Five minutes with Grok showed Walt three bets he didn’t know he was making

Today's Tip: Choosing the right heating oil contract before winter

I used to think choosing a heating oil plan was easy… you call your supplier, sign the contract, and stay warm. I was wrong.

Let me explain…

My neighbor Walt is a 68-year-old retired electrician in Waterbury, Connecticut. He's heated his 1,600-square-foot Cape Cod with oil for 32 years… and this year, the contract offer that landed in his mailbox stopped him cold.

The federal Winter Fuels Outlook, published October 6, projects heating oil will average $5.26 a gallon from November through March… up 34% from last winter. Oil-heated households are expected to spend an average of $2,115 this season. Connecticut's first weekly price survey came in at $6.015 a gallon. Walt burned about 650 gallons last year, and on a pension plus Social Security totaling around $3,400 a month, that kind of increase isn't an inconvenience… it's a crisis.

Walt opened grok.com and typed, "Is $5.89 a gallon a good deal for heating oil?" And Grok gave him the national average and told him "it depends." Sound familiar?

So Walt tried something different. He pasted his supplier's actual offer terms into Grok… the price per gallon, the $0.20 cap fee, the $150 cancellation penalty, his square footage, last year's usage, and his monthly budget. Everything except his account number.

Grok built him a side-by-side table comparing three plan types with his own numbers. A fixed-price plan locks a set rate for the season… predictable, but if prices fall, you're stuck paying the higher locked-in rate. A cap-price plan sets a ceiling… if market prices drop below the cap, you pay the lower price, but that protection costs an extra $0.10 to $0.30 per gallon. And budget billing spreads the total into equal monthly payments but does not protect the per-gallon price at all.

The federal forecast projects prices falling from $5.79 in November to $4.84 by March, so the cap plan would let Walt ride that decline while keeping a ceiling if prices spike. But the $0.20 cap fee added about $130 to his total season cost. Grok calculated the break-even point, flagged the $150 early-termination penalty, and generated eight specific questions for Walt to ask before signing anything.

Here's the prompt that did it…

Walt told me the comparison table alone was worth the five minutes. He could see, in his own numbers, that the cap plan would cost him about $130 more than a fixed plan if prices dropped as the forecast projects… but it would protect him if prices spiked above $5.89. That's a $130 insurance policy on a $3,400 winter fuel bill.

And when he called two more suppliers armed with those eight questions, one of them offered a lower cap fee. What are the odds he'd have known to ask for that on his own?

I know what you're thinking… "I'm not going to let a computer pick my heating plan."

And you shouldn't. The AI doesn't decide for you. It organizes the math and the fine print so you can see all three options side by side with your numbers. You still pick the plan and sign the contract.

But now you walk in knowing what to ask… instead of nodding along to whatever the sales rep says.

Pull out your supplier's contract offer tonight, open grok.com (or claude.ai or gemini.google.com), and paste the terms into the prompt above.

Five minutes now could save you hundreds before spring.

★ Quick Wins

Skill Builder: Comparison Shopping With AI

Most people use AI to compare options the same way they'd use a search engine… they type "best heating oil company" or "best mattress" and hope for something useful. But the AI doesn't know your budget, your location, your deal-breakers, or your deadline. So what do you get? A list of generic advice that helps nobody in particular.

The fix is what I call constraint loading. Every personal detail you add to a comparison prompt… your budget, your zip code, your must-haves, the things you will not accept, your deadline… eliminates options that don't fit and forces the AI to compare only what matters to you.

Here's the difference. A vague prompt like "Which is the best heating oil company?" gets you a list of national tips you could find anywhere. But a loaded prompt like "Compare these three heating oil plans for a 1,600-square-foot home in Connecticut that used 650 gallons last year, on a $3,400-a-month fixed income, Plan A costs X, Plan B costs Y, show me total cost, monthly cash-flow impact, and what I risk if prices drop 15%" gets you a personalized comparison table you can act on today.

Five constraints make almost any comparison useful: your budget, your location, your must-haves, your deal-breakers, and your timeline. Load all five into the prompt before you hit enter… and the AI has no choice but to give you a specific answer built around your situation.

Save this prompt somewhere you can find it… a note on your phone, a printout by the computer, whatever works. It handles heating oil plans, Medicare supplement quotes, new phones, contractor bids, or anything else where you're stuck choosing between two or three options and the fine print makes your head spin.

The more constraints you load in, the less generic the answer. That's the whole trick.