Information Is Abundant. Verification Is Scarce.

Imagine a company announcing that a new industrial project will transform a region. The announcement appears in several publications. An industry newsletter includes it in a market overview. An AI system draws on those pages and describes the region as a rapidly developing opportunity. The conclusion reaches a board presentation, now supported by an apparently reassuring collection of sources.

Nobody needs to have invented anything for the process to become misleading. The announcement may be genuine. The reporting may accurately repeat it. The presentation may still allow a plan to be mistaken for operating demand. Five references can trace back to one statement. The number has grown. The evidence has not.

This is the problem I find more interesting than the familiar complaint that AI sometimes makes things up. A business can reach a weak conclusion using real documents, correct quotations and entirely respectable sources. It simply asks those sources to support more than they know.

Any executive can now produce in an afternoon an analysis that would once have taken a team a month. Country comparisons, regulatory summaries, competitor profiles, all fluent, all structured, all persuasive. The reason is architectural: these systems are trained to be helpful, and helpful, in practice, means agreeable. They absorb the framing of your question and return it to you polished. That is genuinely useful, and we use it constantly in our own advisory work. But the polish is the product, and the consequence is that the persuasive layer of information has inflated while the verifiable layer has not grown at all.

Consider what this means for a concrete case, a feasibility study for entering an ASEAN market. The study states that demand is strong, that regulation is manageable, that a capable partner has been identified. Each sentence may be true, and each may also be the output of a chain of unverified inputs. Demand measured how, by whom? Regulation manageable according to which interpretation, the published rule or the administered one? The partner identified by whom, and checked against what? A decision taken on the sentences without the answers is not analysis-driven. It is formatting-driven, and the market will grade it eventually, on its own timetable.

I described this shift in my research framework as the Great Epistemic Reversal. The scarce resource is moving from the ability to assemble an answer towards the judgement required to know what to believe. Clear structure, fluent prose and a plausible conclusion are now cheap to produce, which means they have stopped being evidence of the work behind them. The harder questions concern whether the information is current, whether apparently separate confirmations are independent, and whether the conclusion survives contact with the thing being described.

Return to the industrial project. An announcement establishes that a project has been announced: an intention, a proposed scale, an expected timetable. A supplier deciding whether to establish a service operation needs to know rather more. Has the relevant part of the project moved forward? Who will purchase the equipment, and when will support be needed? Will that work be available to a new supplier, or does it sit inside an existing global agreement? The announcement can justify investigating those questions. It cannot answer them merely by being repeated.

The same discipline applies everywhere. A corporate profile establishes recorded identity. A customer reference speaks to one particular experience. A site visit reveals what was present on the day of the visit. Each contributes something. None is a universal guarantee. This is where experience matters: someone who understands the industry knows which apparently minor detail changes the conclusion and which impressive figure adds very little. The useful question to ask an experienced adviser is not ‘Which large firm said this first?’ It is ‘What are you seeing here, and what could change your view?’

I call the mechanism an epistemic echo: apparent confirmations that lead back to the same source or assumption. A company’s announcement, an article quoting it and an AI summary of that article are three presentations of one piece of evidence. Asking another system widens the search, but agreement between systems does not by itself establish independent confirmation. People are part of the loop too. Once a proposition appears in the planning document, it is repeated as something already settled, and later readers inherit the confidence without seeing how it arose. An assumption becomes difficult to challenge because nobody wants to reopen a question the organisation believes it has answered.

The remedy is not to distrust everything familiar. It is to notice when familiarity is doing work that evidence should be doing. Information that arrives through one channel, however polished, is an opinion with good formatting. Information that survives contact with independent, differently motivated sources begins to deserve the name knowledge. The test is uncomfortable and simple: who benefits if I believe this, and have I asked anyone who does not?

There is a mistake at the opposite extreme. Faced with uncertainty, an organisation can demand so much documentation that nobody is willing to offer a view. A report can be impeccably referenced and almost useless if it never says what the author thinks matters. Clients do not need an adviser to transport other people’s conclusions into a new document. They need judgement about what those conclusions mean for their business.

I see no contradiction between a strong opinion and serious verification. The opinion gives the investigation direction. Verification tests the parts of the argument that depend on facts. The boundary is whether we can distinguish the observation from the claim of proof. ‘This is the risk I would pay attention to’ is professional judgement. ‘This project has secured the necessary approval’ is a factual statement. Both belong in the same conversation. They should not be mistaken for each other. An adviser should be able to make the first statement without borrowing another organisation’s authority, and should be prepared to substantiate the second.

Verification is not a demand for certainty before every decision. That would be an efficient way to prevent a business from doing anything. An exploratory conversation, an exclusive appointment and a substantial investment carry different consequences, and they deserve different levels of investigation. Spend the effort on the uncertainty that could actually alter the commitment, rather than giving every sentence the same attention. Sometimes one conversation with the right person resolves more than another hundred pages of background. Sometimes the records contradict a persuasive conversation. Sometimes the honest conclusion is that a claim remains unverified, and the company must decide whether to proceed at a smaller scale or wait.

There is also a discipline in stopping. Once the material questions have been answered well enough for the decision at hand, more research can become a way of avoiding responsibility. Verification should improve a decision, not postpone it indefinitely.

If your next consequential decision is resting on a beautifully written analysis, do one thing before you act on it. Identify the three claims the decision is load-bearing on, and ask of each: who verified this, independently, and what would I see if it were wrong? Install the habits that make the asking routine: name a challenger for every major recommendation, someone whose standing improves if the plan fails rather than if it succeeds. Require every significant document to carry its uncertainties, written by the author, distinguishing what is known from what is inferred from what is guessed. Reward the early flaw-finder in public, and price the late surprise accordingly.

If the answers to those three questions are clear, proceed with confidence. If they are not, you do not have a decision yet. You have a well-formatted hope, and hope, however fluently expressed, has never once survived contact with a market on its own.

At TAPiO, verification is not a step in our process. It is the process. If you are making a consequential decision in Malaysia or ASEAN, we would rather be the ones who checked it than the ones who polished it.

Let us talk.

TAPiO Management Advisory is an international advisory firm headquartered in Kuala Lumpur, supporting companies, government agencies and institutions with market entry, investment facilitation and strategic projects across Malaysia and ASEAN.