For organizations looking to automate business processes, the buzzword “process mining” has been gaining popularity of late. It is the next logical step after robotic process automation (RPA), enterprise service management (ESM), and the various approaches taken to automate tasks involving enterprise resource planning (ERP) systems. At the core of this push toward automation is the reality that many enterprises are combinations of hundreds, thousands or even millions of “processes” and that IT’s biggest headache is replacing the rusted bailing wire and hardened chewing gum that keep them all stuck together.
The idea behind process mining is to port all the clever work that has gone into “data mining” and bolt it onto an RPA or ERP system to do a better job looking for panic triggers. Process mining software can be smart enough to recognize when something bad is happening.
But there’s more. Process mining code may be smart enough to recognize and define just what qualifies as a “process” so we don’t need to spend months or years writing out spec files. Process mining code can just grok how the data moves to understand how order data flows link with shipping data flows and then turn into banking data flows so everyone can get paid. If your jobs have unique and consistent ID numbers, the tools can track the work from process to process.
The insights that come from these AI algorithms are not final and they don’t come from a mysterious black box. The tools dump their results into flowcharts describing the processes, often the ones used by RPA tools, becoming the starting point for you to tweak and refine. The AIs may not be clever enough to know what you’re looking to understand about your enterprise, but they can do plenty of basic work to lay a foundation.
The insights aren’t also earth shattering but they’re still very valuable. A common result can flag bottlenecks and process interruptions, details that are fairly obvious to any human who is paying attention, but the point of the systems is to relieve humans from the constant guard duty. The deeper insights about reorganization of the whole workflow are left to you and your team of humans but the code will give you plenty of hints as to which traffic jams to fix.
Here are 6 process mining tools that automate those chores and take the burden of constant watchfulness off of our shoulders.
Celonis Execution Management System
Execution Management System (EMS) from Celonis adds a layer of AI to an enterprise by organizing business events from ERP, CRM or accounting systems into a form that can be tracked and analyzed by the core AI tools. The foundation is a collection of pre-built data collectors for both the big tools in the server rooms (SAP, Oracle, Salesforce, etc.) and the little ones on the desktops (Outlook, Excel, etc).
When tracking data is available to link events, Celonis EMS can discover the flow of data and flag where “friction” enters the system. Process Discovery begins with visualization and adds a layer of statistical tracking for identifying deviations from the norm that can identify the root cause. AI starts to appear as Jupyter Notebooks that can access the process events and analyze them using various open source Python models.
Celonis EMS is available in several different versions. The main Enterprise product harnesses the collection integration with all data collections and the price is available from the sales team. An App Store offers additional modules and collections of routines that can speed integration. Academic users and small businesses can explore free versions tailored to either lab research or businesses with smaller data flows (500MB from flat log files).
Kofax Insight offers a collection of tools for streamlining business processes with RPA and what they call TotalAgility. Many of the tools, such as Invoice Portal, OmniPage OCR and SignDoc, help juggle the endless flow of paper while replacing it in digital form where it can be more efficiently followed and watched. The Mobile Capture Platform, for instance, integrates cell phone cameras with document processing to speed the job of front-line workers.
The analytics for tracking behavior is bundled in a product called “Insight” built from three layers: data gathering (Analytics Server), numerical summarization (Metrics Mart), and presentation (Dashboard). The tools are optimized for code-free development so more team members can turn raw data into useful metrics. The Dashboard integrates several dozen types of charts for fast visual tracking.
Kofax prices are available from their sales team.
myInvenio offers a multi-level mining tool that’s built to make sense of low-level subprocesses, such as invoicing or inventory tracking, as well as higher-level processes, joining them all together where failures may also occur. The heart of the algorithm is the Decision Rule Miner that looks first for rules that explain what is happening and then analyzes how accurate the rules might be.
Many users will want to turn on the tool’s simulation feature, which enables you to play with the model and its rules to see how it performs in extreme situations. Could a surge of orders crash some tools? Are there any dangerous bottlenecks? The company calls this a “Digital Twin” of the organization for experimentation.
myInvenio offers a free trial and two different billing billing models that charge by either process or event.
PAFNow is integrated with Microsoft’s enterprise offerings and deeply embedded in tools such as the Office 365 suite. The main product is built on top of Power BI adding artificial intelligence to help analyze process flows. Users can run the tool locally within Office 365 or in the cloud making it accessible for anyone used to standard Office tools.
The core can be extended to gather information from other systems, such as SAP (“Connectors”), or to trigger other workflows when the numbers justify it (“Action Connectors”). There are also deeper thematic extensions focused on particular jobs such as the Internal Audit Edition. The results appear in customized dashboards, PDF reports, or PowerPoint presentations.
PAFNow offers two editions. Premium is aimed at businesses with more localized data sources, while Enterprise is more deeply integrated with larger data sources such as SAP. Prices for both are available from the sales team.
Signavio Process Intelligence
Business Transformation Suite from Signavio is a collection of tools that turns raw data into collaborative dashboards and reports. The part responsible for applying AI to the data to sense trouble and track changes intelligently is called Process Intelligence. This tool adds an additional dynamic layer of algorithms to “The Hub,” Signavio’s main visual tool where widgets are dragged into a flowchart tracking your enterprise data flow.
The tool encourages teams to collaborate on modeling the workflow through the Process Manager and setting the thresholds that trigger alerts. Process Intelligence widgets watch the model and offer real-time, interactive insights into the data flowing through the network.
Pricing for Business Transformation Suite is available from the sales team.
UiPath Process Mining (formerly ProcessGold)
UiPath has one of the most elaborate tools for robotic process automation with a complete set of graphical UI tools for creating document- and web-focused automation that can be deployed to a cloud. The workflows are designed in the Studio, handed over to the Orchestrator and then monitored by the Automation Cloud. The company has already been using AI tech such as optical character recognition and natural language processing to guide the robots working on documents and forms.
UiPath’s Process Mining extension works with log files to sketch out workflows automatically, something the company calls taking an “x-ray” of your enterprise. If the log files have a unique case ID attached to each event and a timestamp, the code can extract a useful graph. Even if the results don’t lead to something that can be turned over to a robot, managers can still learn about the bottlenecks and issues inside their operation from this x-ray.
The company is also working to expand the scope to what it calls “task mining,” a closely related approach in which pattern-matching algorithms identify opportunities for automation in the work done by some employees.
Information about pricing comes from the sales team.