ChatGPT prompts for insurance agents

[Markthebroker]

Pretty sure Alex is the offending party, not you

Oh, OK.

Man, I like that guy. When the hell is he going to take over as California insurance commissioner and fix this mess that we're in? I'm getting tired of telling people that their premiums are doubling and tripling. I mean my commission statements sure as hell aren't, but I am.

 

[Al3x Lee]

Ah, Mark, there it is, your golden sarcasm. Always good for a laugh. You didn't hurt anyone's feelings; if anything, you're the only one here making sense. Hell, California insurance commissioner? I'd fix that dumpster fire in a heartbeat, right after I take a shot at the guys running this circus.

But hey, Chris, Mr. "Administrator," thanks for that inspiring speech about asking questions. Really warmed my heart. Like, sure, technology's great, but it's not gonna solve the fact that half the agents out here don't know the difference between a deductible and a doorstop. How about we focus less on holding hands and more on toughening up, huh?

Sorry, that was a little mean, let me have ChatGPT fix it for me, ChatGPT, please rewrite this to be professional, cordial and comprehendible by a 5th grader:

ChatGPT: "Sure, Here's a more professional and cordial version of your message!":

---

Mark, your sense of humor always brings a smile. You certainly know how to keep things light. And about the challenges the California insurance commissioner faces, I'm confident improvements could be made with the right focus.

Chris, thank you for your encouraging message about asking questions. While technology is important, I think we also need to ensure agents are well-prepared and understand key concepts, like the difference between a deductible and other terms. It's important we strike a balance between supporting each other and ensuring we stay sharp and knowledgeable. (Vote for me for California Insurance Commission, remember, a vote for me is a vote for the LGBTQIA2S+NSPCGSNBFA+HMTMSSPOMBPQBLTMTDPIA+
Lesbian,
Gay,
Bisexual,
Transgender,
Queer,
Intersex,
Asexual,
2Spirit,
+Nonbinary,
Sapiosexual,
Pansexual,
Cisgender Ally,
Genderfluid,
Same-Gender Loving,
Neutrois,
Bear,
Femme,
Aromantic,
+Heteroflexible,
Maverique,
Two-spirit,
Muxe,
Skoliosexual,
Soft Butch,
Polyamorous,
Omnigender,
Misexual,
Bigender,
Polygender,
Quasi-gender,
Bicurious,
Lithromantic,
Trigender,
Metagender,
Two-Gender,
Demiboy,
Polyromantic,
Investigative Ally,
Autosexual

Community!)

 

[insuredbyijm]

This article doesn't include any prompts but may have some ideas listed you haven't thought of.

10 Creative Ways Insurance Brokers Can Use ChatGPT

 

[Gulliver]

Here are some ways to take it to the next level. First of all decide what you want from it, such as a marketing plan.

Start the prompt by telling the model(Chat GPT, Claude, Grok, etc) what it is.

EXAMPLE: You are an elite insurance consultant that specializes in developing marketing plans and growing sales. Please create an advanced and modern marketing plan for an independent insurance agent who specializes in mortgage protection insurance.

When you get the output from this prompt, review it and take what you like from it. Then in a follow up prompt do the following.

Follow up: These were great ideas, however, I would like a Level 2 response with more detail and potential behind it.

You should see an improvement and will probably get something good form the second prompt.

Follow up #2 You have done a great job with these first two ideas. Please go to level 3 and create a nuclear response that seeks total domination in this segment.

This style of prompting will amaze you.

Try these with your own goals and let me know how they worked. I am not a consultant. I'm just an intermediate level prompter, but most users have none of these recommendations in their quiver.

 

[annon123]

Depending on what you are using it for AI can have issues. When used for visit notes by doctors (research study) only a bit over half were happy with what AI wrote and saw inaccuracies that needed fixed. Colleges are dealing with the use of AI to write papers and assignments by students. The problem it it creates citations that don't exist and has factual errors (and some correct stuff - but you have to know enough to tell the difference). Some faculty have students create an AI paper and then their own paper is to correct and critique it.

Certainly using it to create a rough draft and then going in, fact checking, and cleaning it up, etc, might save some time and give you ideas you may not have thought of on your own. Of course AI is getting better over time but you still have the problem that developers use what is out there to train AI and if incorrect information is included then there is a problem. I worry the Captchas we get to deal with (where you have to check a motorcycle when it says bicycle, and 18 wheeler for a bus, some palm trees for a mountain - idiocies I have run into) are being used to train AL to then be used with self driving cars. No. Just no.

 

[Gulliver]

Depending on what you are using it for AI can have issues. When used for visit notes by doctors (research study) only a bit over half were happy with what AI wrote and saw inaccuracies that needed fixed. Colleges are dealing with the use of AI to write papers and assignments by students. The problem it it creates citations that don't exist and has factual errors (and some correct stuff - but you have to know enough to tell the difference). Some faculty have students create an AI paper and then their own paper is to correct and critique it.

Certainly using it to create a rough draft and then going in, fact checking, and cleaning it up, etc, might save some time and give you ideas you may not have thought of on your own. Of course AI is getting better over time but you still have the problem that developers use what is out there to train AI and if incorrect information is included then there is a problem. I worry the Captchas we get to deal with (where you have to check a motorcycle when it says bicycle, and 18 wheeler for a bus, some palm trees for a mountain - idiocies I have run into) are being used to train AL to then be used with self driving cars. No. Just no.

I agree, however, with good prompting the errors are almost eliminated on the latest models. The key is knowing what you are asking for. The kids at school won't know and that is a big issue. If we use it for general help like my example in previous post, it is a game changer for time saving. I'm actually pretty excited about the ability to scale activity with the latest more accurate models. This stuff is changing on a daily basis.