Dato Seri Ivan Teh: Applied AI Works Best When the Outcome Comes Before the Algorithm

Image: https://www.abnewswire.com/upload/2026/08/228468cbf85289fca4b3d6827a320d4e.jpg

KUALA LUMPUR, Malaysia - August 25, 2026 - Organizations can get more from artificial intelligence by starting with a useful result, keeping accountability visible and designing pilots for sustainable use.

The business case for artificial intelligence is often framed around the sophistication of a model or the novelty of a tool. A more practical starting point is a simple question: What useful outcome should improve, and how will that improvement be recognized?

An editorial explainer published on Ivan Teh [https://ivanteh.com.my/articles/applied-ai]'s website presents that question as the foundation for applying AI responsibly. The central argument is that AI is most useful when a clear outcome, human accountability and a measurable benefit are defined before a model or tool is introduced.

That principle is particularly relevant as organizations move from experimentation toward everyday use. A successful AI initiative is not defined by the presence of an advanced system. It is defined by whether people can use the system to make a task more understandable, repeatable, timely or accessible.

Start With the Outcome, Not the Novelty

AI adoption can become performative when the technology arrives before the problem has been defined. A pilot may appear impressive in a demonstration while offering little value in the workflow where it is expected to operate.

An outcome-first approach reverses that sequence. The organization first identifies the task that needs to improve, the people affected by that task and the evidence that would show whether the change is worthwhile. Only then does it assess whether AI is an appropriate means of achieving the result.

The distinction matters because not every difficult or repetitive process requires artificial intelligence. Some problems may be solved more effectively through clearer procedures, better data access, improved training or conventional software. AI earns its place when it contributes a specific improvement rather than simply adding technological complexity.

Possible outcomes include accelerating a review, helping a team detect recurring patterns or giving specialists more capacity to focus on work that requires judgment. These use cases are less about replacing people than about directing human attention toward the parts of a process where it has the greatest value.

Before a pilot begins, four questions can help turn a broad ambition into a testable proposition:

Question

What a useful answer should identify

What task should improve?

A defined activity rather than a general desire to "use AI"

What would success look like?

A measurable change in speed, quality, consistency, access or understanding

Who remains accountable?

A person or role responsible for reviewing decisions and outcomes

What evidence is available?

Relevant data, a baseline and clear limits on what the system can determine

Keep Human Accountability Visible

A responsible AI application does not allow accountability to disappear behind an automated output. It makes responsibilities clearer.

People using an AI system should understand what the tool is designed to do, what evidence informs its output and what it cannot determine. Just as important, the decision-making process should identify who reviews the result and who remains responsible for the final decision.

This is not only a governance issue. It is also a practical adoption issue. Employees are more likely to use a system appropriately when its purpose and limitations are understandable. Customers, colleagues and other stakeholders are better placed to assess an outcome when the role of the technology is explained rather than hidden.

Human judgment remains essential wherever context, consequences or competing priorities cannot be reduced to a model output. The presence of AI may change how information is prepared or how options are surfaced, but it does not remove the need for people to interpret evidence and take responsibility for decisions.

A Pilot Becomes Practical Only When It Can Be Sustained

A demonstration can show that a system works under controlled conditions. It does not prove that the system can become part of normal work.

The transition from pilot to practice requires several foundations: relevant data, a clearly defined workflow, sufficient skills to operate the system and an honest method of assessing whether the original outcome has been achieved. Without those foundations, an AI initiative may remain a showcase rather than become a durable capability.

The workflow is especially important. If the output is not delivered at the right point, in a form that people can understand and act upon, technical performance alone will not create operational value. The surrounding process may need new review steps, escalation rules, training or documentation.

Sustainability also requires an answer to a less visible question: What happens when the system is uncertain or wrong? A mature implementation defines how unusual cases are handled, when a human must intervene and how feedback is used to improve the process. These safeguards turn AI from a one-time experiment into a managed capability.

Applied AI Is a Management Discipline

The source article places applied AI within a broader story of digital progress. Technology, skills, governance and adoption need to reinforce one another if an improvement is going to last.

That framing shifts attention away from the search for a universal "best" model. The more important decision may be whether the organization has identified the right task, prepared the right evidence and designed the right operating context.

It also offers a useful way to distinguish activity from progress. Activity is measured by the number of tools tested, prototypes built or demonstrations completed. Progress is measured by whether a defined task has become more reliable, more timely, more understandable or more accessible without weakening accountability.

In that sense, applied AI is as much a management discipline as a technical one. Leaders must decide where AI can assist, where human judgment is indispensable and how the organization will learn from the results. Technical teams must make the system usable and transparent. Operational teams must integrate it into routines that people can follow.

From AI Hype to Useful Capability

The most durable applications of artificial intelligence may not be the most spectacular. They may improve review processes, reveal recurring patterns, support better decisions or make expertise easier to access.

Starting with the outcome provides a practical filter for deciding where AI belongs. Keeping human accountability visible helps preserve trust. Designing for the move from pilot to practice ensures that a promising demonstration has a path into everyday work.

The broader lesson is straightforward: artificial intelligence becomes valuable not when it is merely deployed, but when people can understand it, use it and measure the benefit it creates.

References

[1] How should artificial intelligence be applied? Start with a useful outcome - Ivan Teh Editorial Profile [https://ivanteh.com.my/articles/applied-ai]

About Ivan Teh

Ivan Teh is a Malaysia-based technology and digital transformation commentator whose editorial work explores artificial intelligence, digital transformation, data and the practical application of emerging technologies in organizations.

For more information and to read the editorial, visit https://ivanteh.com.my/

Media Contact
Company Name: Ivan Teh
Contact Person: Media Relations
Email:Send Email [https://www.abnewswire.com/email_contact_us.php?pr=dato-seri-ivan-teh-applied-ai-works-best-when-the-outcome-comes-before-the-algorithm]
Country: Malaysia
Website: https://ivanteh.com.my/

Legal Disclaimer: Information contained on this page is provided by an independent third-party content provider. ABNewswire makes no warranties or responsibility or liability for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this article. If you are affiliated with this article or have any complaints or copyright issues related to this article and would like it to be removed, please contact retract@swscontact.com



This release was published on openPR.