Synapse Insights — A New Analytics Framework for Observing Organizational Intelligence

Synapse Insights is publicly available on GitHub:https://github.com/masao-watanabe-ai/Synapse-Insights

Many conventional organizational analytics systems aggregate:

“What happened?”

For example, they examine organizational conversations and activities through metrics such as:

• Number of posts and messages
• Meeting participation and activity rates
• Reactions to messages
• Task-completion and progress rates
• KPI performance
• Rankings by member or team
• Turnover and retention rates

However, organizational conversations contain information that cannot be understood through activity volume alone.

Who consults whom?
Whose comments influence decisions?
Where do knowledge and decision-making inputs originate, and where do they flow?
Why does the same topic keep recurring without reaching a conclusion?
Are questions and coordination requests becoming concentrated on a particular person, creating an organizational bottleneck?

What matters is not simply counting the volume of conversations.
What matters is understanding the structure of relationships, knowledge, and decision-making through organizational interaction.

This makes it possible to observe:

• Who trusts whom and seeks their advice
• How knowledge flows and where it becomes stagnant
• Where friction or disconnection is emerging across departments and teams
• Whether dependence on specific key individuals is becoming excessive
• At which issues, relationships, or approval stages decision-making is slowing down
• Which teams appear active on the surface but are gradually losing their capacity to collaborate

In other words, what matters is:

The intelligence structure of the organization

This is why we developed
Synapse Insights.

What Is Synapse Insights?

Synapse Insights integrates:

Semantic Interaction
+
Temporal Graph
+
Organizational Intelligence

into an Organizational Intelligence Analytics tool.

It is not a system that merely aggregates and visualizes organizational interactions and activities.

It interprets the relationships that emerge through conversation, collaboration, decision-making, and knowledge sharing, and captures how the state of an organization changes over time.

Its essential purpose is:

To observe how organizational intelligence flows, from whom to whom, and how that flow evolves over time.

The Limits of Conventional Analytics

Conventional analytics primarily focuses on:

• activity
• engagement
• ranking
• volume

But in real organizations:

High activity does not necessarily mean a healthy organization.

For example, conditions such as:

• burnout
• hidden dependency
• silent experts
• communication breakdown
• knowledge bottlenecks
• unresolved friction

cannot be adequately identified through simple activity analysis alone.

Semantic Organizational Graph

In Synapse Insights, interactions are treated as semantic edges.

For example:

• reply
• mention
• question_to
• proposal_to
• reaction_insight
• reaction_disagree

are not treated merely as connections. They are treated as graph edges with meaning.

For example:

reaction_insight
→ knowledge propagation

reaction_agree
→ trust propagation

question_to
→ dependency relationship

reaction_disagree
→ friction signal

This transforms the graph from a simple network of connections into:

An organizational structure with meaning

Temporal Intelligence Graph

Equally important is the concept of the Temporal Intelligence Graph.

Synapse Insights does not only observe the current state of an organization. It also captures change over time:

t0 → t1 → t2

This makes it possible to observe patterns such as:

• changes in trust
• growth in influence
• burnout risk
• emerging leaders
• declining channels
• deteriorating collaboration

In essence, this is:

Monitoring the living dynamics of an organization

Emerging Leaders

Conventional organizational analysis often reveals only:

Who is influential now

But what may be even more important is:

Who is beginning to become influential

By observing changes over time in:

• influence delta
• trust growth
• contribution growth

Synapse Insights can identify emerging leaders.

Hidden Experts

Another important category is that of hidden experts.

In many organizations, the person who posts the most is not necessarily the person who possesses the most valuable knowledge.

For example, there may be people who:

• provide exceptionally accurate answers in a specific domain
• appear only when difficult discussions arise
• have high trust but low visible activity

Conventional analytics often fails to identify these individuals.

However, by using a Semantic Graph, Synapse Insights can observe knowledge propagation and reveal their role in the organization.

Burnout Risk

Burnout risk is also an important signal.

For example, a person may show:

• high activity
• high contribution
• declining trust or reactions from others

This may indicate a condition in which:

The person is continuing to work hard, while their connection with the organization is beginning to weaken.

This kind of signal cannot be detected through activity analytics alone.

Friction Detection

In organizations, friction is an important signal.

Examples include:

• disagreement
• repeated clarification
• unresolved discussion
• dependency conflict

Many analytics systems treat these simply as negative events.

In reality, however:

Friction is often an early signal of organizational change.

Synapse Insights treats these patterns as semantic friction.

Dependency Concentration

One major cause of organizational fragility is dependency concentration.

This is the condition in which:

An organization becomes overly dependent on a particular individual.

Examples include:

• All reviews being concentrated on one person
• Only a particular member being able to respond to incidents
• A small number of people holding critical tacit knowledge

This is closely related to the well-known bus factor problem.

Synapse Insights can observe dependency concentration through interaction patterns such as:

• question_to
• reply
• knowledge propagation

Why “Synapse”?

A synapse is the junction through which nerve cells connect and transmit signals.

What matters is not merely the individual node.

What matters is:

How signals propagate between nodes

Organizations work in the same way.

What matters is not individuals in isolation, but how knowledge, trust, and decisions flow through the organization.

Applications in Open-Source Communities

Synapse Insights is also highly compatible with open-source communities.

It can help observe patterns such as:

• maintainer burnout
• review bottlenecks
• hidden contributors
• issue propagation
• contributor onboarding difficulty

Open source is not merely code.

At its core, it is:

A network of knowledge and trust

Applications in Learning Communities

It can also be valuable in education.

For example, it can help identify:

• emerging learners
• mentors
• silent learners
• question dependency
• collaborative learning patterns

Rather than looking only at scores, it makes it possible to observe:

Learning interaction

Applications in Government and Public Communities

Synapse Insights can also be important for government, public-sector, and regional communities.

For example, it can help observe:

• participation inequality
• consensus formation
• trust deterioration
• fragmentation in policy discussions

This is close to understanding:

The structure of social intelligence

Connection with GNNs

Synapse Insights is naturally compatible with Graph Neural Networks (GNNs).

Its structure resembles a:

dynamic heterogeneous temporal graph

Nodes may include:

• User
• Team
• Channel
• Topic
• Knowledge
• Decision

Edges may include:

• trust
• dependency
• friction
• contribution
• proposal

And the graph evolves through time:

• t0
• t1
• t2

This can connect naturally to approaches such as:

• Temporal GNN
• TGAT
• DySAT
• TGN
• EvolveGCN

What It Can Enable in the Future

Over time, Synapse Insights can evolve toward capabilities such as:

• anomaly detection
• organizational forecasting
• collaboration optimization
• knowledge-bottleneck prediction
• team-health monitoring
• organizational simulation

For example, it may be possible to identify early signals such as:

“This team may be approaching a burnout cluster.”

“Dependency is becoming excessively concentrated.”

“A friction cluster is increasing rapidly.”

The Fundamental Shift

The important point is not to treat an organization as merely a collection of people.

What matters is:

Observing organizational intelligence

Synapse Insights integrates:

meaning
+
interaction
+
time
+
trust
+
knowledge
+
decision

This is not merely analytics.

It is:

Organizational Intelligence Analytics