From structured process data to operational intelligence and automation decisions.
Process Intelligence is a portfolio project that transforms structured operational process data into process visualization, operational analytics and automation recommendations.
The project explores a practical question:
What if process mapping could become structured data that can be validated, analyzed and used to support process improvement and automation decisions?
Instead of treating a process map only as documentation, Process Intelligence connects process design with operational data, workload indicators, Lean analysis, business rules and automation assessment.
V1 MVP completed
The current version supports the complete flow:
Excel Template
|
v
Structural Validation
|
v
Process Data Model
|
v
Process Visualization
|
v
Operational Analytics
|
v
Automation Assessment
|
v
Automation Prioritization
|
v
Executive Recommendation
Operational processes are often documented across spreadsheets, presentations, interviews and diagrams.
These formats are useful for describing how a process works, but they frequently make it difficult to consistently answer questions such as:
- Where is operational effort concentrated?
- Which activities generate waiting or rework?
- Which parts of the process deserve further investigation?
- Which activities have characteristics suitable for automation?
- Which automation opportunities should be evaluated first?
- Are process stages and connections structurally consistent?
Process Intelligence addresses this by turning process documentation into structured process data.
The application uses a standardized Excel workbook as its process data source.
The main sheets are:
PROCESSOETAPASATIVIDADESCONEXOESREGRAS
Additional guidance sheets help users understand how the template should be completed.
A downloadable example template is available directly from the application.
Before the process can be analyzed, Process Intelligence validates its structure.
Current checks include:
- Required sheets
- Required fields
- Duplicate identifiers
- Invalid stage references
- Invalid connection references
- Self-referencing connections
- BPMN-type consistency
- Start and end events
- Branch probabilities
- Process-flow structural consistency
Critical structural errors prevent the analysis from running.
The application automatically generates a process map from the structured Excel data.
The visualization currently supports:
- Swimlanes by responsible area
- Start events
- End events
- Tasks
- Exclusive gateways
- Sequence flows
- Conditions
- Branch probabilities
- Rework and return flows
- Zoom controls
- Fit-to-screen visualization
The map uses conventions inspired by BPMN.
It is intentionally described as a simplified BPMN representation and is not yet a formal BPMN 2.0 model or interoperable BPMN file.
Process Intelligence estimates monthly operational workload at activity level.
The calculation considers:
Execution Time
x People
x Monthly Volume
x Occurrence Rate
-----------------
60
The result represents estimated monthly human effort in hours.
Waiting time is analyzed separately and is not automatically considered human workload.
Waiting Time != Human Effort
This distinction prevents process lead time from being incorrectly interpreted as human effort.
Activities can be classified as:
- Value Added
- Necessary
- Waste
- Unclassified
Instead of simply counting activities, the Lean distribution uses estimated monthly workload.
This helps show where operational effort is actually concentrated.
The application identifies activities that may deserve additional investigation.
The assessment combines signals such as:
- Workload
- Waiting time
- Criticality
- SLA proximity
- Rework
- Waste
- Manual execution
These results are intentionally presented as bottleneck signals, not definitive bottlenecks.
A high-workload activity may simply be a high-volume core activity.
Operational context is still required before making a final decision.
Each activity is evaluated using characteristics such as:
- Repetitiveness
- Rule-based execution
- Structured data availability
- Human judgment requirements
- Execution type
The result indicates whether an activity has low, medium or high automation suitability.
Automation suitability alone does not define implementation priority.
Process Intelligence separates automation suitability from operational impact.
Automation Suitability
+
Operational Impact
=
Automation Priority
The prioritization model combines both dimensions before ranking opportunities.
This prevents an activity from becoming a top priority only because it is technically easy to automate.
The application converts analytical results into initial recommendations.
Possible approaches include:
- Deterministic validation
- System integration
- Workflow automation
- Assisted automation
- Process redesign before automation
The recommendation engine is deterministic and rule-based in the current V1.
It does not automatically claim financial savings or guaranteed automation benefits.
Actual benefits should be validated through process-owner review, technical assessment and pilot measurement.
The provided workbook represents a synthetic operational request process containing:
- 12 process stages
- 18 operational activities
- 13 process connections
- 5 business rules
- Multiple responsible areas
- Decision points
- Rework flows
- Waiting time
- Operational workload
- Automation characteristics
All example data is synthetic and intended only for demonstration and portfolio purposes.
Excel Workbook
|
v
Excel Reader
|
v
Process Data Parser
|
+-----------------------+
| |
v v
Template Validator Process Flow Model
| |
v v
Process Analytics Process Visualization
|
v
Automation Scoring
|
v
Priority Engine
|
v
Executive Recommendation Engine
|
v
Next.js User Interface
The current version processes the uploaded workbook directly in the browser.
No backend is required for V1.
- Next.js
- React
- TypeScript
- Tailwind CSS
- read-excel-file
- SVG-based process visualization
- Git
- GitHub
The process map is generated without requiring an external BPMN visualization library.
process-intelligence/
|
+-- docs/
| +-- data-contract.md
| +-- solution-design.md
|
+-- examples/
| +-- Process_Intelligence_Template_V2.xlsx
|
+-- web/
| +-- public/
| +-- src/
| | +-- app/
| | +-- components/
| | +-- lib/
| |
| +-- package.json
|
+-- README.md
+-- LICENSE
Clone the repository:
git clone https://github.com/YuryBdS/process-intelligence.gitEnter the application directory:
cd process-intelligence/webInstall dependencies:
npm installStart the development server:
npm run devOpen:
http://localhost:3000
On Windows environments where PowerShell blocks npm.ps1, the equivalent commands can be executed with:
npm.cmd install
npm.cmd run devThe commands above are installation instructions for someone cloning the project. They do not need to be executed when editing this README.
V1 has been validated with:
Production build OK
TypeScript validation OK
ESLint 0 errors / 0 warnings
Production dependencies 0 known vulnerabilities
Validation commands:
npm run lint
npm run build
npm audit --omit=devThe dependency security statement above refers specifically to production dependencies at the time of V1 validation.
Automation decisions should start with a clear representation of the process, activities, rules, inputs, outputs and exceptions.
When deterministic rules can reliably solve the problem, they should be preferred before introducing generative AI.
High workload is an analytical signal, not automatic evidence of a bottleneck.
Waiting time affects process lead time but should not automatically be counted as operational workload.
When an activity exists because of rework or process inefficiency, redesign may be more appropriate than automation.
Every recommendation should be traceable to structured process data and explicit decision rules.
The current version includes:
- Excel-based process modeling
- Downloadable example template
- Structural process validation
- Automatic process visualization
- BPMN-inspired notation
- Swimlanes
- Operational workload analysis
- Lean analysis
- Bottleneck signals
- Automation suitability assessment
- Automation prioritization
- Executive recommendations
- Local browser-based processing
Possible future versions may include:
- Formal BPMN 2.0 generation
- BPMN file export
- Process simulation
- Capacity simulation
- Scenario comparison
- Automated reporting
- PDF or presentation export
- Process benchmarking
- Historical process analysis
- Integration with operational systems
- AI-assisted process discovery
- AI-assisted process documentation
These capabilities are intentionally outside the V1 scope.
Process Intelligence was created as a portfolio project focused on the intersection of:
Process Management + Data + Operational Analytics + Automation
The objective is not only to draw process maps.
The project demonstrates how structured process information can be transformed into explainable analytical signals and better automation decisions.
Yury Batista
Business processes, operational analysis, data and automation.