Has Your ERP Invested in Process Mining?

Over the past five years, many of the major enterprise platforms holding operational data bought a process mining capability. SAP bought Signavio in 2021. Microsoft built Process Advisor and then bought Minit in 2022. UiPath bought ProcessGold in 2019. IBM bought MyInvenio in 2021. ServiceNow built process mining into the Now Platform and then went on to buy UltimateSuite in 2024. Salesforce acquired Apromore in late 2025 to underpin its Agentforce vision. Vendors who otherwise compete reached the same conclusion within a few years: their existing analytics weren't enough.

What's interesting is that they bought rather than built. Process mining uses different data structures to transactional ERP systems: event logs, object-centric models, and a process-aware query layer. It's a fundamentally different toolkit from a system of record, with different engineering skills and different research disciplines. Bolting it onto an existing ERP isn't just an engineering project. It's a capability the enterprise platform vendors didn't have in-house and concluded they couldn't realistically grow fast enough. Buying a vendor who'd spent a decade building the expertise was a very practical route.

ERP Analytics and Process Mining Aren't the Same Thing

ERP-embedded analytics are mostly built on relational data: entities, attributes, aggregates, and statistical patterns across the records the ERP holds. They describe the state of the business: what things are, where they are, and how values trend over time. What was last quarter's revenue by region? When will inventory run out at this rate? Which orders are at highest risk of late shipment given the current backlog? What should our reorder quantity be next month? As AI gets better, these answers become more accessible and more predictive. The system starts to self-adjust shipping priorities and re-order points. It becomes a system of execution alongside a system of record.

Process mining makes similar promises, but it's built on a different foundation. Where ERP analytics sees metrics and trends, process mining adds flow: how work is moving through a business as sequences of activities over time. It can also be used to describe, predict, and prescribe, but it starts from observed flows rather than from a pre-defined notion of how work is supposed to move. It shows every path an order took through your business, including paths you didn't know existed. It compares how a process was supposed to run against how it actually ran. It can answer the same questions about stock-outs, late delivery, and cycle time drift, but the diagnosis is different.

ERP analytics will show you which orders shipped late and predicts which ones are at risk. Process mining will also show you the path those late orders took.

Each leads to a different response. You can react to late shipments by adjusting reorder points, escalating at-risk orders, and tightening forecasts. And some of those adjustments may genuinely need to happen. But without a flow-based analysis, you can't see whether you're adjusting a real constraint or working around a broken process. A flow-based analysis shows you where the process itself may need to change.

That's why every major enterprise platform decided process mining needed to be acquired alongside the analytics they already had. When Salesforce announced its acquisition of Apromore in late 2025, their President and Chief Product Officer, explained the reasoning directly: process mining gives customers the end-to-end visibility they need to understand how their business actually operates, and that visibility is what lets agentic automation measure, optimise, and act on real processes rather than assumed ones.

Mid-Market ERPs and Process Mining

We haven't seen as many process mining acquisitions amongst the ERPs serving the mid-market, the ones used by manufacturers and other businesses operating at a smaller scale. A couple of vendors at the top of the segment have moved, for example QAD acquired LiveJourney in 2022, and Infor built process mining into its platform. But their offerings are again aimed at the larger end of the mid-market.

Below that tier, the picture is different. The ERPs serving smaller businesses have either bought or built adjacent capabilities like demand forecasting, predictive maintenance, and inventory optimisation, but nothing publicly framed as process mining.

What's visible from the outside is significant investment in better analytics and AI assistants, but not in process mining specifically. Customers buying these systems today are getting better dashboards and better systems of execution. But they're not being offered process mining as part of the deal.

The practical effect is that a smaller business running a mid-market ERP is operating without a capability their larger peers now consider standard.

Final Thought

The major enterprise platforms looked at process mining and decided it was important enough to acquire alongside their existing analytics stack.

Below that, a few platforms serving the upper mid-market have started to move into process mining, but the broader mid-market segment hasn't.

Which means the interesting question for the next few years isn't whether the mid-market will get process mining. It's who will bring it to them.