Latest / The Insurance Marketer / Unlocking AI’s Potential in Marketing w/ Ilya Filipov
Transcript
- 0:00SPEAKER_00: 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_00: Let's welcome today's guest.
- 0:15SPEAKER_02: Hi there, I'm Jim Flynn from Onefire, along with my co-host, Leslie Castillo from the American Equity Underwriters.
- 0:22SPEAKER_02: Welcome to the IMCA Peer-to-Peer Podcast.
- 0:25SPEAKER_02: In today's episode, we will be exploring the nuts and bolts of implementing AI tools into your marketing strategy.
- 0:32SPEAKER_02: And to help us dig into this topic, we're joined today by Ilya Philipov, who leads Go-to-Market Strategy for North America for Cognizal and the Alliance Initiative for Amelia AI, the first provider of affirmative general liability insurance for AI models.
- 0:50SPEAKER_02: With over 20 years of experience across insurance and insure tech sectors, he has held leadership roles in sales, product, customer service, marketing strategy, underwriting, and digital transformation.
- 1:03SPEAKER_02: Prior to Amelia, he helped the go-to-market growth for insurance at Total Expert and Talk Desk.
- 1:10SPEAKER_02: He also spent more than a decade at leading PNC in life and health carriers, launching innovative marketing strategies.
- 1:18SPEAKER_02: Originally from North Methodonia, which is nowhere near Cleveland, Ohio, where he currently is from.
- 1:27SPEAKER_02: Ilya holds an MBA from the University of Notre Dame, Go Fighting Irish, and a BSC from the University of Arizona.
- 1:35SPEAKER_02: He's an avid long-distance runner and has completed eight marathons and counting.
- 1:42SPEAKER_04: Wow.
- 1:43SPEAKER_04: That's impressive.
- 1:44SPEAKER_02: Yeah, baby.
- 1:46SPEAKER_04: Okay.
- 1:46SPEAKER_04: Well, I detected uh before we started talking a distinctly Ohio accent.
- 1:54SPEAKER_04: So um I I I always wondered, you know, we've been on lots of calls together.
- 1:58SPEAKER_04: I always wondered where you were from, and I wasn't really sure.
- 2:00SPEAKER_04: So North Macedonia.
- 2:01SPEAKER_04: Am I saying that right?
- 2:02SPEAKER_04: Yeah, absolutely.
- 2:03SPEAKER_04: It's uh what brought you to this side of the pond?
- 2:09SPEAKER_04: I guess that's how you say it.
- 2:13SPEAKER_04: Yeah.
- 2:14unknown: Yeah.
- 2:14SPEAKER_04: What brought you all the way to Ohio?
- 2:16SPEAKER_01: I mean, short answer is a Delta airplane, but uh Fair enough.
- 2:22SPEAKER_01: Yeah.
- 2:23SPEAKER_01: And yeah, this was pre-9-11.
- 2:25SPEAKER_01: But uh I came to the US for college, and you know, typically we're gonna start one thing led to another, a job, two jobs later, grad school, and um I stumbled in insurance because it was my childhood dream to work in insurance in the US.
- 2:40SPEAKER_01: And uh is a mecha for insurance, so here we are.
- 2:44SPEAKER_01: That was just fun, well, no, absolute.
- 2:46SPEAKER_01: I don't know anyone that's dreamed to work in insurance.
- 2:48SPEAKER_04: I was like, is that a real thing?
- 2:49SPEAKER_04: The first person I've ever met.
- 2:51SPEAKER_01: Ever had.
- 2:52SPEAKER_01: Yeah, the first person I've ever met.
- 2:54SPEAKER_01: Although I have it, it's my life goal to meet the first person who's dreamed to work in insurance as a child, who's a five-year-old uh dreaming to let us know when you meet them and we'll have them on the on the pod.
- 3:09SPEAKER_02: We have known each other for years, and I did not realize that we shared uh marathoning in common.
- 3:14SPEAKER_02: Oh, I didn't know you were a fellow marathoner.
- 3:17SPEAKER_02: I I've I unfortunately I have an ankle injury, so I'm stopped, but I'm so I'm at five and stopped.
- 3:23SPEAKER_02: Uh but I just oh, I just love distance running.
- 3:26SPEAKER_02: Yeah.
- 3:27SPEAKER_01: One of my bucket list items is to run uh through your neighborhoods gym over in Chicago.
- 3:31SPEAKER_01: Marathon is Oh yeah.
- 3:33SPEAKER_01: That's a great that's a great marathon.
- 3:35SPEAKER_01: Oh yeah, it's one of the greatest.
- 3:36SPEAKER_02: Yeah, yeah, absolutely.
- 3:37SPEAKER_02: And it's a fast one, too.
- 3:38SPEAKER_02: Oh yeah.
- 3:39SPEAKER_02: No hills.
- 3:40SPEAKER_02: They like to break the world record there.
- 3:42SPEAKER_02: So yeah.
- 3:43SPEAKER_02: Yeah, absolutely.
- 3:44SPEAKER_02: So ChatGPT has been around 992 days as of today.
- 3:53SPEAKER_02: Wow.
- 3:54SPEAKER_02: Okay, so um, you know, it really launched um I I mean, certainly arguably lots of stuff was happening in AI before that, but from um, you know, a day-to-day people being involved in it, it really started the revolution.
- 4:09SPEAKER_02: You know, and you know, how have you seen the role of AI evolve in in the insurance marketing landscape really since then?
- 4:19SPEAKER_01: Oh boy, um yeah, it's been uh 992 days and a lot has changed.
- 4:24SPEAKER_01: So I mean there's been four big shifts.
- 4:26SPEAKER_01: So the first one is uh we went from where LLM AI was a toy to where it's a tool.
- 4:33SPEAKER_01: So we started with basic copy helpers, uh maybe replacing Google here and there and doing a research, and now um AI is inside other tools we already use, whether that's email suggesting how to write an effective copy, your CRM, your content management system, design.
- 4:52SPEAKER_01: So that way you can plan, produce, and measure all in one place.
- 4:57SPEAKER_01: Uh also uh the search changed the rules.
- 5:00SPEAKER_01: I mean, uh as marketers, we relied on paid ads a lot.
- 5:04SPEAKER_01: So uh right now, Google's AI overviews and AI mode answer more questions on the result page that people typically consumers would not be necessarily scrolling down to look at the web pages.
- 5:16SPEAKER_01: So it reduces the clicks to websites, where the role and the value that marketers bring is actually has been adopting with more intent-driven content and helpful on-site QA versus focusing as much on SEO as we have in the past.
- 5:33SPEAKER_01: So recent studies show that AI overviews are widely uh present uh and can uh cut click through to organic listings uh significantly.
- 5:43SPEAKER_01: In the next five years, I think that's going to be a significant uh significant platform shift for us as marketers.
- 5:50SPEAKER_01: Uh the third one is our great friends at compliance.
- 5:54SPEAKER_01: I mean, they grew teeth.
- 5:55SPEAKER_01: Compliance is becoming a little bit more relevant and they're getting a hold of AI.
- 6:00SPEAKER_01: So regulators now expect uh guardrails around the usage of AI, which uh is why uh working for a company like Armilla has been super exciting.
- 6:10SPEAKER_01: Uh the NAIC, for example, issued a model bulletin of how insurers can use AI.
- 6:16SPEAKER_01: And many states that have adopted are adopting it already.
- 6:20SPEAKER_01: New York has issued their own guidance, Colorado has specific rules for life insurers on using external data and models, but essentially compliance is getting a little bit more clear picture around the guardrails of how we can use AI in a safe way.
- 6:35SPEAKER_01: And last but not least, this is something that has uh becoming more and more relevant over the past uh several months, but we are already seeing some of the fears with AI also come true.
- 6:47SPEAKER_01: So there's a term AI psychosis, which is gaining of foothold.
- 6:51SPEAKER_01: That's where consumers are becoming a little bit more delusional and over-reliant on their AI chatbot.
- 6:57SPEAKER_01: I mean, we're already hearing stories about uh people dating their LLM like a chat GPT, thinking that it's their imaginative partner, and that recommending that hey, maybe you should make some certain life changes.
- 7:09SPEAKER_01: There are major lawsuits right now happening around that, and as funny as it is, it can become really problematic.
- 7:15SPEAKER_01: But uh that also has evolved the need of having guardrails in the deployment of AI uh models.
- 7:23SPEAKER_01: And companies, ChatGPT including, are starting to implement those guardrails so that AI doesn't necessarily uh go rogue and uh cause some stuff and that we have certain lawsuits already happening where uh an LLM model maybe went a little bit too rogue uh and changed people's lives a lot.
- 7:41SPEAKER_04: It is crazy to hear some of the ways that it's helping us but also hurting us.
- 7:47SPEAKER_04: Um so so yeah, I mean ChatGPT definitely brought AI, at least the way that we're we're defining AI today, you know, to the mainstream.
- 7:56SPEAKER_04: But um since then, you know, many tech companies have have really jumped into the fray.
- 8:00SPEAKER_04: There's this um website, it's a it's a website in a newsletter.
- 8:04SPEAKER_04: I think actually I think I learned about it at IMCA Ignite.
- 8:07SPEAKER_04: It's called There's an AI for that.
- 8:09SPEAKER_04: I don't know if you have you guys ever heard of this.
- 8:11SPEAKER_04: Anyway, it's got like literally thousands of AI tools.
- 8:14SPEAKER_04: It's super overwhelming.
- 8:15SPEAKER_04: And when you go to it, you almost get FOMO when you look at it.
- 8:19SPEAKER_04: Like which, you know, which tool is the best?
- 8:21SPEAKER_04: Like, yeah, what am I missing out on?
- 8:24SPEAKER_04: So I was personally really excited that we were gonna have this conversation with you, Ilya, so that you could tell us the best tools or the ones that you like.
- 8:32SPEAKER_04: Um, so anyway, we all end up with kind of our favorite tools.
- 8:36SPEAKER_04: What are what are yours right now?
- 8:37SPEAKER_01: Oh gosh, I mean it can become overwhelming to your point, Leslie.
- 8:40SPEAKER_01: It's uh there's a thousands and thousands of tools.
- 8:43SPEAKER_01: I mean, uh, and the uh exciting part is pretty much anyone can create a tool right now that's based off of ChatGPT and just call it an AI, stick an AI sticker.
- 8:52SPEAKER_01: And uh but uh so I would like to group them by the jobs to be done.
- 8:56SPEAKER_01: So uh let's just say when it comes to researching and outlining, I like to use perplexity and uh also chat GPT.
- 9:05SPEAKER_01: I mean, taking a first tab at creative briefs or like when you paired it with uh human fact checkers, it definitely saves you a few hours on uh developing basic documents or doing research.
- 9:16SPEAKER_01: ChatGPT has also uh released the deep research uh function where you can actually it will do the web scrubbing for you so that that that way you uh don't necessarily have to spend hours and hours uh researching a topic.
- 9:29SPEAKER_01: Uh when it comes for writing and editing, I mean Microsoft Copilot is the obvious tool because it's embedded in everyone's word.
- 9:35SPEAKER_01: Uh same with Google Gemini, they're both embedded in your office suite.
- 9:38SPEAKER_01: Then you have Grammarly for Quick Polish, um, which is also a basic tool.
- 9:43SPEAKER_01: Uh then uh when it comes to designing uh landing pages, uh, I mean Canva has AI magic features, you have Adobe Express Firefly for brand safe assets.
- 9:55SPEAKER_01: Uh the listeners of the podcast can pretty much go and uh open up a free account and start testing them.
- 10:01SPEAKER_01: Uh also for when it comes to audio video repurposing, uh, that's not necessarily a full disclaimer, my area of expertise, but I've kind of played for some private projects with Descript and uh Riverside for editing and clipping.
- 10:15SPEAKER_01: Then uh CRM automation, you have obviously HubSpot uh AI systems, you have sales for Zeppstein, uh Copilot that uh and uh lead an engagement scoring that we use at um at both Cognizal and Armila.
- 10:28SPEAKER_01: And then uh the few things uh uh you also for CRM and automation, you can do HubSpot CIS systems, you have sales for Einstein.
- 10:36SPEAKER_01: I also like to use a tool called Zapier, uh which can uh develop an agent and stitch everything.
- 10:42SPEAKER_01: So if you think about it, developing an executing marketing campaign and uh taking all the steps from uh developing creative brief, developing a strategy document.
- 10:50SPEAKER_01: And so for Zapier, you can actually combine all those steps uh with a very easy click-through uh environment and actually analyze the data, suggest the suggest the appropriate way, combine different LLMs and uh speed it out.
- 11:06SPEAKER_01: I would encourage everyone on this podcast, I mean, go try Zapier.
- 11:10SPEAKER_01: It's free, it's very easy to use.
- 11:12SPEAKER_01: You don't necessarily have to use your private data, you don't, but try it, it's a very powerful tool.
- 11:17SPEAKER_01: And one thing that uh for personal projects that I've started experimenting is a tool called Replit.
- 11:24SPEAKER_01: So I'm not a programmer at all.
- 11:28SPEAKER_01: Like I don't know how to code.
- 11:30SPEAKER_01: So what Replit essentially is the best way to describe it, it's the ChatGPT for coding.
- 11:35SPEAKER_01: So if you can imagine any app that you want to create, a mobile app, pretty much you can give a prompt to Replit like you would do to ChatGPT, and it would do the actual coding for you, the debugging, creating the database, the API integrations.
- 11:48SPEAKER_01: It's a very powerful tool, and it actually makes everyone um pretty much a programmer.
- 11:55SPEAKER_01: So those are the tools that I want to use.
- 11:57SPEAKER_01: Uh yeah, I highly encourage the listeners of the podcast, just go try it, experiment it because they are very powerful.
- 12:04SPEAKER_01: And I mean you you you come you can come up with ideas as it goes.
- 12:07SPEAKER_04: It's a good list.
- 12:08SPEAKER_01: No kidding, what a list.
- 12:11SPEAKER_01: And it's still only half a percent of everything out there.
- 12:13SPEAKER_01: Yeah, oh yeah, yeah, absolutely.
- 12:15SPEAKER_04: But those are your faves.
- 12:16SPEAKER_04: Okay, okay.
- 12:17SPEAKER_02: You know, some people, you know, jumped into all of this like headfirst, you know, like I'm all and and other people, you know, are afraid of it.
- 12:26SPEAKER_02: And I mean it's kind of a whole spectrum.
- 12:29SPEAKER_02: Yeah, but for those that you know that that have really embraced things and and started using different tools, can you tell us one thing that you see out there that you wish my marketers would stop doing with AI?
- 12:43SPEAKER_01: Good question.
- 12:43SPEAKER_01: Well, one thing, uh, there are a lot, but if I there's one thing and one thing only, stop publishing AI-generated content without human review.
- 12:55SPEAKER_01: That's I mean, whether that's chat GPT, and we talked about uh the AI psychosis, like whether that's chat GPT, Cloud, Gemini, it can definitely hallucinate.
- 13:08SPEAKER_01: And as humans, as marketers, we do add value to the table.
- 13:12SPEAKER_01: So do not publish anything that AI has done without having someone review it.
- 13:17SPEAKER_01: AI tools can whip up content really fast.
- 13:19SPEAKER_01: I mean, taking a first step at a document, which could have taken hours right now, can be done in seconds.
- 13:26SPEAKER_01: But without a human in the loop, you risk errors, you release a blind tone, or even worse, a blog post that once accidentally displayed, I cannot fulfill that request instead of usable copy, it can leave a bad impression and damage your credibility.
- 13:39SPEAKER_01: It can lead to losses.
- 13:41SPEAKER_01: So make sure that you do not publish anything that AI has created without reviewing it into details.
- 13:49SPEAKER_01: And at the end of the day, it's easier to edit than to uh to write.
- 13:53SPEAKER_01: So I would imagine uh it still would be a great value add, but do not publish anything without human review, if there's one thing.
- 14:01SPEAKER_01: I mean, there are other stuff as well, but that is the one key item.
- 14:04SPEAKER_04: Yeah, that's really good advice, especially uh in the insurance space.
- 14:08SPEAKER_04: So um I feel like that has kind of been a theme uh whenever we've had different conversations about AI, is just always make sure that you've got the human element incorporated.
- 14:18SPEAKER_04: Jim and I have had several conversations about AI with different, you know, from different vantage points.
- 14:25SPEAKER_04: And one another theme that has emerged a lot is that AI can it can lighten your mental load, right?
- 14:32SPEAKER_04: So it's not just uh developing content, it's it's it's lightening the mental load.
- 14:37SPEAKER_04: So for marketers, when I think about the mental load, you know, it would be developing marketing campaigns from soup to nuts, you know, developing automation workflows, that kind of thing, because my goodness, those things really do require a lot of brain power.
- 14:52SPEAKER_04: So with that in mind, what would it look like to use AI to do those things for you, like to build out a campaign or or lead nurturing automation?
- 15:01SPEAKER_04: Like what tools would you suggest for those things?
- 15:04SPEAKER_01: So Leslie, the first of all, we've got to remember the steps are still the same.
- 15:07SPEAKER_01: They wouldn't change necessarily.
- 15:09SPEAKER_01: You gotta first pick one audience plus one offer, uh, for example.
- 15:12SPEAKER_01: So let's just take cyber coverage for 52,500 employees for manufacturers.
- 15:17SPEAKER_01: Uh then you've got to draft the journey, and here AI can help.
- 15:21SPEAKER_01: Something like either whether that's a chat GPT or Gemini, they can draft the journey, do the initial research, and let's just say create a landing page plus three email nurtures plus sales handoffs.
- 15:34SPEAKER_01: AI can be embedded in all of those steps.
- 15:37SPEAKER_01: You can have whether that's Chat GPT, whether it's Gemini, whether it's Cloud, uh, develop those emails and uh develop the and then the sales handoffs can also happen with something like a Zapier.
- 15:48SPEAKER_01: Uh then uh AI would just save the time, it would write the first drafts, it would write uh generate the three ad variants, it would summarize maybe webinar clips into social posts, but at the end of the day, we have to still review them.
- 16:00SPEAKER_01: And last but not least, uh connect it with your CRM data to the extent that your IT and legal team feel comfortable, but uh the CRM data would help you actually segment the customers.
- 16:14SPEAKER_01: So uh in this example, the by industry, by the company size, by last action, or if you don't have a CRM by your take that your policy admin data or your agency management system data, which already hosts that uh information and can help you uh and AI can just dig in and help you analyze it.
- 16:32SPEAKER_01: And last but not least, like I mentioned, and I cannot overstate this, is have those guardrails, have um guardrails, have uh uh require a human review, it require uh for you to cite the resource the sources so that that way it doesn't necessarily make up information and don't necessarily put any uh personally identifiable information PII into public tools and um measure uh one winner metric, uh like let's just say that's quotes uh given, or uh or maybe uh first uh discovery calls if you're working for a brokerage and um go from there.
- 17:10SPEAKER_01: And the one good thing is you can ship this in days and not weeks, because AI will, I mean, pretty much will handle the first drafts, which is in my mind the most time consuming uh step and uh summaries while your team is tuning the messages and so forth.
- 17:25SPEAKER_02: Yeah, yeah, that that's a that's kind of a great overview.
- 17:28SPEAKER_02: You you mentioned segmentation as as part of that, and I know that's an area that uh some people struggle with.
- 17:36SPEAKER_02: Uh I think it uh for many it takes longer than at least we wish it would.
- 17:44SPEAKER_02: Can you tell us a little bit about you know what specific tools and techniques you recommend for that kind of work?
- 17:52SPEAKER_01: Yeah, so predictive AI is um moving already towards anticipating customer needs before they raise their hand.
- 17:59SPEAKER_01: So a few practical use cases that I've used in the past and currently using are that are working right now.
- 18:04SPEAKER_01: Uh first one would be uh if you think about lead engagement and scoring, there's a tool uh that's very commonly used in the tech world and it's taking a strong foothold in the insurance tech world as well.
- 18:18SPEAKER_01: It's called Sixth Sense or Demand Base, where they don't just tell you which leads are hot, but they also predict which companies are in the market for insurance products right now based on buying signals.
- 18:30SPEAKER_01: So they would uh track uh website visitors, what are they searching on the web, and that would help the sales and marketing team that identify a group of uh consumers that might be uh working in uh risk management or business owners, uh and they would that are searching for insurance, so they would already create an audience for you, and that way you already have something that's uh pre-qualified to target.
- 18:57SPEAKER_01: Uh then uh if it comes to another use case, churn and renewal risk detection, uh platforms like Retention X or ChernZero, they use AI uh spot subtle warning signs like the Chris Portal logins is a first uh example that comes to mind or slower responses to communications or claim chairs that would predict a customer that might not renew.
- 19:18SPEAKER_01: In today's hard market environment, that might be not as helpful because chances are if a customer can find who to insure it with, they're gonna renew very likely the carrier because it's uh difficult to switch carriers, uh, especially in a hard market environment.
- 19:33SPEAKER_01: But still, this tool could give marketers time to uh run targeted safe uh campaigns.
- 19:39SPEAKER_01: When it comes to uh content and offer personalization based on uh segments, uh tools such as Persado or Seismic Aura AI go beyond the subject line testing.
- 19:49SPEAKER_01: They can also predict which tone, emotion, or message will resonate with a given audience segment based on uh what you've uh found already, and then automatically tailor that content to them.
- 20:00SPEAKER_01: And uh then uh tools like Seventh Sense or iterable AI optimizer, they would do optimal timing and channel uh selection, and they could forecast uh the best moment and channel when to send an email, an SMS, a LinkedIn message, because uh a person, not every person would uh read a LinkedIn message at the same time.
- 20:22SPEAKER_01: So you want to stay at the top, you want to kind of segment it based on everyone's characteristics.
- 20:28SPEAKER_01: And when it comes to uh for insurance, this is a little bit uh avant-garde for lack of a better term, but they're product recommendation engines that are engines that are inspired by e-commerce.
- 20:40SPEAKER_01: Um, next best product tools like uh the Conveyor Relevance Cloud that predict cross-sell opportunities, like bundling cyber with ENO or life insurance alongside voluntary benefits.
- 20:52SPEAKER_01: Things that an agent or a marketer can do maybe on a case-by-case basis, but when you're talking about hundreds of millions of rec hundreds of thousands of millions of records, this tool can actually predict what is the next best product for every single of those million records that you're having based on the similar segment behaviors.
- 21:13SPEAKER_01: Wow.
- 21:14SPEAKER_02: I think I think ridiculously and I both said, wow, at the same time.
- 21:19SPEAKER_04: There's just so there's just so much, it's so overwhelming, it's so exciting and overwhelming all at the same time.
- 21:25SPEAKER_02: Yeah.
- 21:26SPEAKER_02: We'll take a quick break and come right back.
- 21:29SPEAKER_03: Are you looking to take your company's growth to the next level?
- 21:33SPEAKER_03: Look no further than OneFire, the RevOps-focused marketing agency that specializes in helping insurance companies grow.
- 21:40SPEAKER_03: Our expert team of strategists, marketers, and HubSpot specialists will help you optimize your marketing efforts and maximize your ROI.
- 21:48SPEAKER_03: From lead generation to conversion and retention, One Fire has you covered with our deep insurance industry expertise and modern approach to growth.
- 21:57SPEAKER_03: Don't let your competitors leave you in the dust.
- 22:00SPEAKER_03: Ignite your growth with Onefire.
- 22:02SPEAKER_03: Visit us today at onefire.com.
- 22:07SPEAKER_04: I'd like to spend a few minutes in talking and to talk about predictive modeling because that's something that comes up a lot in the AI conversation.
- 22:16SPEAKER_04: And I'd love I'd love to hear your perspective on how that can start to lighten the load for marketers as well.
- 22:24SPEAKER_04: So how do you see predictive modeling being used to enhance the development of marketing strategies?
- 22:31SPEAKER_04: And what specific tools do you think are the most effective?
- 22:36SPEAKER_01: Think practical predictions.
- 22:39SPEAKER_01: So the way I would characterize the word insurers, marketer would typically care about is things like lead engagement scoring to prioritize sales follow-up.
- 22:49SPEAKER_01: So a hotter lead obviously would have to uh potentially get more communication from their marketers and may and potentially from the producers as well, or uh or underwriters or wholesalers, or uh depending on what your sales model looks like.
- 23:06SPEAKER_01: Then churn and uh renewal risk uh to uh trigger um safe plays.
- 23:13SPEAKER_01: So like in a prior example, if a customer is likely to churn, potentially have a different communication.
- 23:19SPEAKER_01: Uh so that's an additional uh uh practical prediction that uh a marketer can implement.
- 23:25SPEAKER_01: And last but not least, uh send time and content suggestions, which are uh uh to lift email clicks uh for um for the targeted audience and making sure that you predict of when this person is most likely to respond and their tools like that.
- 23:39SPEAKER_01: So uh a marketer can start with built-in uh options.
- 23:42SPEAKER_01: So if a company is on HopSpot, there's HopSpot's predictive scoring, which is likelihood too close, then you have Salesforce Einstein Lead, which is still in their infancy stage when it comes to uh insurance, but they are evolving rapidly and they will be even getting more and more um stronger foothold.
- 24:01SPEAKER_01: Uh there wouldn't there's no data science required by using with using these tools.
- 24:05SPEAKER_01: I mean, I've used uh HopSpot's uh predictive scoring uh already, and uh in the past life I've also used Salesforce Einstein, and I'm not uh programmer by any chance.
- 24:14SPEAKER_01: Then uh another uh the tools that I mentioned in the previous question also uh would fit into this predictive modeling as well.
- 24:22SPEAKER_01: And a tool that I forgot to mention that I used when I was at Westfield and has significantly evolved is a tool called uh Resonate, and they also can help you identify when a customer is very a typical consumer is likely to open an email or respond to a phone call and an ad and so forth.
- 24:41SPEAKER_01: So so by uh having a practical prediction and know what are you using for, whether it's leading engagement score, insurance your risk, or send uh time and content suggestions, so you can actually start mapping each of these tools across that.
- 24:54SPEAKER_01: So I would start with what's uh what does your uh CRM or policy admin system have out of the box and even or AMS, and they all have certain tools, and then I would map them across that those three uh buckets, and then if there's something that's not been fulfilled, I'll start going to some of the other tools that I mentioned.
- 25:14SPEAKER_02: Let's let's talk about uh chatbots.
- 25:18SPEAKER_02: Chatbots and and conversational AI.
- 25:22SPEAKER_02: Um there's a lot of that out there.
- 25:24SPEAKER_02: There's a lot being launched.
- 25:26SPEAKER_02: It seems like you know, every time you have some need for customer support in almost any part of your life, you're going on a website and there's a new chatbot out there.
- 25:39SPEAKER_02: And uh some of them are uh are really great, some of them are wildly frustrating.
- 25:47SPEAKER_02: And um, I mean, it's talk about a turnoff from uh you know a customer standpoint.
- 25:52SPEAKER_02: I mean, I've had some of them that and you can't get out of them.
- 25:55SPEAKER_02: Those are the ones that drive me crazy.
- 25:57SPEAKER_02: You can't get out, I want an agent.
- 25:59SPEAKER_02: I just I need a a real person because you obviously don't understand what I want, you know.
- 26:04SPEAKER_02: Um so you know what what of those in that space, which of those are out there now that are achieving the the most success?
- 26:14SPEAKER_01: And chatbots are also, I mean, there's chatbots within the chatbots, and there's different uh groupings as well.
- 26:19SPEAKER_01: I mean, there's a voice chatbot, there's a text chatbot and so forth.
- 26:23SPEAKER_01: So uh when it comes to insurance, I've seen most success when the tools are simple and focused.
- 26:28SPEAKER_01: Get the customer to human quickly when they need it.
- 26:32SPEAKER_01: Insurance is a very, very specific industry.
- 26:37SPEAKER_01: It's not necessarily uh you you gotta know insurance in order to deploy chatbot.
- 26:43SPEAKER_01: So it has to have the workflows and the knowledge specified to insurance.
- 26:47SPEAKER_01: It also has to integrate with some insurance-specific systems.
- 26:50SPEAKER_01: I mean, let's just for well, I mentioned Salesforce and Hopspot, but the majority of insurers today are are still not necessarily using a CRM.
- 26:58SPEAKER_01: They have their guide wires, the uh they have the DuckCrigs, they have uh the applied epics, the AMS360s, uh Easy Links, you name it.
- 27:05SPEAKER_01: And uh you would have to integrate with those back-end systems as well.
- 27:09SPEAKER_01: And so what I've seen the most success is with again when these tools are simple and uh focus on insurance.
- 27:15SPEAKER_01: So when it comes from texting for claims and service, or even for new business, uh one tool that comes to mind and it got a strong foothold in the on the carrier side, is uh high Marley.
- 27:29SPEAKER_01: I mean, they're winning because they use plain SMS, they uh it's not another app or chatbot maze, they integrate with uh the majority of the systems that insurers use, and policers can I mean pretty much use it, text a photo uh damage, get real-time updates, and an insurance company can easily deploy it.
- 27:46SPEAKER_01: So high Marley, in my opinion, is uh definitely one of the top brands when it comes to insurance.
- 27:52SPEAKER_01: Then uh when it comes to uh small commercial quoting, I mean you have tools like Tarmica that can streamline multi-carrier quotes for agencies and they can definitely engage the customers that side that way.
- 28:03SPEAKER_01: Then you have uh when it comes to website chats for uh lead capture, I mean you have light payer assistants uh like Drift or similar that do well when they're limited to few but few tasks only.
- 28:14SPEAKER_01: So to your point, Jim, is they can't necessarily handle more complex stuff, but like answering FAQs, maybe booking meetings, uh routing to someone or to the right producer.
- 28:24SPEAKER_01: And recently we've also seen voice assistants take a hold, and there's a tool called Poly AI, and that's uh like talking to a person, poly AI, unlike the previous tools that I mentioned, is not necessarily insurance focused, but it's fairly easy to deploy within insurance.
- 28:43SPEAKER_01: And uh I'll also here put a shameless plug to a team that I used to work with, Talkdesk, that's also um has a state-of-the-art uh chatbots integrated with the majority of uh insurance uh backend systems, and they they are really leading in this space as well.
- 29:02SPEAKER_01: Nice.
- 29:02SPEAKER_04: That's a great list.
- 29:04SPEAKER_04: I think that um one of the things I'm taking away from this conversation is just like I love how you're kind of curating this for us.
- 29:10SPEAKER_04: Like there's so many tools out there, and you're telling us, you know, which ones you see working the best.
- 29:15SPEAKER_04: Um, throughout this conversation, you've mentioned HubSpot and Salesforce uh several times.
- 29:20SPEAKER_04: And um so I want to talk a little bit about those tools or tools like them.
- 29:26SPEAKER_04: I know they're kind of the the the two big ones, but in in recent years, or maybe even what's that?
- 29:31SPEAKER_04: The grill, yeah, for sure.
- 29:32SPEAKER_04: The grills.
- 29:34SPEAKER_04: Yeah.
- 29:34SPEAKER_04: Um so in recent years or recent months, we've started to see platforms like that that are that are already quite robust um start to add on AI functionality to make them even more powerful.
- 29:48SPEAKER_04: Do you recommend, and you you kind of alluded to this a minute ago, but I just I almost want to be really, really clear.
- 29:55SPEAKER_04: Do you recommend starting with a platform like that, like an all-in-one platform with AI already included?
- 30:04SPEAKER_04: Or do you think that marketers should sort of cherry pick the AI solutions that work best for their needs and then integrate those tools uh I guess into those softwares or or a different software altogether?
- 30:19SPEAKER_01: So, yeah.
- 30:20SPEAKER_01: So, and by the way, uh Leslie, uh, Salesforce and Hubso definitely do have AI plays, but some of the insurance-specific players like GuideWire, Verti4, Applied, they're also building similar AI plays, so they definitely also have it.
- 30:33SPEAKER_01: But whether it's insurance or any regulated industries, crawl walk run is a key method of deploying anything, whether it's AI or name it.
- 30:43SPEAKER_01: So I would start with crawl walk run, and in the crawl phase stage, I would turn on the AI tools that you already pay for.
- 30:50SPEAKER_01: So whether it's your HubSpot, whether that's your guide wire, whether it's your Salesforce.
- 30:55SPEAKER_01: I mean, Microsoft and Google is very easy because uh every office is already having a tool, has a uh has copilot in it or uh Google Gemini in it.
- 31:07SPEAKER_01: And uh start with those, turn them on, work with uh the legal and compliance team around setting the guardrails, the processes on how you can use that content.
- 31:17SPEAKER_01: Then in the walk phase, start adding point tools where you have you see having a gap.
- 31:23SPEAKER_01: So whether that's uh creating a video, descript for video or Canva for being fast for developing fast creative or something that's within the Salesforce GuideWire ecosystem, but identify the gaps and start adding them as pilots at a low cost level.
- 31:39SPEAKER_01: So uh I think uh don't quote me on this, but a basic license for Canva is like 25 bucks.
- 31:44SPEAKER_01: So start experimenting there once you see that there's a gap, but you gotta first use the tools and crawl that you already have paying for.
- 31:52SPEAKER_01: And then when it comes run, once you're starting to build the AI acumen and uh volumes are growing, and you're considering uh a customer data platform, start uh to uh connecting them through Museoft, Boomi, Zapier, and start connecting these tools with your back-end systems to unify the data, and there will be the run phase where you're getting something outside.
- 32:13SPEAKER_01: But in the crawling phase, where I would imagine 60% or more of our insurers are, start with whatever we have you currently are paying for, and then as you're getting comfortable on the start adding on other tools as well.
- 32:27SPEAKER_01: But don't necessarily don't necessarily have to go and buy the Cadillac right off the bat, even though I'm sure someone's going, a salesperson is gonna try to convince you to do so.
- 32:35SPEAKER_04: Well, I think there's it just feels very um logical, right, to to assume that if you buy the Cadillac, you're gonna get all the bells and whistles.
- 32:44SPEAKER_04: But the reality is, I guess what you're saying is you might not need all those bells and whistles, right?
- 32:48SPEAKER_04: So start with what you need.
- 32:51SPEAKER_04: Oh yeah.
- 32:51SPEAKER_04: And go from there.
- 32:52SPEAKER_04: Okay.
- 32:52SPEAKER_04: That's good advice.
- 32:53SPEAKER_01: You'd be surprised uh the brands that we mentioned and the CRM uh policy systems providers, how many are not using the AI features at this time?
- 33:01SPEAKER_01: I mean, and that's in my mind uh opportunity lost, but yes.
- 33:06SPEAKER_02: Well, yeah, that's uh that's really smart um looking at it from that perspective.
- 33:12SPEAKER_02: I mean, the other challenge that that not everyone, but darn close to every insurance organization has these siloed platforms, you know, uh the these systems, and it may be, you know, a home-built CRM or a policy administration system or something that underwriting has been bolting together for you know 25 years.
- 33:40SPEAKER_02: And it's hugely frustrating when you try to start pulling these things together.
- 33:45SPEAKER_02: How how is AI playing a role in integrating those tools?
- 33:50SPEAKER_02: And and which tools do you see having the highest degree of success to help pull those together?
- 33:56SPEAKER_01: That is a uh very specific insurance opportunity and a problem, having those back-end systems uh uh integrating, pulling the data.
- 34:05SPEAKER_01: Because if you don't pull that data out of there, I mean the AI tools or any tools are will be definitely a little bit more uh not as useful as they could be.
- 34:14SPEAKER_01: So there are a few things that I would mention.
- 34:17SPEAKER_01: One uh currently there are connectors like Mulusoft, Boomi, Workado, I mentioned Zapier that can integrate with some of the uh more common systems and typically on the CRM size, like Salesforce or Hubspot, out of the box and uh move data between a CRM system and the user.
- 34:35SPEAKER_01: Now we also are coming to but most insurers are also using agency management systems and policy admin systems that are not as easy to integrate with uh out of the box.
- 34:47SPEAKER_01: But those uh companies, whether that's uh Vertif4, whether that's GuideWire, whether that's Duck Creek or Applied, or you name it, they have an ecosystem of vendors that are already have invested in integrating with that back office and also uh being able to pull the data out.
- 35:06SPEAKER_01: Um, one thing that we've also used in the past is using uh a process called uh RPA, uh which stands for robotic processing and application.
- 35:17SPEAKER_01: Uh I it escapes what the A stands for, but it stands for something.
- 35:21SPEAKER_01: But uh there are tools out there that can also pull that information out of uh any system, and it can be a homegrown system and embed it into your marketing applications as well, and that's very AI, that's um AI based as well.
- 35:36SPEAKER_01: Uh Cognizol, the comp one of the companies that I work with, um that's what gets me excited.
- 35:42SPEAKER_01: When I was a Total Expert, we use Cognizole to uh integrate with a very difficult to integrate agency management system for a client.
- 35:52SPEAKER_01: And there are vendors at Cognizol.
- 35:54SPEAKER_01: I know Cognizole um shameless plug, but they I I have very them verified that they can execute that integration as well, they have the talent to do that.
- 36:03SPEAKER_01: Uh, but using a um an integration partner can also be very helpful.
- 36:09SPEAKER_01: But those are the tools, and at the end of the day, uh insurance marketers, anything that we invest in is going to be as useful as the data can use.
- 36:18SPEAKER_01: And uh unfortunately, to your point, Jim, we have systems that have been built from 2030 for when Jimmy Carter was a president.
- 36:27SPEAKER_01: I mean, there are certain systems that are kind of like that, and uh and those tools can definitely um can definitely uh help you uh pull out and integrate and uh plug it in.
- 36:39SPEAKER_02: We have a client we were working with that uh you know had all of the systems that were built in COBOL.
- 36:44SPEAKER_02: Oh, yeah.
- 36:45SPEAKER_02: You know, and the biggest problem they were having was all their COBOL uh programmers were retiring, you know, or dying because no one's done COBOL for a long time.
- 36:55SPEAKER_01: And no one's teaching it.
- 36:56SPEAKER_01: I mean, that's yeah, that is a very common problem.
- 36:59SPEAKER_01: There was a insurance company I was I used to work with as a client when I was at uh talk desk that had uh the system, and the system at the end had uh 78, and we are was this the 78th version, and they go, No, that's the year when the system was actually built.
- 37:16SPEAKER_01: So you can imagine it's wild.
- 37:20SPEAKER_04: The Soviet Union was around when uh as with most of our conversations about AI, there's so much to say.
- 37:33SPEAKER_04: I feel like we could probably talk for another hour about all this.
- 37:36SPEAKER_04: Um, you really are a wealth of knowledge.
- 37:38SPEAKER_04: As we start to wrap up this conversation, what is one AI experiment or use case that you would recommend every marketer try within the next, call it 90 days?
- 37:53SPEAKER_01: Oh gosh, Leslie, there are a lot of use cases, but when I thought about it, I was um the one key factor would be the friction points that as you're trying to implement, they will be there.
- 38:05SPEAKER_01: So you don't necessarily want to have a use case that will be very reliant on integrating with the system that was built in the 70s, or uh they will be very, very they would have a lot of uh non-marketing, non-uh-business insurance uh professionals trying to affect it because they would delay it and would take beyond 90 days.
- 38:24SPEAKER_01: So uh out of the many use cases that are there, the one thing that I would definitely recommend is uh which is a low-hanging fruit in my opinion, is using AI to upgrade your marketing content into multiple formats because for several reasons.
- 38:38SPEAKER_01: One, you don't have to integrate with your back-end data right off the bat.
- 38:41SPEAKER_01: So you don't necessarily have to go through IT demand prioritization and have it uh and sacrifice some other projects.
- 38:49SPEAKER_01: Uh, you would have to have your legal and companies team being involved.
- 38:52SPEAKER_01: You obviously have to have the marketing and creative team and potentially the sales team involved as well.
- 38:57SPEAKER_01: But uh using that to upgrade uh your content into multiple formats is really key.
- 39:02SPEAKER_01: So take one piece of content that as a marketer that you already have, like just take a blog post, a webinar, a product flyer, and use AI tools like Canva Magic or the script or reply-powered scripts to quickly turn it into um short LinkedIn posts, uh one-page client handout or 60-second video clip.
- 39:23SPEAKER_01: Basic license with these tools is I mean, it's less than 50 bucks.
- 39:26SPEAKER_01: So it's not going, you're not going to break the bank, so you're not necessarily going to submit a have to submit a business case for the executive committee to approve for a variance report.
- 39:36SPEAKER_01: You also uh I mean the few things that you have to do is probably embed your marketing brand guideline there, which is not necessarily a uh a privately held uh document, so you can do that and just start taking something you've already published, repurposing it, and seeing how it works.
- 39:54SPEAKER_01: Uh it's a low-cost, uh, high-benefit uh project, in my opinion, that you can learn a lot and then start slowly deploying and expanding beyond that.
- 40:04SPEAKER_04: I think we can handle that.
- 40:05SPEAKER_02: Yeah, that's that's really a good one.
- 40:08SPEAKER_02: Because, yeah, I mean that yeah, you already got some piece of content or many pieces of content that you can do that with.
- 40:15SPEAKER_02: I I like I really like that.
- 40:17SPEAKER_02: I'm I'm uh finding that I'm hyperventilating here because uh we have gone through a lot.
- 40:24SPEAKER_02: This is uh this is really great stuff.
- 40:26SPEAKER_02: Um if if our if our listeners want to connect with you and pick your brain a little bit more about that, what's the best way for them to do that?
- 40:35SPEAKER_01: The best way is probably LinkedIn.
- 40:37SPEAKER_01: Okay.
- 40:37SPEAKER_01: That one's constant always.
- 40:39SPEAKER_01: Uh I mean my photo there doesn't have a beard.
- 40:42SPEAKER_01: I've grown beard ever since, but everything else is pretty much uh consistent.
- 40:46SPEAKER_01: So uh uh you can add me.
- 40:48SPEAKER_01: It's uh my name is spelled it's I L Y A F I L I P O V.
- 40:52SPEAKER_01: Uh it's on the podcast.
- 40:53SPEAKER_01: Feel free to add me, and I'd love to connect and I mean just chat.
- 40:56SPEAKER_01: I'm very passionate about AI.
- 40:58SPEAKER_01: I know that it will drastically change our industry in a variety of ways in the next five years, and I'd love to discuss and help.
- 41:07SPEAKER_01: I mean, because uh as in the one thing I like about the IMCA, it's an organization that unites marketers, and everyone's here to help without any huge asks.
- 41:17SPEAKER_01: And I same here.
- 41:18SPEAKER_01: I mean, I'd love to help any way I can or provide advice or direction or connect you with anyone in my network.
- 41:25SPEAKER_01: But feel free to add me on LinkedIn and yeah, we can go from there.
- 41:29SPEAKER_02: Awesome.
- 41:30SPEAKER_02: That's uh that's a wrap on today's episode of the IMCA Peer-to-Peer Podcast.
- 41:36SPEAKER_02: Uh, we hope that this conversation, albeit uh rapid fire, sparked some new ideas and left you feeling inspired about implementing AI tools into your marketing strategy.
- 41:48SPEAKER_02: If you enjoyed today's episode, don't forget to subscribe, give us a five-star rating, and lead us a review.
- 41:56SPEAKER_02: Your feedback helps us reach more marketers and communication professionals who want to grow their careers and improve their work.
- 42:03SPEAKER_02: And if you would like to share your thoughts on topics you'd like us to cover in future episodes, shoot them over to us at ideas at IMCAPeer, the number two peer podcast.com.
- 42:16SPEAKER_02: Once again, this is Jim Flynn from OneFire, along with my co-host, Leslie Castillo from the American Equity Underwriters.
- 42:24SPEAKER_02: And our guest today, Ilya, has been a rock star amongst the AI team.
- 42:29SPEAKER_02: Thanks for having me.
- 42:30SPEAKER_02: Oh, thank you.
- 42:31SPEAKER_02: Thank you again for being part of today's episode of the IMCA Peer to Peer Podcast.
- 42:36SPEAKER_02: Thank you.
- 42:37SPEAKER_00: The Peer to Peer Podcast is a presentation of IMCA.
- 42:40SPEAKER_00: To learn more about today's episode and listen to other episodes, visit IMCANet.com.
- 42:46SPEAKER_00: This has been a production of OneFire.