2 posts tagged with "building intelligence"

Once a year, I head to Colorado to cautiously ski down the easy part of the mountain. My gear of choice has been a big fluffy yellow jacket. Upon greeting the family for the holidays, my sister’s welcome was slightly colder than the wind chill:

“You look like the Michelin Man with jaundice.”

Ok then. New jacket it is.

Gemini Prompt:

Armed with a list of brands, I went to the store and met a 33 year sales veteran who calls himself Red:

“I’m looking for this [Insert Expensive Brand] jacket.”

“Why?” said Red.

“Ummm, I want something fashionable?”

Red said: “You don't really want fashion. You want comfort.”

Fifteen minutes later, I had a goose down parka that cost 38% less than I thought. How had I screwed up? Simple: The Prompt. I had over-indexed on "fashion" and neglected the "local expertise" constraint.

From the Slopes to the Runway

This brings us to our Scheduling conundrum of the day.

I recently sat down with Maestrow to analyze a carrier revamping its network in Q1 2026. The goal? Shedding unprofitable routes. My initial prompts focused heavily on frequency reductions. The results were great - graphics, analysis, and intelligence on where the carrier was pulling back.

But I made the "Yellow Jacket" mistake again. I was looking at the cuts, but I hadn’t asked the most important question: “What is happening to the planes?”

When I adjusted the prompts to look for insight into next steps - the "dry" data became a much different tale. The real story was a Fortress Hub expansion. The carrier wasn't just shrinking; they were reallocating to out-schedule on primary routes.

I almost missed this!

Your Intellectual Partner

Using an AI platform like Maestrow is a constant exercise in discovery. Our CIO, Chris, and I often "get into it" (sometimes in a collegial manner) regarding “Teaching the User” vs. “The User Experience.”

Chris argues that the system should be intuitive—and it is. Maestrow gets smarter daily by digesting our documentation libraries and existing prompt histories. However, I have to (grudgingly) admit that Chris is right: Prompt Context is everything. If you ask for a "fashionable" schedule, you’ll get one that looks good on a slide deck.

But if you treat the system as your intellectual partner and ask about the effects of asset reallocation, you get the "Red" version: the truth that saves you 38% and reveals the competitor's next move.

So, what have we learned?

1. Always ask what happens to the planes when the route is cut.
2. The right prompt turns "dry" data into actionable strategy.
3. My doctor has officially confirmed I do not have jaundice.

Yesterday, Maestrow and I were having our daily "strategic briefing" over morning coffee. Our automated news prompt, which monitors airline fleet announcements, delivered a fascinating piece of intelligence: a competitor we track is about to receive an unexpected addition of eight aircraft early next year.

This wasn't just interesting news; it was a strategic trigger. The key question immediately became: How do we find out if these new planes will be used to add new routes, or if they'll be used to shift capacity on existing ones? Manually sifting through thousands of flights at a hub airport to compare schedules would take a team of analysts hours, if not days.

This is where prompt engineering turns a tedious task into actionable market intelligence. We turned to an analytical prompt to do the heavy lifting.

Prompt:

For [Origin Airport Code], compare the current published schedule with 
the previous one. 
Identify and list any new or removed destinations over the next 10 months
Exclude code shares and include the carriers involved and the 
monthly change in flight count. 
Format your response in a clear table."

This single command contains several key elements of a powerful prompt:

1. A Specific Target: [Origin Airport Code] tells the AI exactly where to focus.
2. A Clear Task: compare the current published schedule with the previous one.
3. Precise Constraints: Exclude code shares is crucial for avoiding misleading data.
4. Structured Output: Include the carriers involved and the monthly change in flight count ensures the result is immediately useful for analysis.

Answer:

 There are 2 destinations added and 1 destination removed from your 
 target airport’s schedule. 
 Would you like a detailed breakdown of these changes?

A simple Yes, please from my end was all it took for Maestrow to present the full, detailed report.

This is the real power of modern AI. It’s not about replacing human insight; it's about amplifying it. In a matter of minutes, we had the data needed to understand a competitor’s moves and begin formulating a counter-strategy—all without spending hours on a spreadsheet.

This is just the beginning of what's possible with prompt engineering in airline analytics. In upcoming posts, I’ll share how we’re using similar techniques to compare public schedules against private, non-published ones and how we monitor for the subtle signs of network shifts.

For now, this intelligence is more than enough for a team to take action. And with that, I’m going to have another cup of coffee.