Rules

Rules let you control how Jepto AI behaves in AI Chat. They're instructions you provide to guide the AI's tone, terminology, focus, and analysis approach when generating responses. Rules act as guardrails—ensuring AI Chat outputs align with your brand voice, client expectations, and reporting standards.

Rules apply only to AI Chat. They don't affect other Jepto features like automations, anomaly detection, or reporting.

Account Rules vs Client Rules

Account Rules

Apply when you're using AI Chat with "All Clients" selected.

Set Account Rules:

1. Select All Clients in the client picker (top of Jepto)
2. Go to Settings → Rules tab
3. Enter rules in the Custom Rules field
4. Click Update to save
5. When you open AI Chat on "All Clients", these rules will apply

Client Rules

Apply when you're using AI Chat with a specific client selected.

Set Client Rules:

1. Select the client in the client picker
2. Go to Settings → Rules tab
3. Enter rules in the Custom Rules field
4. Click Update to save
5. When you open AI Chat on that client, these rules will apply

How It Works

Rules are scoped to where you set them:

• Rules set on "All Clients" → Used in AI Chat when "All Clients" is selected
• Rules set on a specific client → Used in AI Chat when that client is selected

Each scope is independent. Client Rules don't layer on top of Account Rules—they're used in their own separate context.

‍

Use Cases & Examples

Brand & Tone Rules

For agencies wanting consistent client voice:

• Always refer to the client as "the brand", not their business name
• Use a professional, data-driven tone
• Avoid marketing jargon; explain metrics in plain language
• When making recommendations, focus on strategic impact

For luxury/premium brands:

• Emphasize quality metrics and customer value over volume metrics
• Use sophisticated, refined language
• Always include context about competitive positioning
• Frame recommendations around brand prestige and customer experience

For casual/lifestyle brands:

• Use conversational, friendly language
• Emphasize engagement and community metrics
• Include relevant trends and cultural context where applicable

Metrics & Focus Rules

For performance-focused agencies:

• Prioritize conversion and revenue metrics in all analysis
• When discussing performance, always include ROI or revenue impact
• Lead with what drove growth; secondary focus on what declined

For growth-stage clients:

• Emphasize growth rate and velocity over absolute numbers
• Focus on user acquisition and activation metrics
• Include month-over-month trends in all summaries
• Always suggest optimization opportunities

For budget-conscious clients:

• Always highlight cost-per-result (CPR) and efficiency metrics
• Include spend allocation and budget utilization in analysis
• Flag overspending or underutilized budget immediately
• When recommending changes, include expected ROI

Terminology & Naming Rules

For multi-location/franchise clients:

• Refer to individual locations as "locations" not "accounts"
• Always segment analysis by region when available
• Use "local search" instead of "GMB" or "Google Business Profile"
• Format location names as [City, State] for clarity

For e-commerce clients:

• Use "conversion" to mean completed purchases
• Use "sessions" not "visits"
• Refer to "product category" not "catalog section"
• Always discuss metrics in context of average order value (AOV)

For agency/SaaS clients:

• Refer to customers as "accounts" or "clients" (never "users")
• Use "churn rate" not "customer loss"
• Always measure success by MRR/ARR, not just user count

Compliance & Industry Rules

For regulated industries (finance, healthcare):

• Avoid predictive language; use "historical data" instead of "forecasts"
• Always cite data sources for any claims
• Never mention competitor names or comparative performance
• Include disclaimers on any forward-looking insights

For agencies with multiple industries:

• When analyzing healthcare clients, avoid pharmaceutical claims
• When analyzing finance clients, focus on compliance and regulatory impact
• For legal clients, emphasize confidentiality in all recommendations

Analysis Depth Rules

For executive summaries:

• Lead with the key insight or recommendation
• Keep explanations concise; one sentence per point
• Always include a "so what" or business impact statement

For detailed analysis:

• Always show the full context (month-over-month, year-over-year)
• Include the "why" behind performance changes when possible
• Suggest three potential optimizations minimum

‍

Testing Your Rules

How to Verify Rules Are Working

1. Ask AI Chat the same question twice—once with rules, once without (or in another account)
2. Check for your specific terminology—Do responses use the terminology you specified?
3. Review tone and emphasis—Does the response match your tone and metric priorities?
4. Test edge cases—Ask about performance, recommendations, and analysis to see how rules apply

What to Look For

✓ Responses use the terminology you specified✓ Tone and language match your guidelines✓ Metrics are prioritized as you indicated✓ Recommendations align with your focus areas✗ Rules are being ignored (may indicate they're too vague)✗ Responses are overly robotic or unnatural (rules may be too restrictive)

If Rules Aren't Working

• Rules are too vague. "Be professional" is harder to apply than "Use data-driven language; avoid marketing terminology"
• Rules contradict each other. Review Account and Client Rules together to ensure they align
• Rules are too prescriptive. If rules are extremely specific or numerous, the AI may struggle to apply all of them naturally
• Clear your browser cache. Sometimes cached data can interfere; try Ctrl+Shift+Del (or Cmd+Shift+Del on Mac) and hard refresh

‍

Performance & Limits

Token Usage

Rules consume tokens as part of the AI Chat system prompt. Generally, 5-10 well-written rules add minimal overhead (< 5% token increase). However:

• Very long rules or 20+ rules may increase token usage noticeably
• Longer responses will have proportionally higher token costs regardless of rules
• Token usage is counted toward your AI Credits

Character Limits

• Max per rule: No hard limit, but keep under 200 characters per rule for clarity
• Max total: Practically, 5-15 rules per account/client work best

Best Practices for Performance

• Keep rules concise and specific rather than lengthy explanations
• Prioritize 5-10 key rules over 20+ minor guidelines
• Combine related rules where possible ("Tone: professional, clear, data-driven" vs three separate rules)