that got harder, which is judgement. The billable asset moved. Almost nobody moved with it.

Start with the number, because it is the whole argument in one line. In the D&AD 2026 AI and Creativity Report, published 18 August 2026 and drawn from 197 creative leaders, 70 percent said AI had made parts of the creative process faster and they had not changed their pricing.

The industry read that as a win. Same fee, fewer hours, better margin. I have heard it framed as discipline, as refusing to let clients claw back value, as protecting the perceived worth of creative work.

It is not that. Look at what is actually inside the deliverable.

Two years ago, a campaign fee covered concepting plus a fairly large block of execution: the retouching, the variants, the cutdowns, the deck, the twelve aspect ratios nobody enjoys making. That execution block was the thing you could point at when a client asked what the money bought. It was visible, it took time, it clearly required people.

That block shrank. What did not shrink, and what nobody has been able to automate, is deciding which of the forty machine-generated options deserves to exist, and killing the other thirty-nine in a way you can defend to a client who paid for volume and now sees choice.

So the fee stayed flat, the cheap half got smaller, and the expensive half got bigger. The client is getting a discount on production they never asked for and a free ride on judgement they have never been charged for. Both sides think nothing changed. Everyone is wrong in a slightly different direction.

Why does everyone use AI and almost nobody think it is good for the industry?

Here is the tell that something is broken, and it is the most unbalanced pair of statistics I have seen this year.

In the Creative Boom State of the Creative Industry 2026 survey of 882 creative professionals worldwide, published 29 June 2026, 86 percent used AI tools in their work. In the same survey, 10 percent thought AI’s overall effect on the industry was positive. Fifty-eight percent called it mixed and 28 percent called it negative.

Eighty-six to ten. People are not ambivalent about a tool they use daily because they think it is bad at the job. They are ambivalent because the tool is doing the part that used to be countable, and the part that is left over does not show up anywhere: not in the timesheet, not in the scope, not in the fee, not in the credit.

The rest of that dataset behaves exactly the way you would expect if that were true. Fifty percent felt less financially secure than a year earlier, against 18 percent who felt more secure. Sixty-nine percent reported burnout in the past twelve months, rising to 77 percent among mid-career professionals, the exact group whose value used to sit in fast, reliable execution. Forty-eight percent said they were worried about where the industry is heading, with fewer than 38 percent confident. Thirty-eight percent were considering a job change.

Read that as a market signal rather than a mood. Half of a profession that just got dramatically more productive feels worse off. Productivity gains normally feel like something. These ones feel like nothing because they were priced at zero.

And the appetite for the shortcut has a ceiling that gets ignored in every keynote. The Uppbeat Creator Report 2026, fielded February to March 2026 across 1,792 creative professionals, found only 6.7 percent consider fully AI-generated content acceptable, while nearly 60 percent are willing to integrate AI tools into their workflows. The same report found that respondents aged 16 to 24 were the least likely group to use AI tools at all, which should trouble anyone who assumed the youngest cohort would simply automate the rest of us out of the room.

Sixty percent will use it. Under seven percent will ship it raw. The gap between those two numbers is a job description, and it is the job nobody has written into a contract.

I want to be careful here, because there is a lazy version of this argument that says creatives are simply underpaid and AI is the latest excuse. That is not it. Underpayment is a rate problem and it has existed forever. This is a description problem. The thing being sold in 2026 is not the thing named in the agreement, and both parties are operating off a document that quietly went out of date.

What exactly is the client paying for now?

Break the deliverable into its parts and the mismatch is obvious.

Nothing in that table is about rates. It is about the fact that the line item everybody agreed on in 2024 describes a job that no longer exists in the same proportions.

The client-side numbers confirm this is not an agency-only phenomenon. The McKinsey Global Survey on the state of AI, fielded 4 May to 8 June 2026 across 1,719 participants in 97 nations, found nearly nine in ten organisations report regular AI use in at least one business function, 44 percent say AI is scaling across the enterprise (up from 38 percent a year earlier), and 56 percent use it in three or more functions, up from 51 percent. The most interesting figure in that survey for anyone selling creative services: 32 percent of organisations decided against buying at least one software product because they could build it internally with agentic coding tools.

If a third of your clients are now willing to build rather than buy the tooling, they will eventually ask the same question about the assets. The answer cannot be “our production is faster than yours.” That advantage has a half-life measured in months. The answer has to be the part they cannot generate, which is the call on what deserves to run.

Does Malaysia’s growing ad spend make this better or worse?

Worse, and this is where the argument stops being theoretical.

Malaysian digital advertising is expanding. The Digital ADEX Report from the Media Specialists Association with the Malaysian Advertisers Association and the Malaysian Digital Association, covering January to June 2025 and drawing on 21 agencies that represent about 60 percent of the country’s digital ad spend, recorded Q2 2025 digital adex of RM661 million, up 22 percent year on year. Q1 2025 came in at RM343 million, up 6.4 percent. Social media’s share of digital rose from 44 percent in Q2 2024 to nearly 50 percent in Q2 2025. Separately, more than 75 percent of all Malaysian advertising dollars now go to digital in 2026, according to Sandeep Mark Joseph, co-founder and CEO of Ampersand Advisory, in Marketing-Interactive. Forrester, in the same piece, expects marketers to cut display spend by 30 percent as AI and connected TV redefine engagement.

Put those together. Money is moving into the channel with the highest asset count per ringgit. Social eats half the budget, and social does not want one hero film. It wants fifty things, weekly, per market, per language.

So demand for volume is rising at precisely the moment volume became the cheap part. An agency that quietly absorbed the production discount in 2025 is now absorbing it across four times the asset count, with the review and rejection load scaling linearly while the fee does not. That is not a margin story. It is a slow structural bleed dressed up as growth.

Malaysia also gives us the clearest evidence that the real work relocated, and it comes from outside the industry. The Microsoft 2026 Work Trend Index, published 23 June 2026 and based on 2,000 Malaysian knowledge workers plus trillions of anonymised Microsoft 365 signals, found 92 percent of Malaysian AI users treat AI output as a starting point rather than a final answer. Sixty-nine percent say they produce work they could not have created a year ago. Twenty-four percent qualify as Frontier Professionals, against 16 percent globally.

Then the same study found that 19 percent say they are rewarded for reinventing how work gets done when it does not produce immediate results. And only 32 percent say leadership is clearly and consistently aligned on AI.

Ninety-two percent do the editorial work. Nineteen percent get rewarded for it. That is the agency problem stated in national statistics. The judgement layer is real, near-universal, and invisible to whoever signs off on value.

The behavioural detail underneath it is worth stealing. Among Malaysian Frontier Professionals, 57 percent pause to decide what should be done by a human versus an AI, compared with 39 percent of everyone else. Forty-two percent deliberately do some work without AI to keep their skills sharp, against 33 percent of others. Twenty-six percent report documented, repeatable agent workflows and quality standards, versus 18 percent of others.

Read those three lines again. The people getting the most out of the tools are the ones who spend the most time deciding when not to use them, and who write down what good looks like. That is not production skill. It is editorial standard-setting, and it is the thing that now determines output quality.

Is the problem that the work got worse, or that the valuable part went unnamed?

The second one, mostly.

There is a version of this conversation that blames the output. Everything looks the same, the machine flattens taste, the feed is full of smooth nothing. There is evidence for the slop half of it: Originality.AI’s July 2026 analysis of 5,000 public LinkedIn posts of at least 100 words classified 4,061 of them, or 81.2 percent, as likely AI, up from 53.7 percent in an earlier analysis of long-form posts by influential profiles across January to November 2025. Pangram’s scan of 1,002,627 posts between 24 April and 9 July 2026, in the same roundup, flagged over 40 percent of LinkedIn long-form posts as fully AI-generated.

But blaming the output lets the commercial structure off the hook. The work did not get worse because the tools are bad. It got worse in the places where nobody was paid to have an opinion about it. When production is the billable unit, volume gets attention. When production costs almost nothing, volume gets made anyway, and the only thing standing between a client and a hundred mediocre assets is someone whose judgement is not on the invoice.

That person exists in most agencies. They are usually senior, usually tired, and currently the least legible part of the P&L. The 69 percent burnout figure and the 77 percent among mid-career professionals in the Creative Boom data are what it looks like when a whole layer of work is performed without being described. Founders and studio leaders reported 59 percent burnout in the same survey, notably lower than the mid-career figure, which tells you the pressure is landing hardest on the layer that used to convert time into visible output.

There is a second cost that shows up later. When judgement is unpriced, it is also unstaffed. Nobody hires for it, nobody trains for it, and the juniors watch the senior work happen off the clock and conclude that it is not really part of the job. The Microsoft Malaysia finding that only 26 percent of Frontier Professionals have documented, repeatable quality standards, against 18 percent of everyone else, is a generous reading of an industry that mostly has none. Standards that live in one tired person’s head do not survive that person leaving.

So what actually changed about the job?

The unit of value moved from making the thing to deciding which thing should exist.

That sentence sounds soft until you notice that every dataset in this post points at it independently. D&AD found the process got faster while 70 percent of the commercial framing stayed still. Uppbeat found under 7 percent of creatives will ship raw machine output while nearly 60 percent will work with it, which means the entire profession has silently taken on an editing role. Microsoft found 92 percent of Malaysian AI users treat output as a draft. McKinsey found clients scaling AI into three or more functions and starting to build rather than buy. Creative Boom found the people doing all of this feel less financially secure than they did a year ago.

None of those are pricing findings. They describe a job whose centre of gravity shifted while its description did not.

The uncomfortable part is that the shortcut worked. Things really are faster. The reports that say quality slipped are also right, and both can be true, because speed was captured and judgement was not. You cannot compress the part of the work that was never counted in the first place.

An agency that keeps describing itself by what it produces is describing the cheap half. Malaysia’s ad market will happily buy that cheap half in enormous quantities, at rising volume, at a flat fee, until someone notices what is missing from the description.

The shortcut had a price. It just went on the wrong invoice.

Frequently asked questions

Did AI change what agencies charge in 2026?

Mostly no. The D&AD 2026 AI and Creativity Report, published 18 August 2026 from 197 creative leaders, found 70 percent of agencies had not changed their pricing despite AI making parts of the creative process faster.

Why do creatives use AI if they do not think it helps the industry?

Because the two questions are separate. Creative Boom’s June 2026 survey of 882 professionals found 86 percent use AI tools while only 10 percent call its effect on the industry positive, and 50 percent feel less financially secure than a year earlier.

What part of creative work got harder because of AI?

Deciding what should exist. Only 6.7 percent of creatives accept fully AI-generated content, per the Uppbeat Creator Report 2026, and 92 percent of Malaysian AI users treat AI output as a starting point, per Microsoft’s 2026 Work Trend Index.

Is demand for creative work falling in Malaysia?

No, volume demand is rising. Malaysian digital adex hit RM661 million in Q2 2025, up 22 percent year on year, with social approaching half of all digital spend, according to the Digital ADEX Report from the Media Specialists Association.

AI has changed social media content production in three concrete ways: it removes the cost ceiling on creative variations, it compresses the time between spotting a trend and publishing content around it, and it enables near-real-time optimization of what’s working. For brand owners, this means more content, faster turnaround, and campaigns that can adapt mid-run instead of only being evaluated after the fact.

Where AI is actually changing the workflow

Creative variation at scale
Producing 10 versions of a creative concept used to mean 10x the production cost. AI-assisted generation makes iteration far cheaper, so brands can test more concepts and let performance data — not guesswork — decide what scales.

Faster trend response
Trends on platforms like TikTok and Instagram move in days, not months. AI-assisted content planning and production shortens the gap between spotting a trend and having brand-appropriate content live, which matters most for categories like F&B, gaming, and consumer electronics where cultural relevance drives engagement.

Real-time optimization
Instead of waiting for a monthly report to learn what worked, AI-driven campaign management can surface performance signals during a live campaign, allowing creative and targeting adjustments while the campaign is still running.

What hasn’t changed

Strategy, brand judgment, and knowing your audience still matter — arguably more, since AI removes production as the bottleneck and puts more weight on what to make and why. Brands that treat AI purely as a content-volume tool without strategic direction tend to produce more content that doesn’t move the needle.

FAQ

Q: Does AI replace the need for a content strategist?
A: No — AI increases production capacity, but strategic decisions (positioning, audience, what to prioritize) still require human judgment.

Q: Which platforms benefit most from AI-accelerated content?
A: Fast-moving, trend-driven platforms like TikTok, Instagram Reels, and YouTube Shorts benefit most, since content velocity and trend responsiveness directly affect performance.

Q: How can a brand start using AI in its social content production without a full agency partner?
A: Start with AI-assisted content planning and creative generation for a single platform, measure performance, then scale — or partner with an AI-native agency to skip the setup curve.

CTA: Want AI-driven social content production for your brand? See WildAge’s social media services → or start a conversation →

An AI-native marketing agency is one that builds its core workflow — strategy, creative production, content, and campaign management — around AI tools from the start, rather than layering AI onto an existing agency model. The difference matters because it changes what’s actually possible: turnaround time drops from weeks to days, creative output scales without a proportional increase in headcount, and campaigns can be tested and iterated far faster than traditional production cycles allow.

WildAge is one example of this model — originally a traditional creative studio founded in 2008, now rebuilt around AI-driven creative and campaign execution for brands across FMCG, consumer electronics, home improvement, F&B, and gaming.

What “AI-native” actually means

Being AI-native isn’t about using ChatGPT to draft captions. It means:

• Creative production uses AI-assisted generation and iteration to produce more concepts, faster, without a full production team for every variation
• Content strategy uses AI to identify what’s resonating in near-real time, rather than relying on quarterly trend reports
• Campaign management uses AI-driven optimization to adjust targeting and creative mix during a live campaign, not just in the post-mortem

AI-native vs. traditional agency: a comparison

Why this matters for brand owners

If you’re a solo founder or a small marketing team, the traditional agency model often prices you out of quality creative and consistent content — you either can’t afford the retainer, or you get a junior team stretched across too many accounts. An AI-native model changes that math: the AI does the heavy lifting on production volume, while strategists focus on the decisions that actually require judgment — positioning, messaging, what to test next.

FAQ

Q: Does AI-native mean less human involvement?
A: No — it shifts human time toward strategy and judgment calls, while AI handles production volume and first-draft iteration.

Q: Is AI-native marketing only for large brands?
A: No. It’s often more accessible for smaller brands, since it reduces the cost of producing consistent, quality content compared to traditional production overhead.

Q: How do I know if an agency is actually AI-native versus just using AI tools superficially?
A: Ask how AI is used in their actual production and campaign workflow — not just whether they use AI writing tools for first drafts.

CTA: Want to see what AI-native marketing looks like for your brand? Start a conversation with WildAge →