Predictive Insights

Predictive Insights in MustardHub are interpretive signals designed to help Organizations notice emerging patterns that are difficult to see through day-to-day activity alone.

They surface directional indicators related to workforce dynamics such as:

Potential churn risk
The estimated cost of churn or retention
Changes in sentiment or engagement patterns
Signals that may suggest readiness for more responsibility

[Screenshot: Predictive insights overview screen]

These insights are not guaranteed forecasts or determinations. They do not represent guarantees, instructions, or automated decisions. Their purpose is to provide context, not conclusions.

Predictive Insights work best when viewed alongside what you already observe in the Workspace, such as participation patterns, recognition activity, engagement rhythms, and organizational structure. For that broader observational context, see C.1: Culture Studio.

What questions Predictive Insights help answer

Insights provide directional awareness, not resolution. They are designed to support reflective questions, such as:

Are there emerging patterns we might otherwise miss?
Are certain risks becoming more likely over time?
Where might the cost of inaction be higher than expected?
Are engagement or sentiment trends shifting across teams or roles?

Predictive Insights should be treated as signals worth noticing. They are not designed to replace judgment, context, or leadership discretion.

They are most useful when:

Considered alongside qualitative understanding
Viewed as part of a broader pattern rather than a single data point
Used to inform reflection, conversation, or further observation

What types of insights exist today

Today, Predictive Insights may include high-level signals related to areas such as:

Churn risk: indications that retention risk may be increasing
Cost-based insights: estimates related to the financial impact of churn or retention
Sentiment and engagement trends: directional shifts over time
Growth or responsibility signals: patterns that may suggest readiness for expanded responsibility

[Screenshot: Predictive insights churn risk detail screen]

These insights are presented at a summarized level and are designed to be interpreted in context, not acted upon in isolation. In some cases, insights can be explored at a more granular level to better understand the contributing signals behind them.

For high-level system strength indicators, see C.2: Workspace Strength Score.

For formal reporting and exports, see D.1: Reporting.

What a confidence score represents

Each Predictive Insight includes a confidence score. At a high level, this score represents the system’s estimated level of certainty in the signal being presented.

Confidence is influenced by factors such as:

The amount of historical context available
The consistency and richness of observed patterns
The breadth of signals available within the Workspace

Outputs are only as strong as the inputs available. Connecting external systems such as HRIS, payroll, scheduling, or communication tools can provide additional inputs and organizational context. Over time, these integrations help the system better understand how your Organization operates, which can contribute to clearer signals and more stable confidence scoring.

As more context is accumulated, confidence may increase.

A high confidence score does not mean an outcome is guaranteed. A low confidence score does not mean an insight should be ignored. Confidence is intended to help calibrate interpretation and not to validate or dismiss conclusions.

When insights become more useful over time

On day one, MustardHub has limited context. As the Workspace matures, patterns become clearer.

Over time, the system:

Learns how your Organization participates, communicates, and recognizes effort
Builds historical engagement and sentiment context
Improves its ability to interpret signals within your specific organizational environment

As organizational behavior evolves, so does the system’s understanding of that behavior. This combination of accumulated context and organizational learning can make insights clearer and confidence more stable over time, though never absolute.

Actions taken by the Organization may also influence both the underlying patterns and the insights themselves. That feedback loop is expected and healthy.

Predictive Insights

What questions Predictive Insights help answer

What types of insights exist today

What a confidence score represents

When insights become more useful over time

Recommended articles