Loretta’s neighbor said Medicare would take most of it back

Today's Tip: Calculating your net Social Security raise for 2027

On October 14, Social Security announced a 3.6% raise for 2027… roughly $75 more a month for the average retiree. So why are most people going to see a different number hit their bank account?

Let me explain…

A friend of mine named Loretta… a 67-year-old retired dental hygienist in Dayton, Ohio… saw that headline and did what most people do. She started planning what to do with an extra $75 a month.

Then her neighbor mentioned that Medicare was going to "take most of it back." Loretta remembered what happened last year… the Part B premium jumped $17.90, and that wiped out most of the 2.8% raise she got. Her net increase in 2026 was barely noticeable.

She opened gemini.google.com on her laptop and typed "How much will my Social Security go up next year?" Gemini gave her the same answer she'd already seen on the news… 3.6%, about $75 a month for the average person. But Loretta doesn't receive the average benefit. She gets $1,840. And she had no idea how to figure out her own number.

So she tried again… this time with her own details. She typed in her monthly benefit, her current Part B premium of $202.90, and the $35 dental supplement that also comes out of her check each month. She asked Gemini to calculate her gross raise, subtract the projected Part B increase, and show her what her new deposit would look like in January.

In about 20 seconds, Gemini laid it out line by line. Her gross raise was $66.24 per month. The projected Part B increase was $6.60. Her dental supplement stays the same. Net gain… $59.64 a month, or roughly $716 over the year.

Loretta was keeping most of her raise this time. Gemini also explained the hold-harmless rule… a federal protection that prevents a Part B increase from reducing your net check below last year's level. And it reminded her to set up a my Social Security account at ssa.gov/myaccount by November 18 to see her official notice online before the mailed letter arrives in December.

The difference between Loretta's first prompt and her second was about 30 extra seconds of typing. Here's the version that did the work…

Try out Musk's new AI agent – before his big announcement

Elon just created a device he believes will be "the biggest product ever."

He thinks it could 70x investors' money.

And he's about to make a major announcement…

By the end of this month.

Maybe even tomorrow on X.

He's going to make this game-changing device available to the public for the first time.

He has to sell 1 million to become a trillionaire.

Would you bet against him?

Click here to find out how you can claim a stake in this 70x AI agent.

      ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★
    

★ Copy-Paste Prompt

Works in Claude, Gemini & Grok

I receive $[YOUR CURRENT MONTHLY SOCIAL SECURITY BENEFIT] per month in Social Security retirement benefits. The 2027 COLA is 3.6%. The following amounts are currently deducted from my check each month:

- Medicare Part B premium: $[YOUR CURRENT PART B PREMIUM, or type "standard $202.90"]

- Other deductions (dental or vision supplement, tax withholding, etc.): $[LIST EACH DEDUCTION AND ITS AMOUNT, e.g., "dental supplement $35," or type "none"]

Please do the following:

1. Calculate my gross monthly raise at 3.6%.

2. Estimate my 2027 Part B premium using the projected increase of $6.60 per month.

3. Subtract all known deductions from my new gross benefit to show my estimated net monthly deposit for January 2027.

4. Show me the difference between my current net deposit and my estimated 2027 net deposit, in dollars per month and per year.

5. Explain in plain English what the "hold-harmless rule" means for me.

6. Tell me one thing I should do before December to prepare for this change.

Do not ask me for my Social Security number, date of birth, or Medicare number. Round dollar amounts to the nearest cent.

Why this works: It gives the AI all the numbers it needs without any sensitive data, asks for a specific output in six numbered steps, and builds in the context… the 3.6% COLA, the projected $6.60 Part B increase… so you don't have to look anything up. The numbered steps prevent a wall of text and make each answer easy to check.

One thing to keep in mind… the projected Part B premium of $209.50 has not been officially confirmed yet. Medicare usually announces the final number in November. The Trustees' projection is the best public estimate available right now.

So when Medicare announces the official premium in November, what do you do? Paste the same prompt again with the updated number. Takes less than a minute… and you'll have your final answer weeks before the mailed notice hits your mailbox in December.

★ Today's "Aha!" Moment

The 3.6% COLA is what the government announced. Your raise is what's left after Part B, supplements, and withholding come out… and that number is different for every single person.

I know what you're thinking… "The official Part B premium hasn't even been announced yet. Why should I bother calculating now if the number might change?"

Fair question. The Trustees' projected $209.50 is the best public estimate available, and it has historically been close to the final number. Running the math now gives you a planning figure six weeks before the official notice arrives. And if the actual premium shifts by a dollar or two when Medicare confirms it in November… you paste the prompt again with the new number. Takes less than a minute.

★ One Thing to Watch

AI is doing arithmetic on the numbers you type in, plus publicly available projections. It cannot see your actual Social Security account. If your income is above roughly $112,000 single or $224,000 as a couple, you may owe income-based Medicare surcharges… which means your Part B premium is higher than the standard amount, and the AI will get the math wrong unless you enter your actual premium. Always compare the AI's estimate to your official COLA notice when it arrives in December.

Open gemini.google.com, sign in with your Google account, and paste the prompt with your own numbers. Do it before the official notice arrives in December… you'll already know what to expect.

The headline is 3.6%. Your number is the one that counts.

★ ★ ★

Tool Spotlight ★

Gemini

Site: gemini.google.com

Made by: Google

Cost: Free tier available

Gemini handles today's COLA math with ease… type in your numbers, ask for a breakdown, and the answer comes back in seconds. If you already use Gmail or Google Docs, Gemini fits right into the tools you know, and the free tier at gemini.google.com is plenty for everyday questions like this one.

Google recently started rolling out a feature called Skills… reusable custom instructions you can save and call up by name in any Gemini chat. Skills began appearing in the Gemini app on October 13 and will replace the older Gems feature for personal accounts starting November 17.

Best for: Google users, quick math, plain-English explanations, working inside Gmail, Docs and Drive on paid plans

★ ★ ★

★ Quick Wins

The COLA Notice Signup Reminder — Open Gemini (gemini.google.com) or Claude (claude.ai) and type: My Social Security COLA notice for 2027 will be available online in late November. Walk me through the steps to create a my Social Security account at ssa.gov/myaccount so I can see it before the mailed letter arrives in December. The AI gives you a plain-English walkthrough of the signup process… it takes about five minutes. The deadline to create your account for online notices is November 18, 2026. Works in Gemini, Claude, and Grok.

The Senior Discount Finder — Open Grok (grok.com) and type: I am [AGE] years old and live in [CITY, STATE]. What senior discounts am I eligible for at grocery stores, restaurants, and pharmacies near me? List the discount, the minimum age, and whether I need to ask for it or show ID. Grok searches the live web to find current discounts in your area. Most stores don't advertise these… you have to know to ask. Works best in Grok because it pulls live information.

The Holiday Budget Planner — Open Gemini (gemini.google.com) or Claude (claude.ai) and type: I want to spend no more than $[TOTAL BUDGET] on holiday gifts this year for [NUMBER] people. Help me divide the budget fairly, suggest a spending cap per person, and list three gift ideas under each cap. The AI drafts a full budget and gift list you can adjust in under a minute. It beats staring at a blank page come November. Works in Claude, Gemini, and Grok.

Skill Builder: Writing Specific Prompts

Most people type a question into an AI tool the same way they'd type it into a search engine… a few words, hope for the best. "What's the 2027 COLA?" is a search-engine question. And it gets a search-engine answer… generic, one-size-fits-all, the same thing everyone else sees.

The fix is to fill in your own details before you hit enter. Your benefit amount, your deductions, your zip code, your age… these are the details that turn a generic response into a personal one. And the more of them you provide, the less the AI has to guess.

Here's the difference. A vague prompt looks like this… "How much will my Social Security go up next year?" A specific one looks like this… "I receive $1,840 per month in Social Security. The 2027 COLA is 3.6%. My Part B premium is $202.90. Calculate my gross raise, subtract the projected Part B increase of $6.60, and show me my new net deposit." Same topic. Completely different answer.

The AI doesn't know your situation unless you tell it. Think of every prompt as a form with blank fields… the more you fill in, the better the answer. Here's a prompt you can save and reuse any time you need a personal answer from any AI tool…

★ THE FILL-IN-THE-BLANKS PROMPT

Works in Claude, Gemini & Grok

Before I ask you a question, I am going to give you five facts about my situation. Use only these facts to answer. If I leave any blank, ask me for it before answering.

My topic: [WHAT YOU NEED HELP WITH, e.g., "understanding my Social Security raise"]

My key number: [THE MAIN DOLLAR AMOUNT OR FIGURE, e.g., "$1,840 per month"]

My constraint: [WHAT LIMITS APPLY, e.g., "I pay the standard Medicare Part B premium"]

My goal: [WHAT I WANT TO WALK AWAY WITH, e.g., "my estimated new monthly deposit in January 2027"]

My format preference: [HOW I WANT THE ANSWER, e.g., "a short numbered list with dollar amounts"]

Now answer my question using only these facts.

Why this works: It forces you to think about what the AI needs before you hit enter. The five numbered slots replace the blank-page problem… "What do I even type?"… with a fill-in-the-blank form. And that last line stops the AI from guessing when you leave something out.

Once you use this a few times, you'll start filling in your own details without even thinking about it. That habit… knowing your numbers, your constraints, and your goals before you type… is the difference between generic answers and personal ones.

Save this prompt somewhere you'll find it. It works for medical bills, insurance questions, tax prep, trip planning… anything where the answer depends on your specific situation.

Loretta’s neighbor said Medicare would take most of it back

Try out Musk's new AI agent – before his big announcement

Read More from the AI Advantage