This website uses cookies

Read our Privacy policy and Terms of use for more information.

“Every once in a while, a new technology, an old problem, and a big idea turn into an innovation.”

—Dean Kamen, Engineer & Inventor

The AI Breakdown

The Baggage Behind Every Bot

Every AI chatbot shows up to the conversation with baggage.

Before you type a prompt, tools like ChatGPT, Claude, Gemini, and Grok are following hidden system prompts. These tell the bot how to behave, what tone to use, what policies to follow, when to search the web, how to handle files, and which topics to avoid.

That helps explain why AI can feel inconsistent. The user thinks, "I asked a clear question." The bot is also weighing brand voice, safety rules, copyright limits, tool access, formatting preferences, and instructions from the company that built it. The answer is not just generated. It is routed through a quiet chain of command.

The Washington Post

A dealership would never drop a new employee into the showroom with only one customer question and no context. Yet that is how many stores use AI. They ask it to write a lease email, summarize a call, coach a BDC rep, or draft a service follow-up without giving it the store's rules of engagement.

Custom instructions are the closest thing users get to their own system prompt. They will not override the model's deepest rules, but they can shape the layer that matters most in daily work: tone, audience, process, and judgment.

The Washington Post

The practical move is to treat your AI setup like onboarding. Give it the role, the audience, the standards, and the boundaries. Then watch where it still creates extra work and tighten the instructions.

A better bot usually starts with a better manager.

Prompt of the Week

Now that you know there is a prompt behind the prompt, the next move is learning what you can actually control. This helps you audit how your AI behaves, then turn those findings into custom instructions that make it more useful for your daily work.

I want to better understand how you behave as an AI assistant so I can customize you for my work.

Explain the user-visible parts of how you operate:

1. What kinds of instructions from me are you most likely to follow well?
2. What kinds of requests are you likely to refuse, redirect, or handle cautiously?
3. How do you decide when to be brief versus detailed?
4. How do you handle missing information?
5. When should I ask you to reason, summarize, draft, critique, or format?
6. What formatting preferences can I give you that will improve your outputs?
7. What tone instructions are most effective?
8. What mistakes do users make when trying to customize you?

After that, help me write a set of custom instructions I can paste into my AI settings.

Base the custom instructions on this context:

My role is: [insert role]
My industry is: [insert industry]
My main tasks are: [insert tasks]
My audience is: [insert audience]
My preferred tone is: [insert tone]
Things I want you to avoid: [insert dealbreakers]

Make the final custom instructions specific, practical, and reusable.

Fresh Finds for Auto Pros

  • Finance & Insurance: Kognitos
    An agentic AI platform that lets you automate dealership paperwork and accounting just by typing out instructions in plain English. It reads messy documents like loan forms with perfect accuracy, so you never have to worry about the AI making math mistakes.

  • Marketing & Advertising: AutoFox
    This platform allows anyone with a smartphone to take a photo of a car in a muddy parking lot and instantly place it in a high-end, 3D-rendered virtual showroom. It even uses AI to adjust the "lighting" on the car’s body to match the new digital environment.

  • Content Creation: Bolta
    An AI social agent specifically for auto dealers. Instead of just writing a post, it syncs with your live inventory and automatically generates TikToks, Reels, and Facebook posts. It doesn't just show the car; it creates "Buyer Education" clips using your specific dealership's brand voice.

Hear from the Experts

In this conversation, David Spisak brings the focus back to what has always made great dealerships great: people, purpose, and the discipline to do the hard things well.

As the industry keeps moving through new technology, AI, regulation, and shifting customer expectations, David offers a grounded reminder that progress is not just about adopting what is new. It is about strengthening the culture, talent, and trust that make innovation actually work.

Watch the full interview for a thoughtful, energizing look at leadership, resilience, and why the future of automotive still has a very human heartbeat.

Bits and Bytes

  • Researchers found a 175% spike in ChatGPT using metaphors about goblins and gremlins. 👹

  • Elon Musk announced that Tesla’s AI Vision can now deploy airbags and seatbelt tighteners up to 70 milliseconds before a physical impact occurs. 🦺

  • “Hyperscalers" like Amazon, Meta, and Google are slashing stock buybacks to fund a projected $755B in AI capital expenditures this year. For many, AI spending is now eating 100% of their operational cash flow. 🤑

Parting Pixels

Thanks for reading, Friend! Stay caffeinated, stay curious, and please—for the love of your processor—close a tab or two.

Reply

Avatar

or to participate

Keep Reading