In partnership with 

🛠️ Welcome to Tool Kit Tuesday, friend. This is the one that earns its keep. No buzzwords, no 40-slide decks, just the AI shortcuts worth using. Did someone forward this to you? Sign up here, it's free.

HERE’S WHAT’S IN THIS ISSUE:

✏️ Turn a napkin doodle into a pro flowchart.
📓 This prompt to catches AI’s fake citations.
👓 A free app that reads pill bottles and menus.
📄 iPhone + ChatGPT = a 90-second contract scan.
🥊 Bladed robots sold out a $69 fight night in SF.
💬 “Context window” decoded + the human hack.

But first, trivia. Remember when “press zero for an operator” felt annoying? Now we get website chatbots with the emotional range of a parking meter and the helpfulness of a printer underwater.

That bot looping you in circles isn’t broken. It’s built to wear you down until you give up and close the window. (Rude.) The good news: there’s a back door. Type the right word, and the bot steps aside and hands you to a real, live human.

💬 So what’s the secret password to give a customer service chatbot? A) /please, B) /manager, C) /human, D) HELLO, YOU DUMB TOASTER! Don’t rage-click. The answer’s waiting at the end.

🕵️ Your data has a new buyer: AI companies. 32 data brokers say they sell personal info to firms building chatbots. I use Incogni to get my name, address and phone number off broker lists, and it keeps them off. Get 60% off with code KIM60.* — Kim

I’ve watched a bad whiteboard scribble win an argument it had no business winning. It wasn’t even the better idea. 

Pictures beat paragraphs, always have, and most of us lost that contest because we can’t draw. Nobody remembers what cavemen said, but we remember how they lived, thanks to their doodles. Proof that stick figures have always closed deals.

ChatGPT has crazy good sketch-to-image abilities on mobile, which means the envelope you scribbled on during a phone call is more than a doodle. It’s raw material.

The old rule was that whoever drew the diagram owned the idea, and drawing well was a skill you either had or faked badly. That rule’s gone. Speed owns the idea. The sketch is simply the input.

Take a picture and paste this:

Then read every single word inside the image. Diagrams feel authoritative, so nobody proofreads them. That’s how a typo ends up in a client proposal in 40-point font. The model will confidently invent a label, misspell a product name and turn your “basic user flow” to a “client journey paradigm.” 

A few caveats: Slop in, slop out. Let AI design your chart, but the info is on you. Unless you ask it to flesh out the idea. That’s another ballgame. Don’t feed it customer data or sensitive info. 

This week: Sketch out one thing you’ve explained out loud twice and nobody understood. Drop it into the next proposal or thread with one line: “as visualized here.” Watch how fast the conversation stops being about whether people catch your drift.

🩷 Love this idea? Post this in your group chat: Turned my napkin notes into a real diagram with one free AI prompt. Kim Komando shows you how. Sign up free at GetKim.com.

Your name, address, phone number and what you buy. It’s all for sale. Now AI companies are in the checkout line.

Stanford researchers dug into California’s 2026 data broker registry. They found 32 brokers say they sell personal data to companies building generative AI, the tech behind chatbots and image generators. Some brokers sell more than 10,000 types of data about you. (Ten thousand. I can’t name 10,000 things about myself.)

Here’s what you need to know:

Once your info trains an AI or lands in a scammer’s database, good luck getting it back. The smart move is getting it off the market before the next buyer shows up.

That’s why I use Incogni. (I’ve been a subscriber since 2023.) It sends removal requests to over 420 data brokers and people-search sites for you, then keeps checking so your info doesn’t sneak back.

Get your info off the market before the next buyer shows up.

✅ Get 60% off Incogni with code KIM60. Lock your data from AI bots now here → You need to take action now before the big AI companies have all your data.*

I used to think a citation was the safe and boring part of a document. Then I read what GPTZero found inside a KPMG report, and now I check footnotes like I’m examining a used car.

Imagine the meeting. You’re presenting, the numbers look great, the sources look airtight, then someone in the room clicks one of your links and says, “This page doesn’t say that.” Your credibility leaves the building and it doesn’t take the stairs as you sheepishly say, “Gee, that’s weird they’d change it right before my talk.” 

That is roughly what happened to a Big Four firm. KPMG published “Total Experience: Redefining Excellence in the Age of Agentic AI.” When GPTZero checked it, only 5 of the 45 citations matched their sources, and about half the factual claims were false, unsupported or pinned on the wrong source. 

KPMG quietly pulled the report from some of its sites. But it’s not just one firm. A federal court tossed an expert’s report over hallucinated citations (LeDoux v. Outliers, Inc.) in August. Once the “expert” was gone, the plaintiff’s case was dismissed with prejudice.

Paste this into your chatbot with search turned on:

Trust but verify is retired. New rule: AI citations are guilty until proven innocent. 

Do the human half. Open every link, hit Ctrl + F (Cmd + F on a Mac) and find that exact sentence yourself, because AI will hallucinate its own fact-check. 

Anything you can’t locate gets cut or labeled an estimate. Don’t paste confidential documents into a personal account. Use your company’s approved tool or only the public claims.

📩 Send this to someone who believes everything a chatbot says. You know the one.

Every issue of this free newsletter brings you one AI term you need to know to sound smart. In a year, you’ll be AI literate!

Context window (noun): The AI’s short-term memory. It’s measured in tokens, and one token is about three-quarters of a word.

Think of tokens like Lego bricks. AI doesn’t read whole words the way you do. It snaps text together from little pieces. A short word like “cat” is one brick. A long word like “unbelievable” might take three: “un,” “believ” and “able.”

Everything in your chat eats up space: your messages, the AI’s replies and every file you upload. When the window fills up, the oldest stuff gets tossed or squished into a summary. Details start slipping well before that. That’s why your chatbot “forgets” what you told it three days ago. (It’s not ignoring you. Its brain is full. Same, buddy. Same.)

The fix: Start a fresh chat and paste in a quick recap. Clean slate, sharp memory.

You’ll hear this term when people complain about long chats going off the rails, an AI ignoring a file they uploaded or contradicting itself mid-conversation. It also comes up when companies brag about “million-token” windows.

Use it in a sentence: “Sorry I forgot your birthday, honey. My context window was full.” Spoiler: That’s one token of affection you can’t afford to drop. 🎂

My friend Christine owns six pairs of readers. All six live in rooms she isn’t in. Last week, she FaceTimed me so I could read her a pill bottle. (Love you, Christine.)

I told her to download Seeing AI. Point your phone at a label, a menu or the electric bill, and it reads the words out loud. No squinting, no hunting for glasses.

Microsoft built it for people who are blind or have low vision, and it’s free on iPhone and Android. Christine now calls it her seventh pair of readers, and the only one she can find.

The catch: It reads labels. It doesn’t interpret them. Dosage questions still go to your pharmacist. Guess who I bumped into on my way to get my glasses fixed? Everybody.

📩 Send this to someone with readers in every room but the right one.

I watched a 12-page vendor contract on my desk get read in 90 seconds, and not by a human.

The ChatGPT iPhone app can scan a stack of paper pages and merge them into one PDF right inside the chat. No scanner app, no emailing yourself 12 blurry photos. Then, you can ask it where the auto-renewal clause is hiding.

Boss wants an answer by lunch? This is how you get one.

Setup takes about a minute, and my walk-through, the exact prompt and the privacy fine print are one tap away.

⏱️ Do the 60-second setup →

California regulators weren’t overjoyed when a robot fought a human. Naturally, the next idea involved weapons. No humans, thankfully. 

REK billed its Oct. 9 San Francisco Deathmatch as a first-of-its-kind humanoid weapon fight. The card featured robots with blades, a T800 matchup and Rektile versus T-REK. General admission cost $69.

It’s BattleBots mixed with the Terminator franchise. But humans remotely pilot REK’s fighters, so the mayhem comes with a flesh-and-blood operator.

Check out the promo here. We’ve invented gladiator combat for machines. Somewhere, a Roomba is updating its résumé. Popcorn not included. Tetanus shot recommended.

Share this now:

Here’s your haul for today. Turn a napkin doodle into a polished diagram, catch AI’s fake citations, learn why your chatbot forgets, get a free app that reads tiny print aloud, scan a contract in 90 seconds and watch robots battle with blades. Not bad for five minutes of reading.”

🍔 The answer: C) /human. On chat platforms that support it, typing /human sends you straight to a real person. No luck? Type “agent” or “representative.” Many bots are trained to hand you off when they see those words.

Now for the insider stuff. These work on chatbots and phone menus alike:

“I need to cancel my service.” Cancel is a magic word. It sets off alarms and often sends you right to the retention team, the people with the power to fix things (and hand out discounts).

“I’m returning your call.” The bot can’t find that issue in its script, so it passes you up the chain fast.

Want to go nuclear? Type “supervisor.” What’s customer service’s favorite word? Unfortunately...

🔒 Before you go: Strangers, scammers and AI companies can’t buy what isn’t for sale. Incogni pulls your personal info off data broker and people-search sites, then keeps checking. I’ve used it since 2023. Save 60% with code KIM60.

💡 The smartest people in the room aren’t the ones who know every answer. They’re the ones who check. Including the robots. Forward today’s newsletter to someone who doesn’t get AI. I’ll see you on Thursday with another issue of this great free newsletter. Appreciate you being here! — Kim

Kim Komando • Komando.com • 510+ radio stations • Trusted by millions daily

Photo credit(s): ChatGPT/Kim Komando, @REK via X

Companies and products denoted by an asterisk (*) within this publication are paid sponsors or advertisements. As an Amazon Associate, the publisher earns from qualifying purchases. Statements regarding products denoted by a double asterisk (**) have not been evaluated by the Food and Drug Administration; such products are not intended to diagnose, treat, cure, or prevent any disease. This newsletter is provided for informational and entertainment purposes only and does not constitute legal, financial, medical, or professional advice of any kind. Readers should consult with a qualified professional before making any decisions based on this content. The publisher disclaims all liability for any loss, damage, or injury resulting from the use of or reliance on the information contained herein.