The best Skill usually starts with a workflow you already repeat. The examples below show how the same building blocks—inputs, data scope, ordered steps, output, and guardrails—can support everything from a quick weekly check to a multi-step management brief.
Start with the closest example, replace its scope with the way your team works, and test it before relying on the result. You rarely need to copy every detail from a larger example into a simpler Skill.
How to use these examples
Example 1: Weekly segment pick-up
Example 2: Three-week forecast
Example 3: GM weekly brief
Example 4: Corporate-account review with a file
Example 5: Unmapped value check
Testing and improving your version
How to use these examples
Each example includes a name, identifier, description, and copyable instructions—the same four fields you complete when creating a Skill.
Before saving your version:
Replace the example property, dates, thresholds, metrics, and output with your actual recurring workflow.
Remove any step you do not genuinely need. A focused Skill is usually faster and easier to maintain.
Keep important definitions and missing-data behaviour explicit.
Run the suggested tests and adjust the instructions where the result drifts.
Example 1: Weekly segment pick-up
Use this compact Skill when you want the same pick-up table and summary each week.
Name: Weekly segment pick-up
Identifier: /weekly-segment-pickup
Description: Reviews the last seven days of pick-up by segment for one property and highlights the strongest and weakest contributors.
Instructions:
PURPOSE Give the user a quick weekly view of which segments are driving or reducing near-term pick-up. INPUTS Ask once for the property if it is not already included. Use anything already supplied in the request. DATA SCOPE Use the Pick-Up Topic. - Update Date: last 7 complete days - Stay Date: today through the next 90 days - Group by: Segment - Metrics: Units Pick-up, ADR Pick-up, Accommodation Revenue Pick-up STEPS 1. Confirm the exact property match if the supplied name is ambiguous. 2. Run the analysis and sort segments by Accommodation Revenue Pick-up, highest first. 3. Identify the two strongest positive contributors and any segment with negative Units or Revenue Pick-up. OUTPUT - Start with one sentence summarising total Units and Accommodation Revenue Pick-up. - Show a table: Segment | Units Pick-up | ADR Pick-up | Accommodation Revenue Pick-up. - End with no more than three concise findings. GUARDRAILS Treat zero pick-up as a valid result. Do not add STLP, forecast, or budget comparisons unless the user asks for them.
Try these tests:
/weekly-segment-pickup for Flagship Austin by FlyrRun without a property and check that the Agent asks once for it.
Run for a quiet period and check that zero activity is reported plainly.
Example 2: Guided three-week forecast
This Skill demonstrates a reusable workflow that asks for a property, applies relative dates, and returns a consistent short outlook.
Name: Three-week forecast
Identifier: /three-week-forecast
Description: Summarises forecast Units, ADR, and Accommodation Revenue for the next three complete stay weeks for a property.
Instructions:
PURPOSE Provide a concise forward view for the next three complete stay weeks so the user can spot changes that deserve a closer look. INPUTS Ask once for the property if it is not already supplied. Accept an optional segment filter. Use any information already included when the Skill starts. DATA SCOPE Use the Forecasts & Budgets Topic. - Stay period: the next 3 complete stay weeks - Group by: Stay Week - Metrics: AI Forecast Units, AI Forecast ADR, AI Forecast Accommodation Revenue STEPS 1. Confirm the property match if necessary. 2. Run the forecast by Stay Week. 3. If the user supplied a segment, apply it consistently to every result. 4. Identify the highest- and lowest-forecast weeks for Units and Revenue. OUTPUT - Table: Stay Week | Forecast Units | Forecast ADR | Forecast Accommodation Revenue. - Three bullets: overall pattern, strongest week, and weakest week. GUARDRAILS Do not add budget, previous-period, or STLP comparisons unless requested. If AI Forecast is unavailable for the requested scope, say so and ask whether the user wants to review User Forecast instead; do not substitute it silently.
Try these tests:
/three-week-forecast for Flagship Berlin by Flyr/three-week-forecast for Flagship Berlin by Flyr, Corporate segment onlyUse a property name that could match more than one property and check that the Agent confirms the exact match.
Example 3: GM weekly brief
This is a more advanced Skill. It combines several analyses into one fixed management document, but each section still contributes to the same job: helping a GM understand the property's current commercial position.
Name: GM weekly brief
Identifier: /gm-weekly-brief
Description: Produces a concise weekly property brief covering current OTB position, pace, segment and channel movement, recent pick-up, and the forward outlook.
Instructions:
PURPOSE Create a consistent weekly commercial brief that a revenue manager can review with a GM. Keep the result focused on evidence and the areas that deserve attention; do not prescribe a pricing or operational decision. INPUTS Ask once for the property if it is not already supplied. Default the main review period to the current stay month and the forward view to the remainder of the calendar year. Use any different period already included by the user. DEFINITIONS - OTB = current confirmed Units, ADR, and Accommodation Revenue for the stay period. - STLP = the comparable OTB snapshot at the same point in last year's booking curve. Use STLP for pace comparisons on current and future periods. - Pace = whether current OTB is ahead of, level with, or behind STLP. STEPS 1. Current position: use the Bookings Topic to show OTB Units, ADR, and Accommodation Revenue for the main review period. 2. Pace: compare those metrics with STLP and state whether the property is ahead, level, or behind. 3. Segment movement: show the same OTB and STLP metrics by Segment and identify the largest positive and negative contributions. 4. Channel movement: compare direct and indirect channel performance using the same period and pace basis. 5. Recent activity: use the Pick-Up Topic to show Units and Accommodation Revenue Pick-up, including cancellations, for the last 7 complete days. 6. Forward view: summarise OTB and STLP by Stay Month from the next month through year end. 7. Synthesis: identify three supported bright spots and one evidence-based area to examine more closely. OUTPUT - Lead with a one-sentence bottom line on the property's current pace. - Use these sections: Current position; Segment and channel drivers; Recent pick-up; Forward view; Three bright spots; One area to examine. - Keep the full response scannable and under 800 words unless the user asks for more detail. GUARDRAILS Use the same property and date scope throughout. Do not replace STLP with final prior-year actuals. If a comparison is unavailable, state which part is missing and continue with the supported analysis. Do not invent a cause for a change or recommend a specific price. FINAL CHECK Make sure every conclusion is supported by a figure already shown and name any missing comparison that lowers confidence.
Try these tests:
/gm-weekly-brief for Serenity Resort Atlanta by FlyrRun it for a property without a complete STLP comparison and check that the limitation is stated.
Add “focus on the next two months” when running it and check that the saved structure remains intact while the forward scope narrows.
Example 4: Corporate-account review with a file
Use this pattern when Insights contains current production but the analysis also needs information your team maintains elsewhere. The Skill asks for the current file each time it runs; the file is context for that conversation and is not stored inside the Skill.
Name: Corporate-account review
Identifier: /corporate-account-review
Description: Compares a supplied corporate-account list with current production for one property and flags contribution, cancellation, and achieved-rate patterns to review.
Instructions:
PURPOSE Give the user a bird's-eye health check across the corporate accounts in their current reference file, using real Insights production alongside the terms supplied by the user. INPUTS Ask once for: 1. The property, if not already supplied. 2. The review period, defaulting to completed stay dates from the start of this calendar year through yesterday. 3. The current corporate-account file, if it is not attached. Read the file as supplied rather than assuming a fixed layout. Identify the account name and any available contracted rate, volume commitment, or notes. Ask one grouped clarification only when missing information prevents a valid comparison. DATA SCOPE Use the Bookings Topic for monthly production by Company Name: Units, ADR, Accommodation Revenue, Ancillary Revenue, Total Revenue, average length of stay, and average booking window. Use the Pick-Up Topic for booked and cancelled Units needed to calculate cancellation rate. STEPS 1. Match all file account names against Company Name without treating capitalisation alone as a difference. List any unresolved matches before analysing them. 2. Compare monthly Units and Total Revenue with the equivalent completed stay-date window last year where available. 3. Calculate cancellation rate from booked and cancelled Units. 4. Where the file includes a contracted rate, compare it with achieved ADR as a directional check and explain that booking mix may affect the comparison. 5. Where the file includes an annual volume commitment, show year-to-date attainment and label it clearly as progress against the full-year commitment, not a forecast of final attainment. 6. Classify each account as Doing well, Watch, or Needs attention and support the classification with figures. OUTPUT - Start with the number of accounts reviewed and unresolved. - Table: Account | Units | Total Revenue | YoY change | Cancellation rate | Commitment attainment | Achieved ADR vs contracted rate | Status. - Omit a column when the file or Insights data does not support it. - End with the strongest account, the account needing the closest review, and the most useful next deep-dive. GUARDRAILS Do not invent contract terms or merge distinct account names without confirmation. If the prior-year period or a file field is unavailable, label the comparison unavailable rather than estimating it. Do not recommend a specific contracted rate or concession.
Try these tests:
Attach a current test file and run
/corporate-account-review for Flagship Austin by Flyr.Run without the file and check that the Agent asks for it before querying.
Include one misspelled or unmatched account and check that it is surfaced instead of treated as zero production.
Example 5: Unmapped value check
This Skill shows how a repeatable data-quality check can identify values that may prevent clean reporting without attempting to change the underlying configuration.
Name: Unmapped value check
Identifier: /unmapped-value-check
Description: Finds blank or unmapped rate-code, channel, and segment values for a property and sizes their impact on recent production.
Instructions:
PURPOSE Identify production that is difficult to classify because a reporting value is blank or unmapped, and show where investigation would have the greatest impact. INPUTS Ask once for the property if it is not already supplied. Default to stay dates in the last 90 complete days and accept a different period from the user. DATA SCOPE Use the Bookings Topic. Review Rate Code, Channel Name, and Segment with Units and Accommodation Revenue for the selected Stay Date period. STEPS 1. Confirm the property and date scope. 2. Find records where Rate Code, Channel Name, or Segment is blank or uses an unmapped value. 3. Group the result by the affected field and underlying value where available. 4. Rank issues by Accommodation Revenue, highest first. 5. Calculate each issue's share of the property's total Units and Accommodation Revenue in the same period. OUTPUT - Lead with the total Units and Accommodation Revenue affected. - Table: Affected field | Displayed value | Units | Accommodation Revenue | Share of property revenue. - End with the three highest-impact values to investigate. GUARDRAILS A blank or unmapped value is a classification finding, not proof that the underlying booking is wrong. Do not invent the intended replacement or claim a root cause. If no affected records exist, say the checked fields are fully classified for the selected period.
Try these tests:
/unmapped-value-check for Flagship Berlin by FlyrRun for a period with no affected values and check that the Agent reports a clean result.
Add “last complete month only” and check that the runtime condition overrides the default 90-day period.
Testing and improving your version
For every example you adapt, test:
The usual request with complete inputs.
A request with one required input missing or ambiguous.
A period with no results or missing optional comparison data.
Review whether the Skill used the correct property, time fields, Topics, metrics, and comparison basis; whether it followed the requested output structure; and whether it disclosed missing information instead of filling the gap with an assumption.
When something fails, add the smallest clarification that explains the intended behaviour and why it matters, then rerun all three tests. Avoid copying every guardrail from these examples into every Skill—the right instructions are the ones that reliably support that Skill's one job.
For the creation steps, field limits, invocation methods, and current feature boundaries, see Skills in Insights.
Need help adapting one of these examples to your workflow? Reach out via the in-app chat — we're happy to help.