Insights Panel

When you select an entity in Intelligence, the right-side panel displays contextual information plus recommendations. Some recommendations are AI-generated by the agent runtime; others are computed by deterministic heuristic rules. The panel labels them differently so you always know which is which.

Customer detail

Selecting a Customer shows:

Contact info: name, email, associated platforms
Timeline: ordered list of touchpoints and interactions
Current stage: position in the campaign funnel with progress percentage
Opportunities: any detected buying signals for this customer
View in Pipeline: button to jump to the corresponding CRM record

Campaign insights

Selecting a Campaign shows:

Reach: number of customers connected to this campaign
Funnel stages: Awareness, Consideration, Decision, Conversion with counts
Attribution: revenue and conversions attributed to this campaign
Recommendations: suggested optimizations (timing, audience, frequency)

Recommendation types

The panel surfaces two kinds of recommendations, visually distinguished:

Badge	Source	Example
`AI INSIGHT`	LLM-generated by an agent buddy. Grounded in the current graph + brand context.	”Your Series-B segment responds 34% better to Tuesday 10am sends. And I’ve noticed three just hired VPs of Marketing in the last week.”
Insights (heuristic flare, no AI badge)	Deterministic rules computed server-side from the graph. Timing, frequency, audience overlap.	”This campaign sent 4× to the same segment in 7 days. Consider reducing frequency.”

The heuristic path is always on; it’s free, fast, and runs on every campaign. The AI path runs on demand or when a buddy is asked for a deeper read.

Types of heuristic recommendations

Waypath’s heuristic engine generates these recommendation categories:

Optimize timing: adjust send times based on engagement patterns
Expand audience: identify similar segments that may respond well
Reduce frequency: detect over-communication that leads to drop-off
Optimize content: flag subject lines / CTAs with low click-through
Adjust audience: prune uninvolved segments from a campaign
Content strategy: longer-horizon content recommendations
Segmentation: flag overlapping segments that should be consolidated

All recommendation types are modeled as Recommendation nodes in the graph with RECOMMENDS_FOR and BASED_ON edges pointing to the relevant Campaign and supporting Touchpoints.