Quick Reference Card

Note: Clari and Salesloft completed a merger in early December 2025. VERIFY current product packaging and branding.

Paste this at the start of any AI prompt involving data, forecasts, or analysis:

“If unsure about any assumption, growth rate, or projection, write ‘VERIFY:’ next to it. Do not invent financial data or benchmarks. Flag all assumptions explicitly. This is for internal planning and is not financial advice. Past performance does not guarantee future results.”

Trend Analysis

“analyse this revenue data and identify: (1) trends over time, (2) top performing segments, (3) areas of concern, (4) recommended actions. Present findings in a table format with supporting calculations.”

Variance Analysis

Cohort Analysis

Pipeline Health Check

Competitive Pricing Review

Lead-to-Revenue Forecast

Sales Capacity Forecast

MRR Waterfall Forecast

Three-Scenario Forecast

Quarterly Forecast Update

“Update the annual forecast based on Q[X] results. Original forecast: [X]. YTD actual: [X]. Provide revised quarterly projections, variance analysis, and updated assumptions.”

Complaint Response Prompts

Guest Complaint Response

“Draft a response to this guest complaint using the HEARD framework. Acknowledge the specific issue, apologize sincerely, offer a concrete resolution, and close with follow-up commitment. Keep tone warm but professional.”

Online Review Response

“Draft a response to this negative review. Keep it brief (under 100 words). Acknowledge the issue, apologize, explain one action taken, and invite private follow-up. Do not be defensive.”

Internal Incident Summary

“Summarize this complaint for internal documentation. Include: guest name, date, issue summary, root cause (if known), resolution provided, recovery gesture, follow-up status, and recommendations to prevent recurrence.”

Meeting and Communication Prompts

Meeting Preparation

Executive Summary

Email Draft

“Draft a professional email to [recipient] about [topic]. Tone should be [formal/friendly/urgent]. Include: context, key message, specific ask, and next steps.”

Part 3: Forecasting

Forecasting Methods

Forecasting Formulas

Growth and Trend

Pipeline

Note: Velocity is a heuristic. VERIFY stage consistency and cycle definition.

SaaS / Recurring Revenue

Forecast Accuracy

If Actual values can be zero or near zero, consider Mean Absolute Error or scaled errors instead of MAPE.

Forecast Categories

Example thresholds (VERIFY your internal definitions):

Time Series Data Requirements

At least 12 monthly points helps estimate a baseline trend

Ideally at least two full seasonal cycles for seasonality

For monthly seasonality, this often means 24+ months

Data requirements are method-specific. VERIFY for your approach.

Stage Probability Guidance

Probabilities should be based on recent win rates

Use data that reflects current process, pricing, and ICP

VERIFY the lookback window that best reflects present conditions

Update probabilities quarterly with recent performance data

Part 4: Service Recovery

Service Recovery Frameworks

HEARD (Detailed Recovery)

LAST (Quick Recovery)

Other Common Frameworks

Most frameworks emphasize listening, apology, action, and closure.

The Service Recovery Paradox

Some studies find that strong recovery can raise satisfaction above no-failure cases, under specific conditions.

The paradox, when it appears, is more likely after:

A one-time failure (not a pattern)

An exceptional recovery (exceeds expectations)

Swift response

Genuine empathy and personalization

Caution: The paradox does not justify allowing failures. Prevention remains the priority.

Script Building Blocks

Empathy Phrases

“I completely understand why you feel that way.”

“That sounds incredibly frustrating.”

“I would feel the same way in your position.”

Ownership Phrases

“Let me take responsibility for getting this resolved.”

“My name is [name], and I will own this until it is fixed.”

“This is on us, and I am going to make it right.”

Clarity Phrases

“Here is exactly what I will do next.”

“You will hear from me by [specific time].”

“Let me confirm: you would like [summary]. Is that correct?”

Boundary Phrases

“While I am not able to offer [X], I can absolutely do [Y].”

“Our policy does limit [X], but here is what I am empowered to do.”

Appreciation Phrases

“Thank you for telling us directly. It gives us a chance to fix this.”

“I appreciate you giving us the opportunity to make this right.”

Do Not Say / Say Instead

Complaint Scripts by Situation

Room or Facility Issue

Opening: “Thank you for letting us know. I am sorry you are dealing with this.”

Resolve: “I can send engineering immediately, or move you to a different room. Which would work better?”

Recovery: “I would like to offer [amenity/late checkout] as an apology.”

Service Delay

Opening: “I am very sorry for the wait. Your time is valuable.”

Resolve: “Right now, I can have [service] with you in [time].”

Recovery: “Given the delay, I would like to [waive fee/provide complimentary item].”

Staff Attitude

Opening: “I am truly sorry you felt spoken to that way.”

Acknowledge: “Regardless of intent, the impact on you matters most.”

Action: “I will address this directly with the team member.”

Billing Dispute

Opening: “Thank you for bringing this to my attention.”

Clarify: “Here is how the charge appears and what it relates to.”

Resolve: “Here are the options: [adjust/remove/explain]. Which feels fair to you?”

Online Review Response Template

Dear [Name],

Thank you for sharing your experience. I am sorry that your stay fell short of your expectations and our standards.

You mentioned [specific issue]. We take full responsibility. Since your stay, we have [action taken].

I would welcome the chance to speak with you at [contact] and personally oversee your next visit.

[Your Name, Title]

Part 5: Empowerment and Escalation

Sample Empowerment Matrix

VERIFY and customise for your organisation.

Key principle: Poor reviews can reduce demand and pricing power. Small recovery gestures can be cost-effective compared with reputational damage.

Escalation Decision Tree

Safety or legal issue? → Escalate immediately

Media or public relations risk? → Escalate immediately

Beyond my empowerment level? → Escalate to next level

Guest requests manager? → Honor the request

Repeat complaint from same guest? → Escalate for review

Within my authority and solvable now? → Resolve and document

Response Time Targets

VERIFY targets for your operation.

Part 6: Benchmarks

VERIFY all benchmarks for your industry, stage, and context.

SaaS Metrics (General Guidance)

Pipeline Coverage (General Guidance)

Coverage needs depend on win rates and deal velocity.

MAPE Targets

Some teams target low double-digit MAPE for stable, mature revenue lines. VERIFY targets by forecast horizon, volatility, and business model.

Part 7: Checklists

Before Using AI Output

Numbers verified against source data

Calculations spot-checked manually

Claims checked for VERIFY tags

Tone appropriate for audience

No hallucinated facts or made-up sources

Compliant with company policies

Before Running a Forecast

Data is current and complete

Definitions are consistent

Outliers identified and explained

Source documented

Before Presenting a Forecast

Assumptions documented and labeled

VERIFY tags on unvalidated assumptions

Sensitivity analysis completed

Ranges provided, not just point estimates

Comparison to prior forecasts included

Risks and opportunities noted

Forecast Accuracy Improvement

Review forecast vs. actual monthly

Identify patterns in errors

Update conversion rates quarterly

Remove stale pipeline deals

Validate assumptions against recent data

Document reasons for significant misses

Service Recovery Completion

Guest felt heard (confirmed understanding)

Sincere apology delivered

Concrete action taken or scheduled

Recovery gesture offered (if appropriate)

Follow-up commitment made

Incident documented for team learning

Data Quality Check

Data updated within acceptable window

No gaps or missing periods

Metrics defined consistently

Outliers identified

Source documented

Part 8: Training Tips

AI Tool Adoption

Start with one tool for one use case

Build prompt templates for common tasks

Document what works in a shared playbook

Review outputs before sharing externally

Update prompts as you learn what works

Forecasting Discipline

Document every assumption explicitly

Assign an owner to each assumption

Set review dates for key assumptions

Update forecasts regularly (weekly for pipeline, monthly for financial)

Present ranges, not false precision

Track forecast changes to identify bias

Service Recovery Training

Role-play scenarios in pre-shift briefings

Build familiarity through repetition

Create a non-blame culture for empowerment

Document and share recovery wins

Clarify escalation paths clearly

Reinforce the “why” behind recovery gestures

Part 9: Common Mistakes

AI Usage Mistakes

Forecasting Mistakes

Service Recovery Mistakes

Part 10: Resources

Tool Websites

Pricing Pages

Further Reading

SaaS Capital Retention Benchmarks

Bessemer State of the Cloud

CFI Financial Modeling Guide

Gartner Peer Insights

G2 Grid Reports

Emergency Contacts Template

Fill in for your organisation:

Five Things to Remember Every Day

Listen first. Let people finish before responding.

Own it. Take responsibility, even when fault is unclear.

Be specific. Vague promises erode trust.

Follow up. Close every loop you open.

Verify. AI helps, but human judgment decides.

Disclaimer

This quick reference card summarizes guidance from AI Crash Course materials. All frameworks, formulas, prompts, and scripts should be customised for your organisation, verified against current policies, and adapted to local regulations. Items marked VERIFY require confirmation before use. Benchmarks are general guidance and vary by industry and context. AI-generated outputs are not substitutes for professional judgment. AI Crash Course is not responsible for decisions made using this reference.