Latest / The Insurance Marketer / Harnessing AI for Marketing Success w/ Dale Steinke
Transcript
- 0:00SPEAKER_01: From the OneFire studio, welcome to the IMCA Peer-to-Peer Podcast, where we dive deep into technology, brand, social media, and other topics with an insurance industry expert.
- 0:13SPEAKER_01: Let's welcome today's guest.
- 0:15SPEAKER_05: Hi there, I'm Jim Flynn from OneFire, along with my co-host, Leslie Castillo from American Equity Underwriters.
- 0:22SPEAKER_05: And it's our pleasure to welcome you to today's episode of the IMCA Peer-to-Peer Podcast, where today we will be talking about best practices for creating great AI prompts that lead to better content.
- 0:36SPEAKER_05: And to help us better understand that journey, our guest today, Dale Steinke, is director for the Agent for the Future Advisors at Liberty Mutual Insurance, SafeCo Insurance, and State Auto Insurance.
- 0:51SPEAKER_05: Their mission is to empower independent insurance agencies to adapt to the rapid changes in technology and evolving customer expectations.
- 1:01SPEAKER_05: They equip their agents with the latest digital tools and strategies to wow clients and drive business growth.
- 1:08SPEAKER_05: As their artificial intelligence authority, he shares valuable insights and actionable strategies on integrating AI into sales, marketing, and other agency operations.
- 1:20SPEAKER_05: Dale considers himself a lifelong learner, continually seeking innovative ways for agents to leverage technology ranging from generative tools like ChatGPT and expanding array of AI-driven agency solutions.
- 1:37SPEAKER_05: Dale and the team at Agent for the Future Advisors are on a mission to help independent agents prosper now and in the future.
- 1:46SPEAKER_05: Today he will share with us some of the concepts he teaches agents with an eye towards what marketers and carriers, MGAs, and programs should consider as they create best practices for their organizations and look towards the future of the independent agent distribution channel.
- 2:07SPEAKER_05: Dale's graciously shared some accompanying slides that support the topics that we'll be discussing today.
- 2:13SPEAKER_05: And you can see the full video version of today's podcast on the IMCA's YouTube channel at YouTube slash at IMCANet slash videos.
- 2:26SPEAKER_05: Welcome to the podcast today, Dale.
- 2:29SPEAKER_04: Thank you so much, uh Leslie and Jim.
- 2:31SPEAKER_04: It's a true pleasure to be here.
- 2:32SPEAKER_04: I'm really excited about this.
- 2:34SPEAKER_04: Thanks for having me.
- 2:35SPEAKER_02: Yeah, we're looking forward to it.
- 2:38SPEAKER_02: Yeah, yeah.
- 2:39SPEAKER_02: So uh so where's home for you today?
- 2:41SPEAKER_02: And and is that where you're from originally?
- 2:43SPEAKER_04: Yeah, I'm I'm in Seattle, Washington.
- 2:46SPEAKER_04: Uh I did not grow up here, but I've been here for the majority of my life.
- 2:50SPEAKER_04: Uh, and I grew up in Milwaukee, Wisconsin.
- 2:52SPEAKER_04: So Midwestern boy.
- 2:54SPEAKER_04: Yeah, transplanted to the west coast.
- 2:57SPEAKER_04: Exactly.
- 2:57SPEAKER_04: Yeah, go pack.
- 2:59SPEAKER_02: What what made you move to Seattle?
- 3:01SPEAKER_04: My wife.
- 3:02SPEAKER_02: Okay.
- 3:03SPEAKER_02: That's a good answer.
- 3:04SPEAKER_04: She's uh she's she's from here originally.
- 3:06SPEAKER_02: That's a good answer.
- 3:10SPEAKER_05: Never look back.
- 3:11SPEAKER_05: Yeah, there you go.
- 3:12SPEAKER_05: Oh man, that's funny.
- 3:15SPEAKER_05: Well, I tell you, it, you know, I I've been thinking about this episode and and kind of framing it in my mind, thinking about the conversations that I've you know had with different IMCA members.
- 3:26SPEAKER_05: And you know, and it's clear that there's a you know, there's there's a lot of uh different ways people are using AI.
- 3:36SPEAKER_05: Uh some have really embraced it and really use it in their day-to-day operations and and so forth, and and others, uh I would hazard to say are kind of um doing it secretively, you know, because maybe the organization hasn't, you know, fully brought it online yet or believed in it or adopted it and so forth.
- 3:58SPEAKER_05: So you know, can you tell us you know more about how you envision marketers using AI today and and what additional opportunities maybe you envision that the that are going to come online in the future as well?
- 4:14SPEAKER_04: Oh yeah, lots of lots of possibilities.
- 4:16SPEAKER_04: You know, I the point you made about somewhat people using it secretly, you know, there's there's studies out there, Microsoft has one that says that 75% of knowledge workers uh use AI, and a lot of them are bringing it to work.
- 4:28SPEAKER_04: So if your business does not have an AI policy, you need to get one because people are using this stuff all the time.
- 4:35SPEAKER_04: So speaking of that, you know, I have a policy that I need to go through really quickly for our company's policy.
- 4:41SPEAKER_04: Uh that, you know, this document may contain, or this is what I'm going through today, is may contain content generated by artificial intelligence, uh both con both uh written content and uh images.
- 4:52SPEAKER_04: And I've reviewed the content for accuracy and revised it where necessary.
- 4:57SPEAKER_04: So you know, we at Agent for the Future uh do a lot of research on our independent agents, and we know that there's a lot of interest in AI, but relatively few uh insurance professionals are using it in marketing or whatever else there is that they're trying to do with it.
- 5:11SPEAKER_04: Um, and for a lot of the people who you know have interest in it, I always hear I think AI can help, but how?
- 5:19SPEAKER_04: And when I think about AI, you know, it can do a lot.
- 5:21SPEAKER_04: It can be an analyst, it can be a strategist, a thought partner.
- 5:24SPEAKER_04: Uh, but um where one area can really shine is in content creation.
- 5:28SPEAKER_04: And for you know, all the marketers on this call, I mean, this is a this is a really sweet spot for you.
- 5:33SPEAKER_04: You know, anybody who does marketing in an agency or a carrier or an MGA or any other business in our industry, you know, the amount of work it can take to create that stuff can take quite a bit of time and it can be quite daunting.
- 5:45SPEAKER_04: Yeah, and so so Leslie and Jim, you know, you've probably been involved in a few marketing campaigns over the years.
- 5:50SPEAKER_03: Just a few.
- 5:50SPEAKER_04: Yeah.
- 5:51SPEAKER_04: Just a few.
- 5:52SPEAKER_04: So, what are some marketing campaign components you wish AI could help you create?
- 5:56SPEAKER_02: Oh, video.
- 5:58SPEAKER_02: Video for sure.
- 6:00SPEAKER_02: Uh blog posts.
- 6:03SPEAKER_04: Definitely.
- 6:05SPEAKER_02: Social posts.
- 6:06SPEAKER_02: Social posts for sure.
- 6:08SPEAKER_02: Yeah.
- 6:09SPEAKER_02: Email campaigns.
- 6:11SPEAKER_04: Oh, yeah.
- 6:11SPEAKER_04: Yeah, definitely.
- 6:12SPEAKER_04: That and the all the words.
- 6:14SPEAKER_02: All the words, all the words.
- 6:15SPEAKER_02: All the words.
- 6:15SPEAKER_02: Yeah, all the words, right?
- 6:16SPEAKER_05: It's like well, and visuals too.
- 6:18SPEAKER_05: I mean, you said video, but still pictures, you know, that don't have seven fingers.
- 6:23SPEAKER_05: Those would be good.
- 6:24SPEAKER_04: That's yeah, that's that's that's a thing.
- 6:27SPEAKER_04: Yeah, still a thing.
- 6:28SPEAKER_04: Uh yes, we see that once in a while.
- 6:30SPEAKER_04: Yeah, so there's there's all those kinds of things out there that you you want to do.
- 6:34SPEAKER_04: It's anything that's creative, you know, it can help you with that.
- 6:37SPEAKER_04: And you know, I think when you're looking, you know, looking forward, uh, you're gonna see you know better photos.
- 6:42SPEAKER_04: They're gonna figure out the seven-finger thing, they're gonna figure out you know, how when you change a photo that like a wisp of hair that's in the previous version doesn't stay there sitting on your shirt while you're you know when the photo's changed.
- 6:54SPEAKER_04: Uh you're gonna see you know higher quality video.
- 6:58SPEAKER_04: Um, and one area I think is really interesting is you're gonna see this um more readily available, you know, hyper-segmented marketing automation down to the individual.
- 7:07SPEAKER_04: So I will know of what Jim is interested in, I'll know what Leslie's interested in, and be able to provide you content automatically that is more relational to you because it talks about the things you care about.
- 7:19SPEAKER_04: Scary it's I think I think the era of sort of that mass marketing uh you know is gonna be you know undifferentiated.
- 7:27SPEAKER_04: I think it's it's it's not necessarily on the way out, but I think it's gonna diminish and be replaced by things like that one-to-one.
- 7:34SPEAKER_04: Yeah, and AI is really helping drive that.
- 7:37SPEAKER_05: You know, I I want to go back before I forget this for a second.
- 7:41SPEAKER_05: Uh you mentioned AI policy and how important that is.
- 7:44SPEAKER_05: Is there a place that you would recommend you know the the modeling it after some particular AI?
- 7:52SPEAKER_04: Yeah, I don't have the resources at my fingertips right now, but there are a bunch of them.
- 7:56SPEAKER_04: If you're an independent agent, uh the big eye, IABA, has a has a policy uh template that you can use.
- 8:03SPEAKER_04: Um, there are a bunch of other business organizations that you can go out there and find.
- 8:08SPEAKER_04: It's pretty easy to Google that.
- 8:10SPEAKER_04: And uh you know, another sort of a meta way of doing it, ask GPT to help you write one or another generative AI model.
- 8:17SPEAKER_04: And it actually comes up with a pretty good uh foundation for you to work from.
- 8:21SPEAKER_02: Well, I'm sure it's been asked many times.
- 8:24SPEAKER_02: So it's got a lot to pull from, right?
- 8:27SPEAKER_02: So um speaking of ChatGPT, so you mentioned that you use ChatGPT for some of the visuals in this presentation, or at least I'm assuming so, based on your the uh disclaimer.
- 8:38SPEAKER_02: What other options are out there?
- 8:41SPEAKER_02: I think it feels like AI has the term AI has become synonymous with ChatGPT, but like that's not that's not real.
- 8:47SPEAKER_02: So what what other options are available and how how can a person determine which one is the best fit for what they're looking to do?
- 8:55SPEAKER_04: So it's funny is I I do a lot of research trying to figure out, hey, when I'm talking with people, is like what's out there that you can use.
- 9:01SPEAKER_04: And every time I look at the different models, you know, they don't have it, and then guess what?
- 9:05SPEAKER_04: Next week they do have it.
- 9:07SPEAKER_04: So, you know, it's it's hard to say now, you know, all of them now.
- 9:10SPEAKER_04: Chat GPT um integrated Dolly, which has been around for quite a while.
- 9:14SPEAKER_04: Uh so I think the the base thing I would say is, you know, whatever generated model you're using, check it out, you know, ask it to create some images for you and and judge for yourself whether it's up to snuff.
- 9:27SPEAKER_04: Uh they some of them have real trouble like creating infographics, putting text on an image.
- 9:31SPEAKER_04: It just it starts to get weird really fast and illegible.
- 9:36SPEAKER_04: Um, you know, and and others just make things that have that sort of AI sheen, you know, that doesn't look quite right.
- 9:42SPEAKER_04: But here, get this.
- 9:43SPEAKER_04: So I want to ask you a question.
- 9:44SPEAKER_04: So, what do you think is the second most chat GPT is the most visited um gen AI model?
- 9:49SPEAKER_04: What's the second most visited one?
- 9:52SPEAKER_04: Is it a Microsoft?
- 9:54SPEAKER_04: It's a it's a it's not what you're probably thinking of.
- 9:57SPEAKER_04: No, it's not Microsoft.
- 9:59SPEAKER_02: I feel like this is a trick question.
- 10:02SPEAKER_02: Like I want to say Google or something, you know, or uh I don't know.
- 10:06SPEAKER_02: I don't know.
- 10:06SPEAKER_04: I'm stumping.
- 10:07SPEAKER_04: Okay.
- 10:08SPEAKER_04: So um according to uh it's like AI tools.xyz, it's Canva.
- 10:14SPEAKER_02: Oh, yeah.
- 10:16SPEAKER_04: So interesting, yeah.
- 10:17SPEAKER_04: Canva, you know, and that what that leads me to sort of another thing that I like to talk about too is that there are these standalone tools like GPT, but then you're seeing it getting baked into things that we use every day.
- 10:27SPEAKER_04: So whether it's your marketing automation tools or your creative tools like Canva, you're seeing AI get sort of built into them.
- 10:35SPEAKER_04: Um so you know, it's a really interesting space.
- 10:38SPEAKER_04: But I do have to say though, too, is that you know, when you're thinking about which ones to use, not only do you look at the quality, you really want to look at the potential copyright or IP or trademark issues that you might run across.
- 10:50SPEAKER_04: You know, the different tools have different standards for what they'll say is you know, you own it or they own it.
- 10:57SPEAKER_04: Um, you know, do they take trademark stuff and logos and brands and repurpose those?
- 11:02SPEAKER_04: Because that's probably gonna get you into trouble.
- 11:04SPEAKER_04: And you know, I'm not a lawyer, so don't take my word for it.
- 11:08SPEAKER_04: But you know, consult a legal expert uh as needed would be my advice.
- 11:13SPEAKER_05: Well, and we just uh did a podcast not long ago with a lawyer who spoke about that.
- 11:18SPEAKER_05: So yeah, no, absolutely there's there's some challenges, and uh I mean, and one of them, as you know, is content that is generated by AI is not copyrightable.
- 11:28SPEAKER_05: Um now, at what point does it become copyrightable?
- 11:33SPEAKER_05: How much do you have to massage it?
- 11:36SPEAKER_05: You know, there's there's a lot out there yet, and I suspect in the coming years there's gonna be a lot of case law that starts to support things.
- 11:44SPEAKER_05: But right now it's uh kind of the wild, wild west, isn't it?
- 11:46SPEAKER_05: It's a little unsettled at this point.
- 11:48SPEAKER_04: So you know, I I can give you some advice a little bit later about some of the things I I think about when I think about images and what you might actually want to do.
- 11:55SPEAKER_05: Yeah, yeah.
- 11:57SPEAKER_05: Well, you know, we see a lot of AI, you know, coverage about how how AI can help, and we also see coverage on you know some of the things you mentioned hallucinations, you know, and that certainly is the case.
- 12:11SPEAKER_05: And if you've used it at all and you've asked it to cite its work, it sometimes comes up with uh some something that doesn't exist.
- 12:20SPEAKER_05: So it's it's a little strange.
- 12:23SPEAKER_05: Definitely see that.
- 12:24SPEAKER_05: You know, what are what are what have you found are some of the best practices to create the types of content that we that we talked about when you asked the question, you know, what do you need for a campaign?
- 12:34SPEAKER_04: Yeah, yeah.
- 12:35SPEAKER_04: It's like, you know, I think about you know, when you're starting off in that campaign work, you know, you're staring at a blank screen, you know, who hasn't been there before?
- 12:41SPEAKER_04: Right.
- 12:42SPEAKER_04: Um, you know, that's where I think that AI assistant really can come into play.
- 12:46SPEAKER_04: Uh, you know, it's I look at it as you know, as a you know, if you're looking at it as a marketing assistant, you know, it can create content and craft it in your brand voice.
- 12:55SPEAKER_04: Um, and just like things created by humans, you know, sometimes it'll get it wrong.
- 13:00SPEAKER_04: And you know, when you think about that, you know, guess who created AI?
- 13:04SPEAKER_04: Humans.
- 13:05SPEAKER_04: So you're gonna have those issues.
- 13:07SPEAKER_04: Uh and uh you know, when I talk about it, um, you know, I I think that, you know, I I kind of focusing to say a little bit more on like the GPTs of the world kind of thing, but a lot of the things we'll talk about are things that uh are gonna work in any of these models, whether it's you know, whether it's GPT or Canva or some of these other tools that are out there.
- 13:27SPEAKER_04: Uh and so I, you know, when you know it's gonna get it wrong, uh, and you know, I always treat it as a first draft.
- 13:34SPEAKER_04: Uh you really have to be super careful.
- 13:37SPEAKER_04: It's don't ever take it at word at its word.
- 13:39SPEAKER_04: You know, you had a legal expert on, legal expert, I don't know if they mentioned the the lawyers, uh fellow lawyers who have been um censured by their the the judges because they had fake case law, you know, that AI generate.
- 13:51SPEAKER_04: They didn't check it.
- 13:52SPEAKER_04: You know, tell me your sources and and then validate that it's real.
- 13:56SPEAKER_05: Yeah, no, absolutely.
- 13:58unknown: Yeah.
- 13:59SPEAKER_05: So so prompts, tell us about the creating the great prompts.
- 14:03SPEAKER_04: Yeah, yeah.
- 14:04SPEAKER_04: So there's a a co-worker of mine, Jules Brown.
- 14:06SPEAKER_04: Um, she came up with this.
- 14:08SPEAKER_04: I really love this.
- 14:08SPEAKER_04: It says the eloquence of your request dictates the richness of the response.
- 14:13SPEAKER_04: And and so I'm gonna I want to illustrate that.
- 14:15SPEAKER_04: Uh let's let's do a little experiment.
- 14:17SPEAKER_04: So, Jim, if I asked you to make me a sandwich, what would it be?
- 14:23SPEAKER_05: Well, let's see.
- 14:24SPEAKER_05: Um, I'm gonna start with I want it to uh uh be something uh crab based.
- 14:31SPEAKER_05: That's it.
- 14:31SPEAKER_05: It's it gonna be a kind of a crab sandwich.
- 14:34SPEAKER_05: Uh maybe add a little bit of maybe a little spicy stuff like some jalapenos and uh this yeah, yeah, yeah.
- 14:42SPEAKER_05: Some maybe some uh a little bit of onion on there, maybe a little bit of smoked paprika.
- 14:48SPEAKER_05: Uh, you know, put it on a nice uh croissant, maybe, you know.
- 14:52SPEAKER_05: Sounds fantastic.
- 14:53SPEAKER_04: Um, you know, and it's I'm on the West Coast here, you know, lots lots of dungeness crab.
- 14:58SPEAKER_04: Oh, now it's a good thing you're giving me that sandwich because my wife hates seafood, so even though she grew up here, at least crab no way.
- 15:06SPEAKER_04: So not a good answer for ask me that question, yeah.
- 15:09SPEAKER_04: For me, it's good.
- 15:10SPEAKER_04: Okay, Leslie, what about you?
- 15:11SPEAKER_02: Wait, okay.
- 15:12SPEAKER_02: I don't know if I understand the question.
- 15:13SPEAKER_02: So am I pretending to be Chad GPT?
- 15:18SPEAKER_04: I'm just saying I ask you the question, make me a sandwich, what would you give me?
- 15:21SPEAKER_04: Sorry, I'm an overthinker.
- 15:23SPEAKER_02: Um I'm kind of like a uh I'm like a Chick-fil-A gal.
- 15:27SPEAKER_02: Maybe it's a chick, maybe a fried chicken sandwich.
- 15:29SPEAKER_02: Also, I live in the south, so I put some pickles on it, some some dukes.
- 15:34SPEAKER_04: Very nice.
- 15:34SPEAKER_04: I like I like those too.
- 15:36SPEAKER_04: That's great.
- 15:37SPEAKER_04: You know, but um, you know, so it's it's kind of funny though, as you think about that.
- 15:41SPEAKER_04: Like I asked you both the same question, but I got very different responses.
- 15:44SPEAKER_04: And, you know, it's because, you know, just like with AI, if I did the same thing with it, you know, I don't know what I'm gonna get.
- 15:51SPEAKER_04: Um but you know, maybe I had something really special in mind.
- 15:55SPEAKER_04: Uh, and so instead of you know, instead of just saying, make me a sandwich, you know, what if I turned it around and said, Leslie, um, you're a fantastic cook.
- 16:04SPEAKER_04: Thank you.
- 16:04SPEAKER_04: Oh, good.
- 16:05SPEAKER_04: See, I'm already buttering you up.
- 16:06SPEAKER_04: This is good.
- 16:07SPEAKER_04: Um, so please make me a sandwich.
- 16:10SPEAKER_04: It should be a Reuben on lightly toasted rye with organic sauerkraut and non-dairy cheese.
- 16:16SPEAKER_04: Oh, and avoid meat because I'm vegan.
- 16:19SPEAKER_02: That sounds terrible.
- 16:22SPEAKER_02: Okay.
- 16:23SPEAKER_04: Well, I don't even know what to do.
- 16:27SPEAKER_04: Yeah.
- 16:28SPEAKER_04: And that's kind of why I threw it out there just to see what I wanted to test AI to see what would come with it.
- 16:33SPEAKER_04: I figured I could stomp it, right?
- 16:34SPEAKER_04: Yeah.
- 16:34SPEAKER_04: And you know, and you're probably thinking I'm pretty micromanage-y there too, and asking for all that.
- 16:39SPEAKER_04: But you know, at least it gives you an idea of what I really want.
- 16:42SPEAKER_04: And chances are you're gonna deliver something close to what I expected.
- 16:46SPEAKER_04: Um, and so, like in the same way that sort of little back and forth went, it's it's the same thing with AI.
- 16:51SPEAKER_04: You know, if you give it specific instructions, you're more likely to get that better um response back.
- 16:57SPEAKER_04: You know, because it's sorting through all the information that it is sucked up over, you know, it's training time and said, hey, there's a bunch of recipes out there, I'll I'll find one that uh will meet your needs.
- 17:09SPEAKER_04: So just like asking a friend to make a sandwich or a coworker to write some content, you know, the better job you do of explaining what you want, the more likely you are gonna get to get the response that you expected.
- 17:21SPEAKER_04: So I did this test with the sample the sandwich scenario with uh AI.
- 17:26SPEAKER_04: And for making me a sandwich, the first time I got a ham sandwich.
- 17:30SPEAKER_04: And um, and then another time I got a turkey club.
- 17:33SPEAKER_04: So, and that's because there's a lot of information about ham sandwiches and turkey clubs online.
- 17:38SPEAKER_04: Now, when I when I thought, okay, I'm gonna stump it with this Rubin thing, right?
- 17:42SPEAKER_04: So I ask it for the Rubin thing, and it gives me a really detailed recipe that actually sounded pretty good.
- 17:48SPEAKER_04: I like to cook, so you know, I look, I read it and I'm like, yeah, that's certainly doable.
- 17:52SPEAKER_04: Turns out that vegan Rubens are a thing.
- 17:56SPEAKER_04: So ChatGPT one, Dale Zero.
- 18:00SPEAKER_04: Um and so then I asked an AI image generator to create um a picture of both of these sandwiches.
- 18:09SPEAKER_04: You know, for the ham sandwich, it looked pretty, pretty nasty.
- 18:12SPEAKER_04: The pastrami sandwich, you know, the vegan, it's got steam rising off of it.
- 18:16SPEAKER_04: And there's a frosty beverage, too, which um concerns me a little bit that it it thought, you know, to give me that as an answer.
- 18:23SPEAKER_04: So I'm like, hmm, how well does it know me?
- 18:25SPEAKER_02: It's thinking you're a vegan, you need some kind of excitement in your life, right?
- 18:31SPEAKER_04: So pretty funny.
- 18:33SPEAKER_04: But uh yeah, so I thought it was kind of interesting.
- 18:35SPEAKER_04: But so so I think the you know, that whole exercise, the point of it is is just like uh a recipe, prompts have specific ingredients.
- 18:44SPEAKER_04: And uh there are um six primary ingredients, and I'll walk you through those.
- 18:49SPEAKER_04: Um the first one is a role.
- 18:51SPEAKER_04: So I said in in the first example was you're a fantastic cook.
- 18:55SPEAKER_04: Um, it could be you're an excellent marketer or an expert marketer, you're a data analyst.
- 19:00SPEAKER_04: Uh the next thing is instructions, make me a sandwich.
- 19:03SPEAKER_04: You know, so I'm putting you down a path.
- 19:06SPEAKER_04: Uh and then there's another one called context.
- 19:09SPEAKER_04: Uh and the idea with that one is you give it information that you've done, uh, say maybe you have a previous output that you really like and you want it to mimic it, you give it that.
- 19:17SPEAKER_04: I couldn't give it a recipe because I didn't know the recipes existed, so that probably wouldn't be such a great idea.
- 19:22SPEAKER_04: Then there are um requirements, you start to get into details, you know, that it's got to have non-dairy cheese and organic sauerkraut.
- 19:29SPEAKER_04: And then there are boundaries, the things you don't want it to do, like no meat, please.
- 19:36SPEAKER_04: And then um there's another one called reasoning, and the idea behind that one is you can either ask the prompt or the AI before it does gives you an answer, tell me how you're going to uh make this calculation or how you're gonna come up with this result, and it'll give you the steps.
- 19:53SPEAKER_04: So then you can look at that and say, okay, that sounds good, go for it.
- 19:56SPEAKER_04: Or after the fact, how did you come up with that?
- 20:00SPEAKER_04: Uh which I find myself using quite a bit.
- 20:03SPEAKER_04: And you probably will too.
- 20:05SPEAKER_05: I mean, is it uh I'm I've never done that.
- 20:07SPEAKER_05: Uh that last part of it.
- 20:09SPEAKER_05: Um, does it say uh I I I searched you know 27 million sites, or uh, you know, how does it um you know it's it's interesting.
- 20:20SPEAKER_04: I have a good example for you.
- 20:22SPEAKER_04: But uh yeah, there they it does uh it'll try and give you some background on how it did it.
- 20:27SPEAKER_04: You know, the reasoning could say that I did this calculation this way, or I pulled it from available sources.
- 20:32SPEAKER_04: Depending on the model, um, some of them give you links back to the sources.
- 20:36SPEAKER_04: And I've had issued times where I've looked at the source and I went, hey, this source has nothing to do with what you said here.
- 20:42SPEAKER_04: Could you look at that again?
- 20:43SPEAKER_04: And it says, Oh, I'm sorry.
- 20:44SPEAKER_04: And it gives you a different response with a link that actually does go back to the information you're seeking.
- 20:51SPEAKER_02: Yeah, that's happened to me a lot.
- 20:53SPEAKER_02: And I I tell that story to to colleagues and to remind them that they have to check the work.
- 21:01SPEAKER_02: Because yeah, inevitably Chat GBT will always say, and they're so polite about it, it'll say, Oh, you're right.
- 21:10SPEAKER_03: Yeah.
- 21:10SPEAKER_02: Actually, that wasn't the right source.
- 21:12SPEAKER_02: I should have done it this way.
- 21:13SPEAKER_02: And you're just like, Well, you didn't tell me that.
- 21:17SPEAKER_02: It's a good thing I thought to ask.
- 21:19SPEAKER_04: Yeah, or you could do, I don't like that response, can you do it again?
- 21:22SPEAKER_04: And it'll give you a different response.
- 21:24SPEAKER_04: You know, yeah.
- 21:24SPEAKER_04: It's a people pleaser, so it's gonna try and give you a response that you like.
- 21:29SPEAKER_04: Not always to the good.
- 21:30SPEAKER_04: So you really, yeah, really have to be careful about that.
- 21:33SPEAKER_05: Yeah.
- 21:34SPEAKER_05: I am always amazed with how rapidly it gives you the response.
- 21:39SPEAKER_05: It's just like, what?
- 21:43SPEAKER_02: I know, I know, I love it.
- 21:44SPEAKER_02: I love it.
- 21:45SPEAKER_02: But I also love that we're framing this conversation around using it as an assistant because that's something that I think is really, really important.
- 21:52SPEAKER_05: We'll take a quick break and come right back.
- 21:55SPEAKER_00: Are you looking to take your company's growth to the next level?
- 21:58SPEAKER_00: Look no further than.
- 22:00SPEAKER_00: OneFire, the RevFox-focused marketing agency that specializes in helping insurance companies grow.
- 22:06SPEAKER_00: Our expert team of strategists, marketers, and house five specialists will help you optimize your marketing efforts and maximize your ROI.
- 22:14SPEAKER_00: From lead generation to conversion and retention, OneFire has you covered with our deep insurance industry expertise and modern approach to growth.
- 22:23SPEAKER_00: Don't let your competitors leave you in the dust.
- 22:26SPEAKER_00: Ignite your growth with Onefire.
- 22:28SPEAKER_00: Visit us today at onefire.com.
- 22:33SPEAKER_02: So the steps that you just described are great instructions on how to rewrite something, you know, to be more appropriate for your needs.
- 22:43SPEAKER_02: What do people need to do with the AI content before they put it out there for the world to see?
- 22:50SPEAKER_04: Yeah, it's review everything.
- 22:52SPEAKER_04: Review, review, review.
- 23:00SPEAKER_04: You don't just go, hey, great, thanks.
- 23:02SPEAKER_04: I'm just going to publish it up on the on the the white.
- 23:04SPEAKER_04: That's going to be our new ad campaign and it's all good to go.
- 23:06SPEAKER_04: Yeah.
- 23:07SPEAKER_04: No, no, no, no.
- 23:08SPEAKER_04: Hold on a second there.
- 23:10SPEAKER_04: Take in and give it a good review.
- 23:12SPEAKER_04: Yeah, good advice.
- 23:13SPEAKER_04: So let me let me jump into something.
- 23:15SPEAKER_04: So like so now we're talking about trying to, we were talking about at the beginning of this trying to build a marketing campaign.
- 23:19SPEAKER_04: You know, sandwich recipes, that's great, but you know, if we're trying to build a marketing campaign, um so you know, I I think about I came up with an idea.
- 23:28SPEAKER_04: I said, okay, I want to create a marketing campaign for uh a fictional uh insurance agency, uh, insurance for Dayton in Dayton, Ohio.
- 23:37SPEAKER_04: And um I asked it to uh write some content for me.
- 23:41SPEAKER_04: And so I said, okay, uh, you're an expert insurance marketer, so I'm telling it its role, first of all.
- 23:47SPEAKER_04: I want you to write an article for my website, Insurance for Dayton, that explains why a prospective client would want to do business with an independent insurance agent rather than using a captive agent or a direct carrier.
- 23:57SPEAKER_04: And if there are any of those folks on here, I apologize.
- 24:00SPEAKER_04: Um that's the basic instructions.
- 24:03SPEAKER_04: Uh and then I say, assume the person reading this knows very little about independent agents, which is massively true.
- 24:11SPEAKER_04: And uh that's not part of the prompt, but yes, they don't know much about agents.
- 24:15SPEAKER_04: And then say, you know, write SEO-friendly content.
- 24:18SPEAKER_04: So I'm trying to think about you know how I want this to be found that's aimed at homeowners and small business owners near Dayton, Ohio who are looking for insurance.
- 24:25SPEAKER_04: So I've I've given it more requirements.
- 24:27SPEAKER_04: I'm looking for my ideal customer.
- 24:29SPEAKER_04: And then I gave it some guardrails and I said in boundaries and said, look, our ideal customers are interested in good coverage, not people searching for cheap insurance with low limits.
- 24:38SPEAKER_04: And so what ended up I ended up getting was a really detailed, really nicely written article uh that gave me about 750 words in a few seconds.
- 24:49SPEAKER_04: Uh, you know, you talk about speed, Jim, you know, it literally is a few seconds.
- 24:53SPEAKER_04: Yeah.
- 24:54SPEAKER_04: And um, you know, depending on the complexity of the thing, it's it's it's spitting it out right away.
- 24:58SPEAKER_04: Yeah.
- 24:59SPEAKER_04: It doesn't have to think about it.
- 25:00SPEAKER_04: Um more complicated things, I've seen the wheel spin for a while, but you know, that's usually you're getting into high data stuff.
- 25:06SPEAKER_04: So talked about you know, choice and flexibility, personalized service, local expertise, advocacy and support, long-term relationships uh that you know we have with our customers, um, carrier network, experience and expertise, customer-centric.
- 25:21SPEAKER_04: It just you know went on and on and on, and it was just like, wow, this is pretty amazing, um, amazing stuff.
- 25:28SPEAKER_04: But you know, I look at that and then I it needs some work.
- 25:32SPEAKER_04: So back to your earlier question, Leslie, you know, it's like I think you know, it's a back and forth conversation too.
- 25:37SPEAKER_04: So um I look at the draft and I say, hey, make it more light and engaging, or make it more serious, or remove that section and replace it with this one, or make it shorter, make it longer, uh, rewrite it using these SEO terms.
- 25:49SPEAKER_04: Uh create a version that sounds like the way I write or the way my brand voice sounds.
- 25:55SPEAKER_04: Uh and uh one of my favorites though is asking it what other information do you need to effectively answer my prompt?
- 26:02SPEAKER_02: I love that question because if you don't ask, it won't, it just it's just like, well, I'll just do it, I'll do what I can with what I've got.
- 26:11SPEAKER_02: It's amazing the questions that it will spit back at you.
- 26:15SPEAKER_04: And and you can use it as a coach too, and say, look, uh, I'm trying to do this thing, I don't know how to ask it.
- 26:20SPEAKER_04: What what kind of information do you need?
- 26:22SPEAKER_04: And it will it will tell you.
- 26:23SPEAKER_04: So when I did this exercise, I thought it was really interesting because it was basically like a textbook for a marketer.
- 26:30SPEAKER_04: It walked down all the different kinds of elements, like target audience details.
- 26:34SPEAKER_04: It asked me for specific demographics, what's their age, income level, et cetera?
- 26:38SPEAKER_04: What are their psychographics?
- 26:39SPEAKER_04: That's one of my favorites.
- 26:40SPEAKER_04: What are their values, their lifestyles?
- 26:42SPEAKER_04: What do they believe?
- 26:43SPEAKER_04: You know, um, that you're trying, you know who are these people you're trying to reach?
- 26:47SPEAKER_04: Um are there key messages that your brand uses?
- 26:49SPEAKER_04: Are there, is there a tone and style that's specific to your agency?
- 26:54SPEAKER_04: And I want to get into that or your business and I'll talk about how you can make that work.
- 26:58SPEAKER_04: There are um, you know, are there calls to action you use?
- 27:01SPEAKER_04: Are there things you don't want me to talk about?
- 27:03SPEAKER_04: Uh, or is there feedback you've gotten before about stuff you've written and said, hey, it should change this, I can do that for you?
- 27:08SPEAKER_04: And then it'll even ask, are there other pieces of content you need?
- 27:12SPEAKER_04: And one of the things I like in there is it talks about um are you doing things on on video platforms?
- 27:18SPEAKER_04: Uh do you need infographics?
- 27:20SPEAKER_04: And if it can't create it, it can at least give you a script for a video or directions and direction for how to do an infographic that would cover these things.
- 27:30SPEAKER_04: So to the voice piece, you hear this all the time.
- 27:33SPEAKER_04: Oh, AI sounds like a robot, right?
- 27:35SPEAKER_04: I hear that often.
- 27:36SPEAKER_04: And you see things like, oh, we're gonna delve into this topic.
- 27:40SPEAKER_04: Who says delve?
- 27:40SPEAKER_04: I'm sorry, nobody.
- 27:42SPEAKER_04: Um you know, we're gonna dive deep on this.
- 27:45SPEAKER_04: It's like it's like business one-on-one awful speak.
- 27:49SPEAKER_04: Um, you know, in conclusion is another one too.
- 27:51SPEAKER_04: It's like it has this great thing, and then it goes in conclusion.
- 27:54SPEAKER_04: No one says that.
- 27:55SPEAKER_04: So so I could solve for that.
- 27:58SPEAKER_04: Uh, and here's how you do it.
- 27:59SPEAKER_04: Um, what you need to do is feed it some text that's representative of what you've written or uh content that matches your brand voice.
- 28:05SPEAKER_04: If you've already got a brand style guide, even better.
- 28:09SPEAKER_04: If you just have the text, you dump it in there, like several hundred, couple thousand words, and ask it to analyze it.
- 28:15SPEAKER_04: So, you know, hey, describe the writing style and tone of this text, and then ask it to write a prompt so that next time when you want to write in that style, you have that prompt ready to go.
- 28:27SPEAKER_05: Oh nice.
- 28:29SPEAKER_05: There's a hack for you.
- 28:30SPEAKER_04: I like that one.
- 28:32SPEAKER_04: It's I love asking it to write prompts for me, you know, because it's like here's the thing I want to do again and again and again.
- 28:38SPEAKER_04: So rewrite it so that you'll know what I'm talking about.
- 28:41SPEAKER_04: And you know, it'll analyze everything, it'll go really deep on the things you do, um, what you're looking for, uh, in that tone and style.
- 28:50SPEAKER_04: And you know, it wrote a really structured prompt for me, four sentence long prompt that I can use every time.
- 28:57SPEAKER_04: And now, depending on the model you're using, some of them you can actually input it as a default.
- 29:02SPEAKER_04: Uh, so that every time you use it, it creates that.
- 29:04SPEAKER_04: Or some of the AI agents now you can start to build uh uh you know an agent that always writes in your tone and style.
- 29:11SPEAKER_04: Nice.
- 29:12SPEAKER_04: Yeah, really super interesting stuff.
- 29:15SPEAKER_02: Yeah, that's super helpful because you so you don't have to repeat yourself every time.
- 29:20SPEAKER_02: You know, asking it to write in a particular voice.
- 29:23SPEAKER_04: Yeah, every new chat you do, it forgets what it did before.
- 29:26SPEAKER_04: So, you know, you really want to make sure that you're you're getting that right.
- 29:29SPEAKER_04: And so, you know, earlier we were talking about getting things wrong, and you know, we're doing a lot of massaging and and tweaking of it to get to this point, but you still need to be editing it and making sure that it's it's all accurate.
- 29:42SPEAKER_05: Makes sense.
- 29:44SPEAKER_05: So AI is iterative, as you've just you know shown us.
- 29:51SPEAKER_05: Um yeah, how do you ask it to rewrite it if it gets it wrong?
- 29:57SPEAKER_05: And for those of you that are listening to the podcast, I would really encourage you to take a look at the full video version of this on YouTube because there's some great slides that are going along with this that have those um rather than probably taking furious notes, you can take a look at these slides and and use them on a on a regular basis.
- 30:18SPEAKER_05: So anyway, getting into this, uh how do you know how do you get it to do things more appropriately?
- 30:24SPEAKER_05: Yeah, you know, so that the you really get what you can use.
- 30:29SPEAKER_04: Yeah, you know, so you're you're going back and forth with it.
- 30:32SPEAKER_04: Uh, you know, you can you can say, hey, I'd I'd prefer it to be written this way instead, and it will change it.
- 30:37SPEAKER_04: So maybe maybe it's getting uh uh too detailed or not detailed enough.
- 30:43SPEAKER_04: Or, you know, I I would say too, is anything that's you know, especially with insurance, highly regulated industry, you really want to make sure the things that it's saying are accurate.
- 30:51SPEAKER_04: Um, so everything goes through you know multiple levels of reviews.
- 30:55SPEAKER_04: Now, there's another concept though, too, is as you're going through this.
- 30:59SPEAKER_04: So, prompt chaining, uh, that's a concept I think that with all of this stuff is super important.
- 31:04SPEAKER_04: You know, you were mentioning before it's iterative.
- 31:06SPEAKER_04: Um and uh prompt chaining is is a uh it's a way to take tasks and break them up into bite-sized pieces, you know, create an article.
- 31:16SPEAKER_04: And once you like that and you've got that thing accurate, then what's the next thing you want to create?
- 31:21SPEAKER_04: Um so it's you know it's just a fancy way of saying let's do things one step at a time.
- 31:24SPEAKER_04: Because if it makes a mistake in the beginning, it's just gonna keep replicating that all the way through the whole rest of your marketing campaign.
- 31:30SPEAKER_04: So you want to fix that before it gets to that next step.
- 31:34SPEAKER_04: Um so yes, I think it's kind of a fun little term to drop in a you know networking conversation.
- 31:39SPEAKER_04: It's like we'll boost your AI cred a little bit.
- 31:44SPEAKER_04: So speaking of that, so of that prompt chain, so the next link in our prompt chain could be taking that article we created earlier and use it as a basis for creating social posts.
- 31:55SPEAKER_04: So I'm in the same chat with GPT as I've created this article, so I don't have to remind it to act like a marketer.
- 32:02SPEAKER_04: Um and I can tell it to use the same content we created earlier.
- 32:06SPEAKER_04: Uh and so um, you know, I'm asking it to write five posts each for LinkedIn, Facebook, and Twitter or X, whatever you want to call it these days.
- 32:15SPEAKER_04: And um I want to ask for reasoning.
- 32:18SPEAKER_04: Tell me what criteria you are using to uh write the post for each platform so you can see is it is it writing in a way that's gonna be most effective on that platform?
- 32:29SPEAKER_04: So it'll explain that to you.
- 32:31SPEAKER_04: Um, you know, I I said five posts of each, but you could say 10 or 20 or 100.
- 32:36SPEAKER_04: You know, you're gonna get tired long before AI ever gets tired of answering your questions or creating different versions.
- 32:43SPEAKER_04: And then, you know, I also ask it to you know highlight key points in the article of most interest to the people I'm trying to attract and follow our social media style and tone and encourage people to reach out to the agency.
- 32:56SPEAKER_04: So I'm trying to give it some really clear instructions on what I want it to do.
- 32:59SPEAKER_04: And you get posts back, and depending on the model you're using, it might include uh hashtags, it might include emojis.
- 33:06SPEAKER_04: You know, some platforms don't like uh hashtags anymore.
- 33:10SPEAKER_04: So hopefully, if it's following the right kind of standards, it would take those out if it's not right for that one.
- 33:15SPEAKER_04: Same thing for Facebook and and you know, Twitter and whatever.
- 33:19SPEAKER_04: And so the other thing you can do with it too is is um is you can say, look, um, I like the first two.
- 33:27SPEAKER_04: So can you take those two and and riff on those and do another 10 or 20?
- 33:32SPEAKER_04: So instead of having that blank screen, you can have a whole bunch of stuff to choose from right away that you can then start keep iterating on to come up with what really is actually gonna work for your brand.
- 33:42SPEAKER_02: It is amazing the the mental energy that that it will relieve you of.
- 33:51SPEAKER_02: And I think about if someone says, Oh, come up with 20 posts, I mean it's like, oh my god.
- 33:55SPEAKER_02: Like, you know, this is gonna take so long.
- 33:57SPEAKER_02: And just the fact that you can just click a button with some appropriate guidance, yeah, just feels so nice.
- 34:04SPEAKER_02: It feels nice to be living in the this world.
- 34:09SPEAKER_04: Yeah, yeah, it's like you were done a brainstorm exercise.
- 34:11SPEAKER_04: It's hard to brainstorm by yourself, right?
- 34:14SPEAKER_04: But AI gives you that that, you know, basically a whole group of people who are gonna throw a bunch of ideas at you, and then you can start you know really moving along pretty quickly.
- 34:21SPEAKER_04: Yeah.
- 34:22SPEAKER_04: Okay, so we've created created an article, created a bunch of social posts, um, say maybe you want to do a marketing automation pitch.
- 34:29SPEAKER_04: Um, and uh I like thinking about what's the next step in their journey.
- 34:33SPEAKER_04: So you've you've convinced them to do business with you, maybe they've already bought from you now.
- 34:37SPEAKER_04: What do you do about um people who say just bought one policy?
- 34:40SPEAKER_04: You could come up with a marketing automation campaign.
- 34:43SPEAKER_04: You know, in this example I was thinking about was a three-part campaign on the value of bundling additional policies with the agency.
- 34:50SPEAKER_04: And it comes up with that campaign, and then here's another step in that sort of iterative process.
- 34:57SPEAKER_04: You know, maybe it feels a little dry, it doesn't resonate with your ideal customers.
- 35:01SPEAKER_04: So, you know, if your ideal customers are millennials, you could ask it to create a campaign that's more upbeat and contains pop culture references that would appeal to millennials.
- 35:10SPEAKER_04: Um, and then also put in a thing there that says, you know, make sure that the pop culture references do not have potential copyright or trademark issues.
- 35:17SPEAKER_04: If you leave that out, it throws in all kinds of names of brands and things like that that are probably going to get you in trouble.
- 35:22SPEAKER_04: Uh and but if you take to have that in there, it gives you mentions about fast food wins, free guacamole, uh binging hit shows, um, and great redemption arcs.
- 35:34SPEAKER_04: You know, there's a mention there about our favorite meal deal, and I'm like, okay, probably not a concern, but if it's a happy meal, you know, some folks with the golden arches might be a little concerned by you using that.
- 35:44SPEAKER_02: Millennials like to eat, apparently.
- 35:47SPEAKER_02: Apparently, yeah, yeah.
- 35:48SPEAKER_02: Food.
- 35:49SPEAKER_04: Love it.
- 35:50SPEAKER_04: I think we're the core demo for all of those fast casual places, right?
- 35:54SPEAKER_02: I love it.
- 35:56SPEAKER_04: That's the so yeah.
- 35:58SPEAKER_04: So, you know, so you're just moving along here, you know.
- 36:00SPEAKER_04: So say you've got that that email part, you've got if you do text marketing, have it right, the text messages.
- 36:06SPEAKER_04: I mean, my gosh, you can crank those things out all day long.
- 36:10SPEAKER_04: Uh, and you know, and in instance.
- 36:13SPEAKER_04: And uh, you know, it'll put the emojis in there too, and that kind of thing.
- 36:16SPEAKER_04: So it makes it look really nice.
- 36:18SPEAKER_04: Uh so you know, and like you were saying earlier, Leslie, it's like that starting, you know, that's like, oh, there's so much to do and so much to, you know, nothing to start with.
- 36:26SPEAKER_04: This really helps you move along really quickly.
- 36:29SPEAKER_04: Um, how long do you think this whole thing probably took us?
- 36:31SPEAKER_04: If if I were just going through start to finish, 10 minutes, maybe, you know, minus, you know, plus whatever editing I do.
- 36:40SPEAKER_04: How long does it take you to type all this stuff out?
- 36:43SPEAKER_04: I mean hours.
- 36:45SPEAKER_04: Yeah.
- 36:46unknown: Yeah.
- 36:47SPEAKER_05: Hours.
- 36:48SPEAKER_05: Yeah, I mean teams of people, hours.
- 36:53SPEAKER_02: Yeah.
- 36:54SPEAKER_02: I mean, it probably took you longer to craft um an a a properly worded prompt than for the, you know, to have these deliverables in front of you.
- 37:04SPEAKER_02: And of course the editing, because it's very important, right?
- 37:07SPEAKER_02: You cannot just take what your assistant does and say it's gospel.
- 37:12SPEAKER_02: So yeah, this is fantastic.
- 37:14SPEAKER_04: Yeah, and I say it's like 60, 80% of the way there, right?
- 37:18SPEAKER_04: Versus 0% of the way there.
- 37:20SPEAKER_04: Uh again, it's not just, you know, generate it and post it, it's generate it and review it and massage it and make it make it your own.
- 37:29SPEAKER_02: Yeah, I think that's probably the best advice you can give anybody using AI.
- 37:33SPEAKER_02: Absolutely.
- 37:34SPEAKER_02: If a part of this campaign included generating images of people, uh, there's a good chance, like we talked about earlier, that they wouldn't have the right number of um fingers.
- 37:46SPEAKER_02: Or arms for that matter.
- 37:50SPEAKER_02: So uh which is kind of funny, right?
- 37:52SPEAKER_02: But then it can be frustrating because you start to lose confidence for me anyway.
- 37:56SPEAKER_02: I mean, I would start to lose confidence in in my assistant when I'm like, you don't even know how many fingers a person is supposed to have.
- 38:02SPEAKER_02: So can you give us some best practices for creating images with AI?
- 38:08SPEAKER_04: Yeah, you know, I think the people one is probably the most problematic of the bunch.
- 38:12SPEAKER_04: You know, you tend to see in some of the models where they they trained on whatever they trained on, it tends to be young models.
- 38:19SPEAKER_04: And, you know, it's not necessarily representative of what your customer base looks like.
- 38:24SPEAKER_04: One of the things that I really like doing is instead of doing the image thing first, I ask it to give me ideas for what a good image would look like for whatever the content is that I've created.
- 38:35SPEAKER_04: Uh, and it'll give you descriptions.
- 38:38SPEAKER_04: Uh so you know, we all have libraries of imagery.
- 38:41SPEAKER_04: Uh you can use that then to help sort of narrow down your search on the things you already own and have the rights to.
- 38:46SPEAKER_04: Uh you know, again, going back to that issue of, hey, is this thing working off of derivative imagery imagery and creating a potential uh copyright or trademark issue?
- 38:56SPEAKER_04: Uh the um, you know, but then, you know, so say you don't have it, then yeah, you can try the the AI stuff, but you know, you have to be really careful, I think, about feeding it information about the kind of demographic you're seeking.
- 39:09SPEAKER_04: Um, you know, what is the what are the the people are they middle-aged?
- 39:12SPEAKER_04: Are they, you know, and even then, you know, um are they senior citizens?
- 39:15SPEAKER_04: It it still does weird stuff.
- 39:18SPEAKER_04: Um, so you may find you're doing a lot of a lot of work to try and get what you want.
- 39:22SPEAKER_04: Now, there is that gray area we were talking to about about the legal side of things.
- 39:27SPEAKER_04: Um, there are tools uh, you know, that are very good at this too.
- 39:31SPEAKER_04: And and frankly, here's another thing is that all these things we're using today are the worst AI you're ever going to use.
- 39:38SPEAKER_04: It only gets better after this, you know, every day.
- 39:40SPEAKER_04: It's the the the the the capability.
- 39:44SPEAKER_04: Yeah, it can only get better.
- 39:46SPEAKER_04: Yeah.
- 39:46SPEAKER_04: Yeah, exactly.
- 39:47SPEAKER_04: And so like the cost and the capabilities are going up about 10x a year right now.
- 39:53SPEAKER_04: Um so it's so early days yet.
- 39:56SPEAKER_04: Um so I think someday we'll eventually get to the point where the imagery is working for us really well.
- 40:00SPEAKER_04: Um, you know, but uh Canva I think does a pretty nice job.
- 40:04SPEAKER_04: Um there's like synthesia is another one too that's pretty good.
- 40:07SPEAKER_04: Dolly.
- 40:08SPEAKER_04: Uh but try them out, see what you think, and just keep that intellectual property trademark copyright thing in the back of your mind.
- 40:15SPEAKER_05: Absolutely.
- 40:17SPEAKER_05: So yeah, I suspect that we could talk about this all day, and I suspect you you would have plenty to talk about.
- 40:25SPEAKER_05: But uh but we can't we can't do that.
- 40:27SPEAKER_05: So uh can you give us a quick recap?
- 40:30SPEAKER_05: Maybe you know, three hacks or something of that nature?
- 40:33SPEAKER_04: Sure, definitely.
- 40:34SPEAKER_04: Um so there's three things.
- 40:35SPEAKER_04: Um, one, remember to use those prompting components, those ingredients that we talked about, um, and to save your prompts so you can do future campaigns more quickly.
- 40:44SPEAKER_04: You know, there's uh most of these tools save your prompts.
- 40:47SPEAKER_04: Uh you can also create a separate prompting library.
- 40:49SPEAKER_04: There are tools out there that allows you to share prompts across teams, which I think is really interesting because then you're magnifying your capabilities as a group.
- 40:58SPEAKER_04: Um make it a back and forth conversation is number two, uh, you know, so you can keep tuning those results.
- 41:04SPEAKER_04: And three, always check the work.
- 41:08SPEAKER_04: Always.
- 41:09SPEAKER_04: And so, you know, I think as you look for ways to streamline your efforts and boost your output and get higher quality results, you know, look to that AI system to help get more done and to make your work and your iteration process easier and more fruitful.
- 41:23SPEAKER_04: Nice.
- 41:24SPEAKER_02: Very good advice.
- 41:26SPEAKER_02: Um, well, Dale, like Jim said, this is a hot topic.
- 41:29SPEAKER_02: I feel like every time we're talking about what episode do we want to record next, it's like AI, AI, AI.
- 41:36SPEAKER_02: We can't seem to get away from it.
- 41:38SPEAKER_02: Um, but we also can't seem to get enough of it because there's always something new to talk about.
- 41:43SPEAKER_02: So I'm sure this is not the last time that we see you in the insurance space in general or maybe even on this podcast, um, especially when it comes to AI.
- 41:51SPEAKER_02: So thank you again for your time.
- 41:53SPEAKER_02: And if any of our listeners want to connect with you about anything that they learned today, what is the best way for them to do that?
- 42:00SPEAKER_04: Well, first of all, I would be honored to be back on a future podcast.
- 42:03SPEAKER_04: Uh, if that happens, that'll be a great day for me.
- 42:06SPEAKER_04: Uh, so um, you can find me on AgentForthefuture.com or on LinkedIn.
- 42:12SPEAKER_04: Um, either one.
- 42:13SPEAKER_04: Uh, you know, LinkedIn is is great because you can message me directly at Agent for the Future has a message thing too.
- 42:19SPEAKER_04: Uh, but yeah, LinkedIn you can read a little bit more about me as well and some of my experience.
- 42:24SPEAKER_04: And I would love to hear from people and to learn uh from everybody who's listening.
- 42:30SPEAKER_04: So this has been a true pleasure.
- 42:32SPEAKER_04: Uh Jimmy Leslie.
- 42:33SPEAKER_04: Thank you so much.
- 42:35SPEAKER_05: Oh, absolutely.
- 42:36SPEAKER_05: The pleasure is uh that is ours.
- 42:38SPEAKER_05: It's it's been uh very interesting, very interesting and enlightening.
- 42:42SPEAKER_05: Well, that's a wrap uh for today's episode of the IMCA Peer-to-Peer Podcast.
- 42:48SPEAKER_05: Remember, you can see the full video version, and I really, really uh encourage you to do that because there's some great prompts there that you can use on a regular basis.
- 42:58SPEAKER_05: Um, and that's available on IMCA's YouTube channel at YouTube slash at IMCA net slash videos.
- 43:07SPEAKER_05: We hope you gained valuable insights into best practices for creating great AI prompts that lead to making better content.
- 43:15SPEAKER_05: And if you enjoyed today's episode, please don't forget to subscribe, give us a five-star rating, and leave us a review.
- 43:23SPEAKER_05: Leaving us a review helps us reach more listeners just like you.
- 43:28SPEAKER_05: And if you're looking uh for uh to give us some ideas on topics that you would like us to cover in the future, please shoot us over an email at ideas at imca peer, the number two peer podcast.com.
- 43:46SPEAKER_05: Once again, this is Jim Flynn from One Fire, along with my co-host Leslie Castillo from American Equity Underwriters.
- 43:55SPEAKER_05: And it's been our honor and a pleasure to spend this time together, along with our guest, Dale Steinke.
- 44:01SPEAKER_05: Thank you again, Dale, for being part of today's episode of the IMCA Peer to Peer Podcast.
- 44:07SPEAKER_05: Thank you so much.
- 44:08SPEAKER_05: Honor was all mine.
- 44:10SPEAKER_01: The Peer to Peer Podcast is a presentation of IMCA.
- 44:14SPEAKER_01: To learn more about today's episode and listen to other episodes, visit IMCANet.com.
- 44:19SPEAKER_01: This has been a production of One Fire.