Article By Aidan Brecknell, Vice President and Managing Director, Pacific Region, Infor
As businesses move towards agentic AI, this becomes more complex. There is a need for clear governance around what agents can access and what decisions they can make independently. Australia has an opportunity to build these governance foundations alongside adoption. This will give businesses greater confidence to move from isolated AI use cases to embedding AI across their operations.
Until recently, manufacturers wanted to understand what AI could do and where it might fit into their operations. Could it genuinely deliver value?
But that conversation has changed. With productivity at 60-year lows, there’s an urgent need for AI to shift the dial.
Distribution is an ideal place for AI to drive transformation, sitting at the heart of the Australian economy. Transport, postal and warehousing alone accounted for 4.5% of GDP in 2024-25, while the movement of goods underpins industries from manufacturing and construction to retail and resources.
The operating environment is also becoming more complex. Global volatility is putting renewed pressure on supply-chain resilience, while changing trade patterns, rising costs and new emissions-reporting requirements are forcing businesses to make faster decisions about how goods are sourced, stored and moved. Efficiency is not just about reducing costs but about resilience and competitiveness.
However, technology alone is not enough. Distributors may have abundant transaction data and complex processes, but they lack IT staff, capital and time. They need a simple way to cut through all the complexity and deploy AI in ways that are practical for their business.
One of the biggest lessons we’ve learned is that AI should no longer be viewed as another software product, but as a service embedded within the business so that it can evolve in tandem with the organisation.
For example, a supply chain model could initially be used to identify potential shortages and recommend alternative suppliers. But as demand patterns change, new suppliers are added or transport routes are disrupted, the data and assumptions informing those recommendations also change.
AI is advancing too fast for the traditional enterprise software model, in which a vendor builds, tests, implements and supports a product in a relatively linear fashion. Instead, manufacturers need an approach that lets them continually identify opportunities, deploy new capabilities, measure results and refine processes over time.
Industry expertise more valuable than ever
As AI becomes increasingly capable, precision matters. Manufacturing requires more than general knowledge: it needs industry context. For distribution, this means understanding the differences between electrical, HVAC, plumbing, building supply and other microverticals to produce accurate recommendations relevant to their operations.
There is also a significant cost consideration. Generic AI models can become expensive as manufacturers provide more context and process larger volumes of data to produce useful industry-specific results. This is where tokenomics increasingly matters. Costs multiply as AI use scales across an organisation. By building industry context into AI from the outset, not only is the relevance of its outputs improved, but the cost of deploying it across the business is also reduced.
Optimisation, not just automation
We see a three-step approach to deploying AI: diagnose, automate and optimise.
Process intelligence is the starting point. Manufacturers should first understand workflows and identify bottlenecks, unnecessary complexity and opportunities for improvement. As the cost of experimenting with AI increases, investment must focus on processes where it can deliver measurable improvements, so costs are controlled and linked to business outcomes, such as gaining market share from competitors.
Not every process requires the same level of human involvement. AI can operate autonomously for low-risk,?routine?tasks like invoice reconciliation, master data maintenance, reordering and basic support enquiries, as errors can be corrected quickly without human involvement.
Conversely, production replanning, supplier contract renewals or dynamic order fulfilment have more material consequences, making an approval-gated approach more appropriate.
Importantly, AI should not simply automate existing processes but help organisations continuously optimise them. As a first step, combining AI with near-real-time location and operational data can allow distributors to move from tracking what has already happened towards anticipating supply-chain disruptions, inventory requirements and fulfilment problems before they occur.
Agentic AI takes this a stage further by enabling systems to act on this information with greater autonomy. Beyond this is agent-to-agent (A2A) communication, where specialised AI agents can work together across different functions. For example, an agent identifying a potential inventory shortage could communicate with agents responsible for procurement and logistics to assess suppliers, availability and delivery options, instead of each step being handled separately.
As AI becomes increasingly autonomous, governance becomes more important. Businesses need clean, well-governed data, clearly defined processes and appropriate controls, permissions and oversight to ensure AI operates within defined boundaries and its actions remain auditable.
For Australian manufacturers and distributors, the opportunity now is to turn AI adoption into measurable productivity gains. The urgency is growing as manufacturers in other markets start using AI and automation to move into higher-value activities. Australian businesses will increasingly be competing with companies that are operating at lower cost while also developing more sophisticated operations.
But distributors don’t need another AI product to manage. They need industry-specific software, governed automation, a more adaptive user experience and a partner willing to stay engaged as AI keeps changing.
