Process mining vs process improvement: the question most banks are asking backwards
139 days. That is how long a Kuwait bank's procurement cycle ran before anyone mapped it. No process mining platform was involved — the team described the workflow, ran DMAIC analysis, and applied the E-S-S-A-M framework (Eliminate, Simplify & Standardize, Automate, Migrate). Procurement dropped to 57 days: a 59% cycle-time reduction, 106.9% efficiency improvement, and 7 sign-offs collapsed to 5, all digital.
No data lake. No six-figure licensing contract. No months of ERP log ingestion.
That result surfaces a question most operations leaders never think to ask: what if you have been buying analytics when what you needed was a fix?
The framing problem nobody corrects
The "process mining vs process improvement" debate is usually framed as a tool comparison — enterprise analytics platform versus structured methodology. But that framing confuses two entirely different jobs.
Process mining does one job: it reads event logs from your ERP, CRM, or core banking system and reconstructs what actually happens, step by step, with timestamps. The output is a visual map and a dashboard showing where time is lost, where the process deviates from the intended flow, and how often.
Process improvement does a different job: it fixes what is broken. It finds waste, redesigns the workflow, documents the new standard, and deploys it so staff can operate differently tomorrow.
These are sequentially related, not interchangeable. Mining shows the symptom. Improvement removes the cause. Buying one when you need the other is an expensive mistake — and the banking sector makes it regularly.
Why banks buy mining first (and what it costs them)
Enterprise process mining platforms typically run $100,000–$500,000 per year in licensing, with 3–6 months of implementation before the first dashboard is live. The pitch is compelling: see your processes as they really run, not as your SOPs say they should.
The problem is that a dashboard is not a fix. It is a starting point.
An operations lead described it plainly after spending 8 months configuring a process mining deployment at a mid-size Singapore bank. The platform produced a detailed map of the procurement flow. It showed a 94-day bottleneck concentrated at the compliance desk. The data was accurate. But the team had no structured method for deciding what to change — and the platform's recommendations were directional, not operational. The dashboard showed that two teams were both reviewing the same due diligence materials at different points in the process. It did not design the fix, write the SOP, or tell the compliance leads how to consolidate the work.
Six months after the dashboard went live, the 94-day bottleneck was still 94 days.
A single conversation with ESSAM — describing the workflow, the two redundant review stages, and the missing template — produced a redesigned SOP with the redundant review eliminated and a standardized due diligence checklist deployed. The fix that analytics had identified in 8 months took one session to action.
This is not an argument against mining. It is an argument about sequence.
The sequencing argument
The question is not "process mining or process improvement?" The question is "do you want to know what is broken, or do you want it fixed?"
If your answer is "both," the order matters more than the tools.
| Condition | What you need first | Then |
|---|---|---|
| Process is undocumented | Map and fix (improvement) | Mining is optional |
| Process is documented but clearly broken | Fix (improvement via DMAIC + E-S-S-A-M) | Mining adds monitoring later |
| Process is optimized and you need ongoing deviation detection | Mine (monitoring) | Improvement runs on alerts |
| Process runs across multiple systems with high transaction volume | Mine first (visibility) | Improvement follows |
Most banks in APAC operate in the first two rows, not the third or fourth. Their processes are partially documented, running on a mix of core banking systems and manual handoffs, with no structured improvement methodology applied. For these banks, mining a broken process produces an expensive map of a broken process. Improvement methodology — DMAIC to find the root cause, E-S-S-A-M to redesign and deploy — produces an operating fix.
Process mining platforms require clean, structured event-log data from ERP or core banking systems. Many mid-tier APAC banks have siloed systems, manual entry stages, and incomplete logs. Mining these environments yields incomplete maps, which require significant data-engineering investment before the analytics are trustworthy. That work costs time and budget before a single process has changed.
ESSAM works differently. The input is a conversation. An operations leader describes a process — the steps, the handoffs, the approvals, the exceptions. ESSAM baselines it, runs the E-S-S-A-M analysis, identifies waste categories, and returns a redesigned SOP with an audit trail. No ERP integration required. No data engineering prerequisite.
What ESSAM produces vs what process mining platforms produce
The output difference matters for banking operations teams. It determines what the team does next.
Process mining platform output:
- Visual process map (discovered from logs)
- Deviation frequency charts
- Bottleneck identification by step and time
- Conformance analysis (how often the process follows the designed path)
- Dashboard for ongoing monitoring
ESSAM output:
- Measured process baseline (cycle time, steps, handoffs, sign-offs)
- Waste map coded to E-S-S-A-M categories
- Redesigned SOP with specific changes (roles, steps, templates, approval routing)
- Deployed documentation your team operates from immediately
- Audit trail for regulatory and compliance purposes
Mining produces analytics. ESSAM produces deployed operations change. Neither replaces the other — but for a bank that needs a process to run differently by next month, one output is immediately usable and the other is a diagnostic starting point.
The Kuwait bank's procurement team needed a process to run differently. They had a 139-day cycle harming vendor relationships and internal planning. DMAIC identified waste at every stage: redundant approvals, missing templates, paper-based sign-offs, unclear ownership. The E-S-S-A-M redesign eliminated the redundant approval layer, standardized templates, digitized all 5 remaining sign-offs, and produced SOPs deployed across the team. Fifty-seven days. No mining platform in the stack.
When process mining is the right call
This post is not a case against process mining. There are genuine use cases where the analytics-first approach is correct and ESSAM is the wrong starting point.
High transaction volume, multiple systems. If a bank is processing 50,000 loan applications per month across three core banking systems and a document management platform, mining the event logs reveals patterns no human analysis can surface. Variance by channel, by branch, by product type: that is legitimate mining territory.
Post-optimization monitoring. After a process has been redesigned and deployed using structured improvement methods, mining provides the ongoing conformance signal. You want to know when the new SOP starts drifting. Mining alerts operations leaders before drift becomes systemic.
Regulatory audit and process verification. Some regulatory environments require documented evidence of how a process actually ran, not how it was designed to run. Mining the ERP logs produces that evidence.
In all three cases, the process should be stable, documented, and operating on clean system data before mining adds value. Mining a poorly documented, high-variance process generates noise before signal.
How to decide: a practical test for your team
Before committing budget to either approach, run this three-question test.
1. Is the process currently documented to step level? If no: start with ESSAM or a structured DMAIC engagement. Document and fix before you mine.
2. Do you have clean, timestamped event logs from the systems that run this process? If no: mining will require data-engineering investment before the analytics are reliable. Factor that cost into the timeline.
3. Are you trying to fix the process now, or monitor it for future drift? If fix: ESSAM. Describe the process, get the redesigned SOP, deploy it. Pricing starts at $40/month. If monitor: mining platforms are designed for this. Expect $100,000–$500,000/year and a 3–6 month implementation before dashboards are operational.
The Kuwait bank answered these questions before committing. The process was partially documented, the event data was inconsistent, and the goal was to fix cycle time, not monitor for deviation. DMAIC plus E-S-S-A-M was the correct choice. It was also the faster and lower-cost choice by a significant margin. See the full outcome in the ESSAM case studies.
The honest limitation
ESSAM's conversational model works for processes that an operations leader can describe and that run on human decisions and handoffs. It is not a monitoring tool. It does not read ERP logs, detect real-time deviations, or alert you when a process starts drifting from its new standard.
If you have 200 branches each running variations of the same process and you need to know which branches are drifting farthest from the new SOP, you need a monitoring layer that ESSAM does not currently provide. That is a legitimate use case for process mining platforms layered on top of an ESSAM-designed operating standard.
The practical sequence for most APAC banks: fix the process with ESSAM first, deploy the SOP, then evaluate whether transaction volume and system-data quality justify adding a mining layer for ongoing monitoring. Buying mining before fixing the process is the expensive path. Buying it after fixing the process is the rational one.
Run the fix before you run the analysis
Most operations teams discover, somewhere in month 4 or 5 of a mining implementation, that the dashboards are showing them things they already knew. The 94-day compliance bottleneck. The redundant approval loop. The missing template that forces rework. The data confirms the intuition — but the intuition was available on day one.
ESSAM's 7-step cycle — Baseline, Analyze, Optimize, Document, Approve, Deploy, Repeat — runs in one session for a single process. The Kuwait result took one structured engagement. The Singapore bank described above spent 8 months to arrive at the same diagnosis, then needed another session to produce the fix.
Sequence matters. Fix before you mine if the process is broken. Mine before you fix only if the process is optimized and you need deviation detection. Knowing which situation you are in is the decision.
Explore ESSAM pricing and plans
Map one broken process to a deployed SOP
Describe one process that is taking too long — the vendor approval chain, the compliance sign-off sequence, the account-opening handoffs. ESSAM returns a measured baseline, a waste map coded against the E-S-S-A-M framework, and a redesigned SOP your team can operate from. One conversation, no data-engineering prerequisite, no consulting retainer.
Turn one slow process into a deployed SOP
Frequently asked questions
What is the difference between process mining and process improvement?
Process mining uses event-log data from enterprise systems to reconstruct how a process actually runs, producing visual maps and deviation analytics. Process improvement uses structured methodologies — such as DMAIC combined with the E-S-S-A-M framework — to identify waste, redesign the workflow, and deploy an updated operating standard. Mining diagnoses; improvement fixes. They are most valuable in sequence, not as alternatives.
Can you do process improvement without process mining?
Yes. Process improvement methodologies do not require event-log data. A structured description of the workflow, analyzed against waste categories (Eliminate, Simplify & Standardize, Automate, Migrate), identifies where time and cost are lost. The Kuwait bank reduced a 139-day procurement cycle to 57 days using DMAIC and E-S-S-A-M with no process mining platform in the stack.
When does it make sense to use process mining in banking?
Process mining is most valuable when a process is already documented, running on clean and complete system-event data, and the goal is ongoing deviation monitoring or cross-branch conformance analysis at high transaction volumes. For processes that are undocumented, partially broken, or running on mixed manual and digital handoffs — which describes most mid-tier APAC bank processes — improvement methodology delivers faster and lower-cost results.
How long does process improvement take with ESSAM compared to process mining platforms?
Enterprise process mining platforms typically require 3–6 months of implementation before operational dashboards are live, at $100,000–$500,000 per year in licensing. ESSAM delivers a baseline, waste map, and redesigned SOP in a single session, with plans starting at $40/month. For banks that need a process to run differently within weeks, the timelines are not comparable.
Does ESSAM replace process mining for banking compliance and audit?
No. Process mining platforms that read ERP event logs provide a specific type of regulatory evidence — a timestamped record of how a process actually ran — that ESSAM does not replicate. For compliance environments that require log-based process verification, mining platforms serve a distinct purpose. ESSAM is designed for process design, improvement, and SOP deployment, not for real-time log-based conformance monitoring.
Related reading: