Projected data centre energy consumption may rise from 415 TWh in 2024 to 945 TWh by 2030, with related emissions climbing from roughly 185 Mt CO₂ to 421 Mt CO₂. For Malaysian SMEs, these figures translate directly into higher operating costs, larger carbon-related liabilities, and greater pressure to justify every RM of technology spend. AI systems must therefore be treated as assets that deliver measurable returns, not merely as capabilities that consume resources. At AINNA, we approach this from a finance and accounting angle: efficiency, control, and clear business value.
1. Detached AI
Detached AI means we should not attach AI to every process by default. From an accounting perspective, this avoids overcapitalising systems and inflating operating expenditure on compute that does not generate commensurate value. Routine tasks can often be handled more cost-effectively through automation, formulas, database queries, or rule-based systems. By reserving AI for processes where it truly adds value, SMEs can improve asset utilisation, lower unnecessary cloud and energy costs, and build a cleaner digital workflow. The ESG benefit is straightforward: less wasted resource and stronger operational discipline.
2. Smart Routing
Smart routing means assigning each task to the most appropriate tool. Lightweight tasks should run on lightweight systems, while larger AI models should be invoked only when the business case justifies the cost. This is essentially capital allocation for compute: it prevents the overuse of high-cost resources and keeps the cost per transaction in check. For ESG reporting, smart routing reduces the carbon intensity of each process and supports more sustainable technology consumption.
3. Segmentation
Segmentation means breaking AI work into smaller, clearly defined components. Rather than relying on one monolithic system, each module can own a specific function such as reading data, validating accuracy, classifying information, recommending actions, or producing reports. From an asset management standpoint, this makes it easier to track costs, depreciation, performance, and return on investment for each part of the system. It also creates a stronger audit trail and greater transparency, which feeds directly into governance and reliable ESG disclosures.
4. Guardrails
Guardrails are the internal controls that keep AI systems safe, accurate, and accountable. They include standardised output formats, fact-checking steps, sensitive data access limits, and human approval for material decisions. In finance terms, these are risk management and control procedures: they reduce the likelihood of errors, rework, and compliance breaches, and they protect the integrity of financial and ESG data. Responsible AI must support human judgment, not bypass it.
Good AI is not about deploying AI everywhere. It is about deploying AI with purpose, cost discipline, proper controls, and measurable business value. For Malaysian SMEs, this is how AI becomes a genuine asset rather than an open-ended liability, and how ESG commitments become auditable outcomes.
Official Cross-Network Resource
