Skip to main content

Make the Most of the AI Query Helper in Insights

Best practices for using the AI Query Helper, helping you maximize the effectiveness of this feature and streamline your analytics workflow.

Ashley Dehertogh avatar
Written by Ashley Dehertogh
Updated today

Introduction

The AI Query Helper in Insights revolutionizes how you interact with your hospitality data, saving you valuable time while delivering deeper insights. This powerful tool transforms complex data exploration from a technical task into a natural conversation, allowing you to focus on strategic decisions rather than query building. By following these best practices, you'll maximize the effectiveness of this feature, dramatically accelerate your workflow, and uncover actionable insights that might otherwise remain hidden.

Getting Started with AI Query Helper

Start in the Right Topic

Before typing your first request, ensure you're building your query in the appropriate topic. This is essential for getting the results you want:

  • To compare current on-the-books (OTB) data to a previous period ➝ use the Bookings topic

  • To view what was OTB as of a prior date or analyze pick-up ➝ use the Pick-up topic

  • To evaluate forecast or budget data ➝ use the Forecast & Budget topic

  • To assess rate shopping data and competitor rates ➝ use the Rate Shopping topic

πŸ“– Need help choosing the right topic? Refer to our Insights Topics Overview for guidance.

Accessing the Feature

  • Location: Look for the "AI Query" button in your Insights workbook

  • Conversation interface: The assistant opens a chat-like interface for your queries

Understanding the Query Building Process

  • Real-time feedback: Watch how your query changes as you refine your requests

  • Field selection visibility: Notice which fields the AI selects to satisfy your request

  • Translating business questions: The AI converts natural language to technical queries behind the scenes

Formulating Effective Queries

Be Specific and Complete

Complete requests yield accurate results the first time, saving multiple iterations

  • Include all key elements in your initial request:

    • Metrics: "Show me ADR, Occupancy, and RevPAR..."

    • Time period: "...for the next 90 days..."

    • Comparison period: "...compared to same time last year..."

    • Breakdown dimensions: "...broken down by segment and channel..."

  • Example: "Show me ADR and Revenue by channel for April 2025 compared to April 2024, sorted by Revenue descending"

Structure Multi-Part Requests Clearly

Clear structure helps the AI parse complex requests correctly, delivering precisely what you need

  • Instead of: "I want to see occupancy, ADR, and revenue year-to-date with monthly breakdowns showing year-over-year changes for my downtown properties focusing on groups versus transient business"

  • Better approach: "Show me YTD performance with monthly breakdown, comparing to last year. Include Occupancy, ADR and Revenue. Filter to downtown properties only and segment by group versus transient business."

Refinement Techniques

Iterative Query Building

This approach lets you build complex analyses incrementally while maintaining control

  • Begin with a basic question and refine with follow-up requests:

    • Start: "Show ADR and Revenue by month for this year"

    • Refine: "Now add Occupancy to the query"

    • Further refine: "Filter to only show weekends"

    • Final touches: "Sort by Revenue descending"

Describing Modifications Clearly

Clear modification requests ensure precise adjustments without rebuilding entire queries

  • Be specific about changes: "Change the date range to last 90 days"

  • Reference existing elements: "Keep the same metrics but change the breakdown to segment instead of channel"

  • Add calculations: "Add a column showing year-over-year growth percentage"

Navigating Mistakes and Making Corrections

Undoing Changes

  • Reset to previous state: "Go back to the previous query" or "Undo the last change"

  • Start fresh: "Start over with a new query" or "Reset this query"

  • Remove specific elements: "Remove the channel filter" or "Remove the Revenue column"

Correcting Misinterpreted Requests

  • Be explicit about corrections: "You misunderstood - I meant Property X, not Property Y"

  • Clarify ambiguities: "I meant revenue per available room, not revenue"

  • Provide context for terms: "By 'direct' I meant direct booking channels, not direct costs"

Version Management

  • Save intermediate versions: Before making substantial changes, save your current query

  • Use descriptive titles: Name saved queries based on their purpose for easier reference

  • Keep track of manual edits: After AI generates a query, note any manual adjustments you make

Understanding Results and Insights

Interpreting Generated Data

  • Check your units: Pay attention to the actual measurements (e.g., currency, percentages)

  • Verify time frames: Confirm that the date ranges match your intended analysis period

  • Understanding variance: When comparing periods, note whether differences are absolute or percentage-based

Getting Explanations

  • Ask for context: "Explain why ADR is higher this month compared to last year"

  • Request analysis: "What insights can you provide about this occupancy trend?"

  • Clarify metrics: "What exactly does RevPAR STLP represent in this context?"

Transitioning Between AI and Manual Analysis

  • Export results: Take AI-generated queries to Excel or other tools for further analysis

  • Save as templates: Use successful AI queries as starting points for similar analyses

  • Manual fine-tuning: Make precise adjustments to AI-generated queries when needed

Understanding Limitations and Working Within Them

Data Availability Constraints

  • Cross-Topic Data: The AI can only access data within the current topic (e.g., Bookings). For analysis combining Forecast & Budget, STR Benchmarking, or Pick-Up data, you'll need to:

Complexity Boundaries

  • Too Many Metrics: Queries with excessive dimensions and metrics may become unwieldy

    • Solution: Break into multiple focused queries

  • Highly Complex Calculations: Some advanced custom calculations may require manual formula creation

    • Solution: Start with AI-generated query, then add calculations manually

  • Performance Impact: Very large datasets or highly complex queries may take longer to process

Special Cases

  • Custom Time Periods: Very specific date comparisons (e.g., "Compare the third Tuesday of each month") may require manual refinement

  • Complex Filtering Logic: Multiple nested conditions can be challenging for AI to interpret correctly

    • Solution: Build filters incrementally, confirming each step

Technical Constraints

  • Field Name Precision: The AI may occasionally confuse similarly-named fields

    • Example: Distinguishing between "Lead Time (Stay Date)" and "Lead Time (Arrival Date)"

    • Solution: Be explicit about which version you want

  • Advanced Visualizations: Some specialized visualizations may require manual configuration

    • Solution: Request the data structure you need, then customize the visualization manually

Industry-Specific Terminology and Concepts

Hospitality Metrics Understanding

  • The AI understands common hospitality terms like ADR, RevPAR, STLP, and MTD

  • Use standard industry terminology to get the most accurate results

  • For specialized metrics, provide a brief definition if needed

Handling Property-Specific Terms

  • Property codes or custom segment names may need additional explanation

  • Custom field groupings or user-defined labels should be explained clearly

  • Consider using the exact names from your property setup for best results

Advanced Usage

Comparative Analysis

  • Period-over-period analysis: "Compare Q2 performance to same period last year and two years ago"

  • Custom date ranges: "Show me June 1-15, 2025 versus June 1-15, 2024"

  • Multiple comparisons: "Show this month's ADR versus STLP and Previous"

Filtering & Segmentation

  • Multiple filters: "Show only high-ADR segments with occupancy below 70% for next month"

  • Nested criteria: "Find dates where weekday occupancy is above 85% but ADR is below $150"

  • Exclusions: "Show all channels except OTAs for this quarter"

Visualization Guidance

  • Request specific chart types: "Show this in a bar chart with months on the x-axis"

  • Ask for pivoted data: "Pivot the query to show months across columns and segments down rows"

  • Request totals: "Add column totals to the query"

Common Pitfalls to Avoid

Overcomplicating Requests

Simpler requests deliver more accurate results and are easier to refine

  • Avoid: Extremely long, multi-part questions with many conditions

  • Better: Break down complex analyses into sequential, focused queries

Ambiguous Terminology

Precise terminology ensures the AI retrieves exactly what you need

  • Avoid: Vague terms like "performance," "results," or "data"

  • Better: Specify exact metrics like "ADR," "Revenue," "Occupancy," or "Units"

Assuming Context

Clear context prevents misinterpretation and rework

  • Avoid: References without context ("Show me the same for last month")

  • Better: Provide sufficient context in each request

Summary

The AI Query Helper transforms data analysis from a technical exercise into a strategic advantage for your property or portfolio. By following these best practices and understanding its capabilities and limitations, you'll unlock deeper insights in a fraction of the time traditionally required. The result? More informed decisions, optimized revenue strategies, and a competitive edge in your market. As you continue to interact with the AI, it learns your specific needs and terminology, becoming an increasingly valuable partner in your revenue management process.


πŸ“Œ Need more information? Reach out to our Advisory team via the chat with any feedback on particularly good examples you've used and/or you need any support!

Did this answer your question?