Territory Advisor uses a conversational interface, so the way you ask questions affects the quality of the result. Good prompts and sensible settings help the system produce more useful analysis.
Prompt best practices
Ask one question at a time
Start broad, then narrow
Use plain language
Ask follow-up “why” questions when needed
Focus on one territory or one issue before expanding
Prompt examples that work well
| Goal | Prompt example | Why it helps |
Start analysis | Analyze my territories | Creates a broad health view before you drill in |
Find a problem | Which territories need the most attention? | Helps prioritize high-impact work |
Investigate a territory | Focus on territory North and tell me what the problems are | Narrows the review to one part of the model |
Generate action | Generate recommendations for territory North | Moves from diagnosis to decision support |
Preview change | Show me the before and after for this recommendation | Makes it easier to validate impact before acting |
Prompts to avoid
Very long multi-part prompts
Requests to plan territories from scratch
Requests for daily route optimization
Questions that depend on data from systems outside Repsly
Requests based on a previous session without rerunning analysis
Settings that most affect results
Territory Stability: how much operating buffer you want
Team Efficiency: how evenly work should be distributed
Rebalancing Aggressiveness: whether you want smaller or larger changes
Overlap Penalty: how strongly to discourage overlapping territories
Risk Penalty: how much to protect strong territories from disruption
Capacity and cadence settings: places per rep, working hours, and visit requirements
Filters: which territories and places are included in the analysis
If recommendations seem too small, too disruptive, or not aligned with reality, revisit your settings before dismissing the product output.
Simple rule of thumb
If the output feels off, check these three things first: your data, your filters, and your assumptions.