FIELD NOTE
Three different operating models
Traditional workflow automation follows explicit triggers, rules, and actions. A copilot proposes or drafts while a person remains in control. An agent selects and sequences actions towards a goal, often across several tools.
Calling every automation an agent makes architecture harder to discuss. Name the level of discretion the system actually receives.
FIELD NOTE
Where deterministic automation wins
Choose rules when the input is structured, the acceptable outcome is clear, and consistency matters: validation, routing, reminders, record synchronisation, scheduled reporting, and approved calculations. These systems are easier to test and explain.
FIELD NOTE
Where an agent can earn its complexity
An agent can help when the path changes with context: researching across approved sources, preparing a case file, resolving a multi-system exception, or coordinating a task that requires several conditional steps.
- Constrain accessible tools and records
- Separate read, draft, approve, and execute permissions
- Set spend, volume, and time limits
- Require approval for material external actions
- Log plans, tool calls, outputs, and overrides
- Design a safe stop and recovery state
FIELD NOTE
The practical test
If the process cannot be explained, owned, and measured without AI, adding autonomy will usually increase confusion. Stabilise the process first, then grant the smallest amount of discretion that creates measurable value.
DIRECT ANSWERS
Questions operators ask
What is the difference between an AI agent and automation?+
Automation follows predefined logic. An AI agent can choose and sequence actions toward a goal using context and tools, which adds flexibility as well as uncertainty and control requirements.
Does every Malaysian SME need AI agents?+
No. Many high-value problems are better solved with reliable integrations, workflow automation, reporting, or a human-assisted copilot.
What controls should an AI agent have?+
Use least-privilege access, action limits, approval gates, data boundaries, complete logs, monitoring, evaluation, and a tested stop and recovery process.
SOURCE LEDGER
Primary sources
Official material is linked directly. Claims are paraphrased and checked against the source status available on 2026-07-21.FRICTION EDITORIAL CONTROL
Original analysis. Visible limitations. No invented certainty.Prepared by Friction Research and reviewed against primary sources. This material is general information, not legal, tax, financial, or regulatory advice. Requirements can change; verify material decisions with the relevant authority or a qualified adviser.Read our editorial policyREADER EXCHANGE
Add to the conversation.
Name and email are required. Suspicious or abusive comments are held before publication.
Loading conversation.