Process Mining Without Execution Is Just Documentation
Process Mining Without Execution Is Just Documentation
A founder perspective on why most European enterprise process mining investments produce dashboards instead of operational change — and the architecture pattern that closes the gap.
By Nikola Dlaka, co-founder, DNA Automation Consulting
Most enterprise process mining deployments produce dashboards. Not change.
A six-figure process mining license becomes a quarterly status update with very nice visuals. The dashboard shows the bottleneck. The variance. The deviation from the canonical path. The procurement officer signs off on the renewal. The process keeps running the way it always ran.
This is not a critique of the technology. The process mining category crossed one billion dollars globally in 2025 and is growing fast for a reason — the discovery problem it solves is real. Without process mining, most enterprises cannot see how their processes actually run versus how they think their processes run. The variance is invisible. Mining makes it visible.
But visibility is not change. And in most European enterprise deployments, change is what was actually bought.
Three things process mining alone cannot do
Process mining is a discovery layer. It does one job, well. The mistake is assuming that the discovery layer alone produces operational change. Three things mining cannot do.
It finds the bottleneck. It cannot change the BPMN that produced it.
The process model that creates the variance has to be redesigned. The redesigned model has to be deployed to an engine. The engine has to execute the new flow against the operational systems. The change is in the redesign and the execution, not in the heat map.
It produces the heat map. It cannot orchestrate the work.
The work runs in the bank's core banking platform, in the energy operator's SAP workflows, in the manufacturer's ServiceNow flows, in the insurer's Camunda runtime. Mining sees the trail of execution. It does not run the execution. That is a different layer of the stack.
It proves the variance exists. It cannot enforce the canonical path.
Enforcement is a workflow engine's job. The engine has the process model, the state, the rules, the exception handlers. The engine routes the work. Mining tells you the work is being routed wrong. The engine makes it route right.
Each of these three things is a different system. Mining is one of the systems. It is not all of them.
Why the gap is architectural
The enterprises that get value from process mining are not the ones with the best mining tool. They are the ones with the architecture that turns the mining output into operational change.
The architecture has three layers.
Discovery is the mining layer. Celonis dominates the European process mining market, with its valuation peaking above $13 billion in 2022 before settling in the high single-digit billions by mid-2024. But the category extends beyond Celonis — UiPath Process Mining, IBM Process Mining, SAP Signavio Process Intelligence, ARIS Process Mining, ABBYY, and a long tail of specialised vendors all compete on similar capability. The mining tool surfaces the variants. It does not change the variants.
Design is the BPMN layer. The redesign of the process model lives in ARIS, in Signavio, in BIC Platform, in a modeling tool the enterprise already owns. The redesigned model is the bridge between what mining showed and what execution will do.
Execution is the engine layer. The engine — Camunda 7, Camunda 8, Concordia BPM, a SAP banking workflow engine, a custom orchestrator — runs the redesigned process against the operational systems. The engine produces the audit trail that proves the new process is in fact running the new way.
Most European enterprise process mining deployments stop at layer one. The mining renewal continues. The findings stay in PowerPoint. The redesign is never executed. The dashboard becomes documentation of a state the enterprise did not change.
The execution claim
A specific version of this problem applies to enterprises that bought process mining at scale.
Several process mining vendors have positioned their category as execution-driving, not just discovery-driving. The framing dates from 2020–2022 — the period when valuations were at peak. The promise was that mining would not stop at discovery; it would drive the execution layer too.
In practice, most large-scale process mining deployments at European enterprises have not delivered the execution claim. The discovery layer works. The dashboards work. The bottleneck identification works. The actual operational change — the redesign of the process, the deployment of a new model to a workflow engine, the enforcement of the canonical path — is a different category of work, and most mining customers do not do that work inside the mining tool. They do it (or fail to do it) somewhere else.
This is not the mining vendors' failure. The mining tool is a discovery tool. The execution work is BPM work. The two layers belong to different stacks, run on different platforms, and require different skills.
What the market needs is a more honest framing of what each tool actually does.
The mining → BPMN → engine pipeline
The pipeline that produces operational change has four stages.
Stage one: mining identifies the process variance. The output is a heat map, a variant analysis, a quantified bottleneck.
Stage two: process redesign happens in a BPMN modeling tool. The output is a new process model — a BPMN diagram that closes the variance the mining found. This is the layer where the consultancy adds value. The mining tool surfaces the problem; the BPMN model defines the solution.
Stage three: engine deployment. The new BPMN model is deployed to a workflow engine. The engine runs the new flow against the operational systems. This is where execution starts.
Stage four: operational change. The new flow runs in production. The operational systems behave the way the new BPMN model defines. The mining tool now sees a different variant pattern — proving the change took effect.
Each stage is a different platform. Each stage requires different work. The whole pipeline produces the change. No subset of the pipeline produces the change.
The DNA position
DNA is vendor-neutral on the mining tool. Celonis fits some buyer scenarios. SAP Signavio Process Intelligence fits SAP-heavy estates. IBM Process Mining fits IBM-aligned stacks. ARIS Process Mining is the natural fit for ARIS modeling estates. The right mining tool depends on the enterprise's existing platform commitments.
What DNA does is what most mining customers actually need next: bridge from mining findings to BPMN redesign to engine execution. Discovery, Design, Automation — the three services DNA delivers, run in sequence across the pipeline above.
DNA operates across ARIS, Camunda, ServiceNOW, SAP Signavio, and Microsoft Power Platform — and is also the developer of Concordia BPM for enterprises that want a maintained Camunda v7-compatible engine with an open AI-powered platform layer on top. The execution work that closes the loop from mining can run on whichever platform fits the enterprise.
The honest framing for European enterprise procurement of process mining: buy the mining tool. Do not buy the promise that mining produces operational change. The change happens in a different layer. Plan for that layer at the same time, or the mining investment will renew without ever producing the result it was bought to produce.
Book a meeting → https://www.dnaconsulting.io/contact
About DNA Automation Consulting
DNA Automation Consulting is an EU-based BPM consultancy specialising in process discovery, BPMN design, and automation delivery for enterprise clients in banking, insurance, energy, telecom, and manufacturing. 75% of client engagements last five years or more.
Founded by Ante Gudelj and Nikola Dlaka. Clients include Proximus, Equinor, Merck, and NEXI.
Frequently asked questions
A founder perspective on why most European enterprise process mining investments produce dashboards instead of operational change — and the architecture pattern that closes the gap.
By Nikola Dlaka, co-founder, DNA Automation Consulting
Most enterprise process mining deployments produce dashboards. Not change.
A six-figure process mining license becomes a quarterly status update with very nice visuals. The dashboard shows the bottleneck. The variance. The deviation from the canonical path. The procurement officer signs off on the renewal. The process keeps running the way it always ran.
This is not a critique of the technology. The process mining category crossed one billion dollars globally in 2025 and is growing fast for a reason — the discovery problem it solves is real. Without process mining, most enterprises cannot see how their processes actually run versus how they think their processes run. The variance is invisible. Mining makes it visible.
But visibility is not change. And in most European enterprise deployments, change is what was actually bought.
Three things process mining alone cannot do
Process mining is a discovery layer. It does one job, well. The mistake is assuming that the discovery layer alone produces operational change. Three things mining cannot do.
It finds the bottleneck. It cannot change the BPMN that produced it.
The process model that creates the variance has to be redesigned. The redesigned model has to be deployed to an engine. The engine has to execute the new flow against the operational systems. The change is in the redesign and the execution, not in the heat map.
It produces the heat map. It cannot orchestrate the work.
The work runs in the bank's core banking platform, in the energy operator's SAP workflows, in the manufacturer's ServiceNow flows, in the insurer's Camunda runtime. Mining sees the trail of execution. It does not run the execution. That is a different layer of the stack.
It proves the variance exists. It cannot enforce the canonical path.
Enforcement is a workflow engine's job. The engine has the process model, the state, the rules, the exception handlers. The engine routes the work. Mining tells you the work is being routed wrong. The engine makes it route right.
Each of these three things is a different system. Mining is one of the systems. It is not all of them.
Why the gap is architectural
The enterprises that get value from process mining are not the ones with the best mining tool. They are the ones with the architecture that turns the mining output into operational change.
The architecture has three layers.
Discovery is the mining layer. Celonis dominates the European process mining market, with its valuation peaking above $13 billion in 2022 before settling in the high single-digit billions by mid-2024. But the category extends beyond Celonis — UiPath Process Mining, IBM Process Mining, SAP Signavio Process Intelligence, ARIS Process Mining, ABBYY, and a long tail of specialised vendors all compete on similar capability. The mining tool surfaces the variants. It does not change the variants.
Design is the BPMN layer. The redesign of the process model lives in ARIS, in Signavio, in BIC Platform, in a modeling tool the enterprise already owns. The redesigned model is the bridge between what mining showed and what execution will do.
Execution is the engine layer. The engine — Camunda 7, Camunda 8, Concordia BPM, a SAP banking workflow engine, a custom orchestrator — runs the redesigned process against the operational systems. The engine produces the audit trail that proves the new process is in fact running the new way.
Most European enterprise process mining deployments stop at layer one. The mining renewal continues. The findings stay in PowerPoint. The redesign is never executed. The dashboard becomes documentation of a state the enterprise did not change.
The execution claim
A specific version of this problem applies to enterprises that bought process mining at scale.
Several process mining vendors have positioned their category as execution-driving, not just discovery-driving. The framing dates from 2020–2022 — the period when valuations were at peak. The promise was that mining would not stop at discovery; it would drive the execution layer too.
In practice, most large-scale process mining deployments at European enterprises have not delivered the execution claim. The discovery layer works. The dashboards work. The bottleneck identification works. The actual operational change — the redesign of the process, the deployment of a new model to a workflow engine, the enforcement of the canonical path — is a different category of work, and most mining customers do not do that work inside the mining tool. They do it (or fail to do it) somewhere else.
This is not the mining vendors' failure. The mining tool is a discovery tool. The execution work is BPM work. The two layers belong to different stacks, run on different platforms, and require different skills.
What the market needs is a more honest framing of what each tool actually does.
The mining → BPMN → engine pipeline
The pipeline that produces operational change has four stages.
Stage one: mining identifies the process variance. The output is a heat map, a variant analysis, a quantified bottleneck.
Stage two: process redesign happens in a BPMN modeling tool. The output is a new process model — a BPMN diagram that closes the variance the mining found. This is the layer where the consultancy adds value. The mining tool surfaces the problem; the BPMN model defines the solution.
Stage three: engine deployment. The new BPMN model is deployed to a workflow engine. The engine runs the new flow against the operational systems. This is where execution starts.
Stage four: operational change. The new flow runs in production. The operational systems behave the way the new BPMN model defines. The mining tool now sees a different variant pattern — proving the change took effect.
Each stage is a different platform. Each stage requires different work. The whole pipeline produces the change. No subset of the pipeline produces the change.
The DNA position
DNA is vendor-neutral on the mining tool. Celonis fits some buyer scenarios. SAP Signavio Process Intelligence fits SAP-heavy estates. IBM Process Mining fits IBM-aligned stacks. ARIS Process Mining is the natural fit for ARIS modeling estates. The right mining tool depends on the enterprise's existing platform commitments.
What DNA does is what most mining customers actually need next: bridge from mining findings to BPMN redesign to engine execution. Discovery, Design, Automation — the three services DNA delivers, run in sequence across the pipeline above.
DNA operates across ARIS, Camunda, ServiceNOW, SAP Signavio, and Microsoft Power Platform — and is also the developer of Concordia BPM for enterprises that want a maintained Camunda v7-compatible engine with an open AI-powered platform layer on top. The execution work that closes the loop from mining can run on whichever platform fits the enterprise.
The honest framing for European enterprise procurement of process mining: buy the mining tool. Do not buy the promise that mining produces operational change. The change happens in a different layer. Plan for that layer at the same time, or the mining investment will renew without ever producing the result it was bought to produce.
Book a meeting → https://www.dnaconsulting.io/contact
About DNA Automation Consulting
DNA Automation Consulting is an EU-based BPM consultancy specialising in process discovery, BPMN design, and automation delivery for enterprise clients in banking, insurance, energy, telecom, and manufacturing. 75% of client engagements last five years or more.
Founded by Ante Gudelj and Nikola Dlaka. Clients include Proximus, Equinor, Merck, and NEXI.
Frequently asked questions
What is process mining and what does it do?
Process mining is a category of software that analyses event logs from operational systems (ERP, CRM, banking platforms, ticketing systems, custom Java) to reconstruct how business processes actually execute in production. The output is a visual representation of the real process flow, including the canonical path, the variants, the bottlenecks, the exceptions, and the deviations from the documented process model. Major vendors include Celonis, SAP Signavio Process Intelligence, UiPath Process Mining, IBM Process Mining, ARIS Process Mining, and ABBYY.
Why doesn't process mining produce operational change on its own?
Process mining is a discovery layer. It identifies what is happening, where the bottlenecks are, and where the variance lies. It does not redesign the process model that produced the variance, it does not orchestrate the operational work, and it does not enforce the canonical path. Each of these is a different layer of the automation stack — BPMN redesign in a modeling tool, execution in a workflow engine, enforcement through the engine's rules and state management. Mining surfaces the problem; the rest of the stack produces the change.
How does process mining differ from BPM?
Process mining is discovery. BPM (business process management) is design plus execution. Mining shows what processes are actually running and where they break. BPM defines the desired process in BPMN 2.0, deploys the model to a workflow engine, and orchestrates the execution against the operational systems. The two categories are sequential, not competitive — mining feeds BPM, and the change happens when BPM acts on what mining surfaces.
How big is the process mining market?
The global process mining market crossed one billion dollars in 2025 and continues to grow at strong double-digit CAGR through 2026 and beyond, with market sizing varying by methodology across analyst firms (estimates for 2026 range from approximately $1 billion to several billion depending on what is counted in the category boundary). Celonis is the most-cited category leader, with significant European market share. The category is highly concentrated at the top, with a long tail of specialised vendors.
What is the mining → BPMN → engine pipeline?
The pipeline that converts process mining findings into operational change has four stages. Mining identifies the process variance and produces a quantified output. Process redesign happens in a BPMN modeling tool (ARIS, Signavio, BIC Platform, Camunda Modeler), producing a new BPMN 2.0 process model. The redesigned model is deployed to a workflow engine (Camunda 7, Camunda 8, Concordia BPM, SAP banking workflow, or another orchestrator). The engine executes the new flow against the operational systems, and operational change follows. Each stage requires a different platform and a different category of work.
How does DNA approach the discovery-to-execution gap?
DNA's three services — Discovery, Design, Automation — run in sequence across the mining-to-execution pipeline. Discovery uses the enterprise's existing mining investment (Celonis, Signavio Process Intelligence, ARIS Process Mining, or another tool) plus DNA's own discovery methodology. Design produces the redesigned BPMN process model in the enterprise's modeling system of record (typically ARIS or Signavio). Automation deploys the redesigned model to a workflow engine and operates it against the production systems. The engine choice depends on the existing stack — Camunda 7, Camunda 8, Concordia BPM, or another platform. DNA's value is bridging the layers, not selling the mining tool.
What is process mining and what does it do?
Process mining is a category of software that analyses event logs from operational systems (ERP, CRM, banking platforms, ticketing systems, custom Java) to reconstruct how business processes actually execute in production. The output is a visual representation of the real process flow, including the canonical path, the variants, the bottlenecks, the exceptions, and the deviations from the documented process model. Major vendors include Celonis, SAP Signavio Process Intelligence, UiPath Process Mining, IBM Process Mining, ARIS Process Mining, and ABBYY.
Why doesn't process mining produce operational change on its own?
Process mining is a discovery layer. It identifies what is happening, where the bottlenecks are, and where the variance lies. It does not redesign the process model that produced the variance, it does not orchestrate the operational work, and it does not enforce the canonical path. Each of these is a different layer of the automation stack — BPMN redesign in a modeling tool, execution in a workflow engine, enforcement through the engine's rules and state management. Mining surfaces the problem; the rest of the stack produces the change.
How does process mining differ from BPM?
Process mining is discovery. BPM (business process management) is design plus execution. Mining shows what processes are actually running and where they break. BPM defines the desired process in BPMN 2.0, deploys the model to a workflow engine, and orchestrates the execution against the operational systems. The two categories are sequential, not competitive — mining feeds BPM, and the change happens when BPM acts on what mining surfaces.
How big is the process mining market?
The global process mining market crossed one billion dollars in 2025 and continues to grow at strong double-digit CAGR through 2026 and beyond, with market sizing varying by methodology across analyst firms (estimates for 2026 range from approximately $1 billion to several billion depending on what is counted in the category boundary). Celonis is the most-cited category leader, with significant European market share. The category is highly concentrated at the top, with a long tail of specialised vendors.
What is the mining → BPMN → engine pipeline?
The pipeline that converts process mining findings into operational change has four stages. Mining identifies the process variance and produces a quantified output. Process redesign happens in a BPMN modeling tool (ARIS, Signavio, BIC Platform, Camunda Modeler), producing a new BPMN 2.0 process model. The redesigned model is deployed to a workflow engine (Camunda 7, Camunda 8, Concordia BPM, SAP banking workflow, or another orchestrator). The engine executes the new flow against the operational systems, and operational change follows. Each stage requires a different platform and a different category of work.
How does DNA approach the discovery-to-execution gap?
DNA's three services — Discovery, Design, Automation — run in sequence across the mining-to-execution pipeline. Discovery uses the enterprise's existing mining investment (Celonis, Signavio Process Intelligence, ARIS Process Mining, or another tool) plus DNA's own discovery methodology. Design produces the redesigned BPMN process model in the enterprise's modeling system of record (typically ARIS or Signavio). Automation deploys the redesigned model to a workflow engine and operates it against the production systems. The engine choice depends on the existing stack — Camunda 7, Camunda 8, Concordia BPM, or another platform. DNA's value is bridging the layers, not selling the mining tool.
What is process mining and what does it do?
Process mining is a category of software that analyses event logs from operational systems (ERP, CRM, banking platforms, ticketing systems, custom Java) to reconstruct how business processes actually execute in production. The output is a visual representation of the real process flow, including the canonical path, the variants, the bottlenecks, the exceptions, and the deviations from the documented process model. Major vendors include Celonis, SAP Signavio Process Intelligence, UiPath Process Mining, IBM Process Mining, ARIS Process Mining, and ABBYY.
Why doesn't process mining produce operational change on its own?
Process mining is a discovery layer. It identifies what is happening, where the bottlenecks are, and where the variance lies. It does not redesign the process model that produced the variance, it does not orchestrate the operational work, and it does not enforce the canonical path. Each of these is a different layer of the automation stack — BPMN redesign in a modeling tool, execution in a workflow engine, enforcement through the engine's rules and state management. Mining surfaces the problem; the rest of the stack produces the change.
How does process mining differ from BPM?
Process mining is discovery. BPM (business process management) is design plus execution. Mining shows what processes are actually running and where they break. BPM defines the desired process in BPMN 2.0, deploys the model to a workflow engine, and orchestrates the execution against the operational systems. The two categories are sequential, not competitive — mining feeds BPM, and the change happens when BPM acts on what mining surfaces.
How big is the process mining market?
The global process mining market crossed one billion dollars in 2025 and continues to grow at strong double-digit CAGR through 2026 and beyond, with market sizing varying by methodology across analyst firms (estimates for 2026 range from approximately $1 billion to several billion depending on what is counted in the category boundary). Celonis is the most-cited category leader, with significant European market share. The category is highly concentrated at the top, with a long tail of specialised vendors.
What is the mining → BPMN → engine pipeline?
The pipeline that converts process mining findings into operational change has four stages. Mining identifies the process variance and produces a quantified output. Process redesign happens in a BPMN modeling tool (ARIS, Signavio, BIC Platform, Camunda Modeler), producing a new BPMN 2.0 process model. The redesigned model is deployed to a workflow engine (Camunda 7, Camunda 8, Concordia BPM, SAP banking workflow, or another orchestrator). The engine executes the new flow against the operational systems, and operational change follows. Each stage requires a different platform and a different category of work.
How does DNA approach the discovery-to-execution gap?
DNA's three services — Discovery, Design, Automation — run in sequence across the mining-to-execution pipeline. Discovery uses the enterprise's existing mining investment (Celonis, Signavio Process Intelligence, ARIS Process Mining, or another tool) plus DNA's own discovery methodology. Design produces the redesigned BPMN process model in the enterprise's modeling system of record (typically ARIS or Signavio). Automation deploys the redesigned model to a workflow engine and operates it against the production systems. The engine choice depends on the existing stack — Camunda 7, Camunda 8, Concordia BPM, or another platform. DNA's value is bridging the layers, not selling the mining tool.
Written by:
Nikola Dlaka
Managing Partner
Share with friends: