1 Synesis: Structuring Knowledge for Data-Driven Decisions

Guide for decision makers and knowledge professionals

1.1 The challenge of the data era

We live in an era characterized by the acronym VUCA: Volatility, Uncertainty, Complexity, and Ambiguity. In this context, making efficient decisions requires more than intuition — it demands the ability to integrate and analyze multiple data sources, identifying not just correlations, but cause-and-effect relationships between observed phenomena.

As highlighted in recent research on Causal Knowledge Graphs, simple correlation between data can lead to wrong conclusions. The classic example: there is statistical correlation between chocolate consumption and number of Nobel Prize winners in different countries — but it would be absurd to conclude that distributing chocolate will increase an organization’s scientific output.

The real challenge lies in capturing causal relationships, not just superficial correlations.

1.2 Where does Synesis language fit in?

Synesis is a structured language for systematic capture of causal knowledge from documentary sources and qualitative observations.

While automated systems like statistical causal discovery work with numerical data, much of organizational knowledge is dispersed in:

Reports and technical documents
Interviews and qualitative surveys
Specialized literature and manuals
Incident records and lessons learned
Engagement and organizational climate surveys

Synesis offers a standardized way to extract, document, and structure causal relationships from these textual sources, making them processable and integrable with other analysis systems.

1.3 The problem Synesis solves

1.3.1 Limitation 1: Qualitative knowledge is not quantifiable

When a causal relationship is extracted from a document (for example: “lack of training causes operational errors”), it is purely qualitative. There’s no way to predict the probability or magnitude of the effect.

How Synesis contributes: By systematically structuring these relationships, they can later be integrated with numerical data to quantify causal effects.

1.3.2 Limitation 2: Statistical discovery can invert cause and effect

Statistical causal discovery algorithms, when applied to biased or incomplete data, can derive relationships where the direction of causality is inverted.

How Synesis contributes: Prior knowledge documented in Synesis serves as “prior knowledge” to validate or correct statistical discoveries.

1.3.3 Limitation 3: Data from different contexts don’t communicate

Traditional causal discovery works with a single dataset. It can’t combine observations from different departments, periods, or conditions.

How Synesis contributes: The standardized structure allows integrating knowledge from multiple sources in a common vocabulary.

1.4 Practical applications

1.4.1 Human Resources Management

Scenario: The HR area needs to recommend initiatives to improve engagement, but each department has different characteristics.

With Synesis: - Document causal relationships from organizational behavior literature - Record observations from engagement surveys with structured interpretations - Identify which factors (career opportunities, vision clarity, teamwork) cause which outcomes (motivation, satisfaction, retention) - Recommend initiatives based on documented evidence

Example record:
- Source: Engagement Survey Q4/2024, Commercial Department
- Observation: "I feel I don't have growth opportunities here"
- Interpretation: Employee expresses frustration with career limitations
- Identified relationship: Lack of Opportunities → REDUCES → Motivation

1.4.2 Failure and Incident Analysis

Scenario: Operations team needs to identify root causes of failures in complex systems.

With Synesis: - Extract causal relationships from technical manuals and troubleshooting guides - Document each incident with the identified causal chain - Build a knowledge base that accelerates future diagnoses

Example record:
- Source: Incident Report #2024-0892
- Observation: Intermittent disconnection during file download
- Interpretation: Pattern consistent with MAC configuration error
- Identified relationship: Configuration Error → CAUSES → Disconnection

1.4.3 Market Analysis and Negotiations

Scenario: Sales team wants to understand why some negotiations succeed and others don’t.

With Synesis: - Record observations from each negotiation (meetings, objections, decisions) - Identify causal patterns between actions and outcomes - Combine with engagement data to understand how internal factors affect external performance

Example record:
- Source: Negotiation Analysis #2024-156 (Success)
- Observation: Client mentioned trust in technical team
- Interpretation: Technical competence demonstration influenced decision
- Identified relationship: Technical Competence → INFLUENCES → Trust → INFLUENCES → Closure

1.5 How it works in practice

1.5.1 1. You document the source

SOURCE @engagement_survey_2024
    access_date: 2024-12-15

All knowledge is traceable to its origin — a document, an interview, a report.

1.5.2 2. You record observations and interpretations

ITEM
    quote: "I don't know where the company is going"
    note: Employee expresses lack of clarity about strategic direction
    chain: Lack of Clarity → REDUCES → Engagement
END

Each item connects textual evidence to your interpretation and the identified causal relationship.

1.5.3 3. You define your conceptual vocabulary

ONTOLOGY Engagement
    topic: Organizational Behavior
    description: Level of employee commitment and emotional connection
END

Concepts are explicitly defined, ensuring consistency throughout the project.

1.5.4 4. The system validates and structures

The Synesis compiler: - Verifies annotation consistency - Generates structured formats (JSON, CSV, spreadsheets) - Enables integration with analysis systems

1.6 Benefits for the organization

1.6.1 Complete traceability

Every conclusion can be traced back to the original evidence. When someone asks “why do you recommend this?”, the answer is documented.

1.6.2 Preserved institutional knowledge

Expert knowledge, which normally exists only in their heads, gets recorded and available for the organization.

1.6.3 Qualitative-quantitative integration

Qualitatively documented causal relationships can be combined with statistical analyses for cross-validation.

1.6.4 Foundation for artificial intelligence

Organizations investing in Causal Knowledge Graphs need structured knowledge as input. Synesis provides this structure from documentary sources.

1.6.5 Grounded decision making

Instead of decisions based on “gut feeling” or superficial correlations, the organization operates based on documented and verifiable causal relationships.

1.7 Comparison with other approaches

Approach	Data source	Relationship type	Limitation
Traditional statistical analysis	Numerical	Correlation	Doesn’t distinguish cause from effect
Statistical causal discovery	Numerical	Causal (estimated)	Requires large volumes; may err on direction
Traditional qualitative research	Textual	Qualitative	Unstructured; hard to integrate
Synesis	Textual	Causal (documented)	Requires human analysis; complements statistics

The ideal combination: use Synesis to structure qualitative knowledge, statistical causal discovery for numerical data, and integrate both in a Causal Knowledge Graph.

1.8 Integrated use scenario

Imagine a company wanting to simultaneously improve: - Success rate in commercial negotiations - Employee engagement - Project delivery quality

Flow with Synesis:

Structured collection
- Engagement surveys documented in Synesis
- Negotiation analysis documented in Synesis
- Quality reports documented in Synesis
Connection identification
- Statistical discovery identifies: Engagement → Delivery Quality
- Synesis documents: Teamwork → Engagement (from literature)
- Synesis documents: Delivery Quality → Customer Trust (from negotiations)
Integrated graph
- Teamwork → Engagement → Quality → Trust → Negotiation Success
Grounded recommendation
- Investing in teamwork initiatives has documented cascade effect up to commercial results

1.9 Getting started with Synesis

1.9.1 For organizations

Identify a pilot domain — HR, operations, or commercial are good starting points
Define initial vocabulary — what concepts are relevant for your decisions?
Train analysts — the learning curve is low for those who already work with qualitative data
Integrate with existing processes — Synesis complements, doesn’t replace, current tools

1.9.2 For individual professionals

Experiment with a small project — a literature review, a set of interviews
Focus on causal relationships — always ask “what causes what?”
Maintain consistency — use the same terms for the same concepts
Document your interpretations — the annotation is as important as the citation

1.10 Frequently asked questions

1.10.1 Does Synesis replace BI or analytics tools?

No. Synesis works with qualitative knowledge and integrates with quantitative tools. It’s a complement, not a replacement.

1.10.2 Do I need technical knowledge to use it?

The syntax is simple and readable. Research, analysis, and management professionals can learn quickly. Programming is not required.

1.10.3 How does Synesis relate to artificial intelligence?

AI systems that work with causal reasoning need structured knowledge. Synesis provides this knowledge from sources that AI can’t process alone with the same quality.

1.10.4 What’s the difference between Synesis and simply using spreadsheets?

Spreadsheets have no consistency validation, don’t natively structure causal relationships, and don’t guarantee traceability. Synesis was specifically designed for causal knowledge capture.

1.11 Conclusion

In a world where decisions need to be increasingly data-grounded, qualitative knowledge can’t be left out. Synesis offers a bridge between expert knowledge — dispersed in documents, reports, and observations — and the analysis systems that support decision making.

The combination of structured causal knowledge with statistical analysis represents the next step in the evolution of organizational intelligence: decisions not only based on data, but based on causal understanding of phenomena.

Synesis: transforming dispersed knowledge into actionable intelligence.