Malaysian SMEs need a workflow owner before AI becomes infrastructure

Gleb Tsipursky

Malaysia’s Artificial Intelligence Malaysia Expo begins today at MITEC with a practical promise for small and medium-sized businesses (SMEs): tools, demonstrations, workshops, business matching, and guidance that can move AI from curiosity into everyday work.

That is useful. It also creates a management problem that gets less attention than the technology itself. Once an AI workflow moves from experiment to routine infrastructure, who owns it?

A business can have an IT administrator, a software vendor, and employees using the tool every day while still having nobody accountable for the workflow’s business outcome. When that happens, problems become easy to pass around. The vendor points to the model. IT points to the business process. Employees work around the system. Managers see faster output and assume the automation is succeeding.

Business News Malaysia recently described AI as a major accelerator for SMEs, especially because smaller firms can adopt tools with more agility than large organizations. That agility becomes a real advantage when companies pair it with clear ownership.

Before an SME scales an AI workflow, one named person should own six decisions on a single page: the workflow’s purpose, what data the system may use, the minimum acceptable output quality, which outputs require human review, which situations trigger escalation, and what evidence would cause the company to pause or stop the workflow.

That owner should come from the function receiving the business value. A marketing workflow needs a marketing owner. A finance workflow needs a finance owner. An operations workflow needs an operations owner. IT and vendors should support the system, but the person accountable for the outcome should understand the real work well enough to recognize when automation creates a new problem instead of solving an old one.

A workflow for Malaysian SMEs

Then run the workflow for 30 days and keep a simple exception log. Record where employees intervene, how long checking and corrections take, which errors recur, whether work gets reopened, and whether customers or colleagues need human rescue after an automated step goes wrong.

The owner reviews that evidence each week and makes one of three decisions: scale, redesign, or stop. That creates a much stronger management signal than adoption rates or the number of prompts employees send.

Clear ownership also helps with employee trust. People are more likely to report failures when they know who can fix the process and when escalation is expected rather than treated as resistance to the technology. That feedback gives leaders a faster route to useful adoption because problems surface before they become embedded habits.

SMEs have another reason to establish ownership early. Successful automation tends to become invisible. Once a process works most of the time, people stop thinking about the decisions built into it. That makes unclear authority harder to correct later.

At AIMX, business owners should ask vendors about more than features and promised time savings. They should ask how the tool handles exceptions, how data boundaries are enforced, what gets logged, how uncertain cases reach a human, and what controls let the business owner pause the workflow.

Malaysia’s SMEs can move faster than larger organizations. The companies that turn that speed into durable value will give every important AI workflow a human owner before the technology becomes infrastructure.

—

Gleb Tsipursky, PhD, a behavioral scientist, CEO of Disaster Avoidance Experts, and author of The Psychology of AI Adoption at Work: From Resistance to Results (Georgetown University Press, 2026). https://disasteravoidanceexperts.com/aibook

#businessnews

From Digital Convenience to Intelligent Living: How AI is Embedded in Everyday Life