40% of enterprise apps will have AI agents by 2026: what that means if you’re running SYSPRO in South Africa

Gartner predicted in August 2025 that 40% of enterprise applications will feature task-specific AI agents by 2026, up from less than 5% in 2025. That is not a five-year roadmap. It is the next 12 to 18 months.

For a manufacturer running SYSPRO in South Africa, this is not about whether AI ERP is coming. That question is settled. The question is what it means for your specific operation, your ERP environment, and the decisions your business is positioned to make right now.

The Lancea team’s view is that the manufacturers who benefit most from this shift will not be those who move fastest. They will be those who understand their position clearly and make deliberate choices about where to build.

What 40% of enterprise apps featuring AI agents actually means

The Gartner projection is specific: task-specific AI agents, not general-purpose AI assistants. These are software components that can perform defined operational tasks autonomously within a business application, retrieving data, triggering workflows, evaluating conditions, and in some cases taking action based on the results.

For manufacturers running SYSPRO ERP, this means the operational environment your business works in today will look meaningfully different within that window. AI components are becoming part of how enterprise software delivers value, including the ERP systems manufacturing businesses have built their operations around.

What this does not mean is that every business needs to act on every AI capability at once. The 40% projection describes what enterprise software vendors will embed in their platforms. It does not prescribe how quickly any individual business should adopt. The more useful question is: of the SYSPRO AI capabilities available right now, which ones are worth your attention, and what does your operation need to have in place to benefit from them?

Key Takeaway

The Gartner projection describes what enterprise software vendors are building. The manufacturer’s question is different: which AI capabilities exist today, and what does your SYSPRO environment need to be ready?

What the South African numbers tell us

The local context matters here. Forty-nine percent of South African enterprises are now creating dedicated AI budgets, according to Dell Technologies’ Q2 2024 Innovation Catalyst Research. That is a majority position, not an early-adopter one. For businesses running ERP systems in South Africa, the question is not whether to engage with AI for manufacturing in your operations. It is how to do so in a way that is grounded in your actual operating environment.

AI and IoT investment in South African manufacturing is projected to contribute R380 billion to manufacturing GDP by 2030, according to Tech in Africa’s 2025 analysis of the South African market. That figure reflects the scale of what becomes available to manufacturers who are well-positioned as the capability matures, not to those scrambling to catch up at the end of the window.

South Africa’s manufacturing sector has specific characteristics that make this moment particularly worth paying attention to: supply chain complexity, real constraints on certain technical skills, and the constant pressure to extract more value from existing systems. These are exactly the conditions where AI’s ability to surface data quickly, identify exceptions early, and reduce the friction of getting answers out of ERP environments delivers practical, measurable value.

“The manufacturers who capture this value won’t necessarily be those who moved fastest. They will be those who understood the territory clearly, and made deliberate choices about where to build.”

AI ERP and the deliberate advantage

Forty-three percent of enterprises globally are now actively investing in ERP integration and AI ERP capabilities, up from 35% in 2024, according to Deloitte’s October 2025 research on ERP and agentic AI. The pace of adoption is real.

The pattern the Lancea team has observed across SYSPRO implementations is consistent with this: businesses that extract real value from new capabilities are those who know their own systems well enough to understand what they are actually adding. Moving deliberately is not moving slowly. It means building on a foundation that can carry the weight of what you add, and making sure the AI ERP capabilities your business invests in are connected to real operational data in a real manufacturing ERP environment, not explored in isolation from the systems your business runs on.

The 18-month Gartner window is a useful frame. It is short enough to require a decision about timing. It is also long enough to get the important things right.

What deliberate looks like in practice

For a SYSPRO manufacturer, deliberate AI adoption starts with three questions. These are the ones the Lancea team has found matter most, and the ones most businesses have not yet worked through in full.

• What AI capabilities are available for your SYSPRO environment right now? AI for manufacturing is not a roadmap item for SYSPRO users. Conversational analytics against live ERP data, AI-assisted operational decision-making, and real-time visibility into production and inventory: these capabilities exist now. Understanding what is actually available today, and what it asks of your current SYSPRO configuration, is the right starting point for any conversation about AI in your business.
• Is your SYSPRO instance configured for the architecture AI requires? AI ERP implementations work best when the boundary between your ERP core and any AI layer is clearly understood and managed deliberately. What belongs inside SYSPRO, what sits in a connected layer alongside it, and how data governance applies across that boundary: these decisions shape every AI initiative that follows. Getting this right at the start saves a significant amount of rework later.
• Do you have the right expertise on both sides of the ERP boundary? The AI capability is one side. Deep knowledge of your SYSPRO environment, how it is configured, where the data lives, and how your business rules are encoded, is equally important. The implementations the Lancea team has seen deliver the most value are those where both sides of that boundary are well understood before any AI initiative begins.

Key Takeaway

Deliberate does not mean cautious. It means understanding what is available now, getting the architecture right before initiatives start, and making sure the expertise available to you spans both the AI layer and the SYSPRO environment underneath it.

These three questions are not a reason to wait. They are the work that makes the next 12 to 18 months count.

Where to start

The window is real. The opportunity for South African manufacturers running SYSPRO is real. The right starting point is not a product evaluation or a commitment to a particular AI direction. It is a clear-eyed look at where your business stands today: what your SYSPRO environment looks like, where the data is strong, where the access friction is highest, and what AI capabilities are already within reach.

The Lancea team works with SYSPRO manufacturers across South Africa and the UK who are at exactly this point. The conversation worth having is about your specific environment, not AI in the abstract. That is the place to start.

Talk to the Lancea team.