Podcast: Play in new window | Download | Embed
Subscribe: RSS
Topics Mentioned
- Why Artificial Intelligence?
- What is Sentiment Analysis?
- Cleaning Data is Difficult
- Is Any of This Live Yet?
- Lessons Learned
Episode Transcript
Craig: I’m so excited to introduce my guest for this episode, Brian McLaughlin, CEO and founder of Redtail Technology. Brian, hey how’s it going?
Brian: Doing good man. How are you? Okay.
Craig: I am excellent. Welcome. Glad you can make it.
Brian: Yeah, yeah, I’m sorry we had to reschedule this a couple times. It’s a great topic we’re going to cover today and I’ve been looking forward to this.
Craig: We had to reschedule a couple of times because you were a little busy. A couple of things going on with your company.
Brian: Surprise!
Craig: It was a surprise.
Brian: It was it was probably the best kept secret out there.
Craig: People were telling me, didn’t you know? I said, no, they kept it a secret didn’t tell anybody.
Brian: Nope, nope. We kept it down to a really small group of people and took care of business. It was super excited to be able to announce our acquisition by Orion. It’s amazing. It’s gonna be a fun journey. I can’t wait.
Craig: I can’t wait to see what you guys do. So we’re not talking about that today. If you want to learn more about Orion, acquiring Redtail Technology, Google it and you’ll hit the press releases and articles and stuff on the trade mags. You can also go to a podcast that I did with my good friend Kristen Schmidt the day it was announced. And Brian, you liked that podcast?
Brian: I did. Good job, Craig. Yeah, I identified a couple things, just to clear some air real quick, you identified a couple things that we weren’t clear on. One is will you be able to get Redtail like August or September or October or whatever on its own standalone? Yes, we will be able to still do that. Is there gonna be some big pricing changes? No. What we’re really gonna be focusing on, Eric and I, are the integration opportunities in the synergies and really tying things together. And that’s a big part of what we’re gonna talk about today. One chunk of it is all the stuff we’ve been doing. I’ve been working on from an AI, ML experience from working on, I get access to like so much more data. If you’ve ever known me or listened to my podcast with you or anybody else, you’ll know, total nerd. I’ll try to go too over the heads and heard buddy but yeah, I love the access to the data. And I love what is going on the other half going forward to me really cool.
Why Artificial Intelligence?
Craig: Well, we’re nerds together, which is why we get along. And I think a lot of people listening to this podcasts are also nerds. So feel free to geek out as much as you want. Definitely want to go deep, so thanks for the kudos on the podcast, Kristen and I really enjoyed doing it. And we just like talking about this stuff. But this podcast is about artificial intelligence, machine learning, natural language processing, predictive analytics, because that’s what’s going on on our podcast this month. We’re talking to a lot of other vendors who are doing similar things, but I really wanted to kind of dive deep. Can you just give us a quick overview of, first of all, why you decided to get into machine learning and natural language processing for Redtail? What was the impetus for you to start looking into this?
Brian: Yeah, so fantastic question. So if we go back a few years now, I think it’s about four years ago when I said, Hey, we’re gonna start a program to start doing a strategy around AI, but really more machine learning and natural language processing those segments of AI. It was really advertised as, hey, we have a ton of data that we’re not really using as well or as intelligently as we could what can we extract out of advisor data to make more meaningful inferences to a client experience or to maybe for marketing purposes, maybe for next opportunities, stuff like that?
Brian: So we started digging in, and one thing that I said back then was we have billions of emails, and it just keeps growing over the last two years, even way higher than that. And so I started parting that as an opportunity to extract information. And we really started saying, hey, if we want to get into the ML space because financial services is not really there yet. It’s kind of in chatbots, a little bit here and there, you get the robo-advisor doing some stuff, but nothing on a large scale. And you have to take so long to ramp up this up so long to get to experience and or training and all the models learned. Now if I say hey, we’re gonna get started. This is gonna be a strategy for 10 years. And now we’re kind of looking at saying, Hey, what are we coming out of it? After three or four years of work, what have we seen? What have we learned? What did we not learn and what have we not learned?
Brian: It’s extremely tough to start there, it is not easy. And the reason I mean the technology itself is relatively as far as how to make it do something, how to have code do X, and say, Come in with this means right? The problem is you can end up with so many false positives and negatives that could skew a potential advice or suggestion wildly.
Brian: And you see this on YouTube, Netflix as well where, you know, looking at one video different than you normally do can change your entire space of recommendations so fast. But in financial services, we’re really really concerned because if I say hey, this person a great candidate for this account, or something like that, or maybe their risk tolerance was not aligned with their accounts and so forth. If I get that wrong, and tell you to do something, everybody’s liable and it just looks, it’s egg on their face, right? It’s just terrible. And it’s not an experience that we want. So really spending a ton of time on trying to identify and weed out the ends of either side of a false positives and negatives, right? Because we would run things through where we would say, hey, let’s do sentiment analysis, which I think it’s gonna be huge for advisors to understand who their detractors and who their influencers are in the business. System analysis really helps with that. But one or two wrong phrases if the models are not trained accurately for your type of speech, can throw it way off.
Craig: So hold on a second, can you describe what you mean by false positive and false negative?
Brian: A false positive is something like it’s almost tone was taken into place, you ever text somebody and it gets misread misread, mis-processed, miscommunicated?
Craig: No, that’s never happened to me. I don’t understand, what do you mean by that? Misread a text?
Brian: Yeah, exactly. So it’s the same way with natural language processing. If you just take the words in the order that they’re in, and when they get ranked, and when they get processed for sentiment, let’s say, you can skew a positivity way high or way low just by like using the word “rate”. Or what if you put an exclamation point, but you didn’t mean to, put a question mark, that influences what that means. So you have to look at a larger scale of data, a larger pool of data to really identify is this truly a positive or negative sentiment? Or most of the cases are neutrals.
Brian: You’re throwing in and out a lot of data and we’re looking for like spikes that Hey, okay, if this way over here a positivity reading you’re like No, wait, that was actually a negative sentiment. How do we adjust for that? How do we accommodate? And we’re gonna say that’s not we have a solid positive. Here’s where you really should almost play like a law of averages. We’re looking at a sentiment and we’re trying to get the accuracy as close as we can. Sentiment is always tough. Pulling out things like entities keywords, much, much simpler. And we should talk about some of that.
What is Sentiment Analysis?
Craig: So going back, you will talk about sentiment analysis. Can you explain what you mean by that?
Brian: Yes. So, sentiment analysis is basically the natural language processing for people to identify if somebody’s talking positively. neutrally or negatively about something. If I say oh, man, that was totally great. Well, we have our tone and our voice that helps us decide, well, that’s kind of flat, right? And so if I say, Hey, that was really great. Yeah, well, that’s pretty positive, Brian, but we don’t have that in text. So we have to use other things in the way the language is written and processed. To understand, is it positive or negative and explanations and punctuation and so forth help us identify that. But basically, what we’re looking for is things like net promoter score and NPS score. You know, they say 1 to 10. You keep the top two, you keep the bottom three availabilities, kind of toss out. Those are your influencers those are your detractors.
Brian: Well, we take sentiment from every conversation we ever had with your client, right? Because we can take that messaging or text messaging, even a letter and take that content. Can we determine whether this is a client that might churn, that might turn over and you lose, right? Because the top kind of negatively, you’re not seeing a lot of positivity, a lot of excitement. So if that’s the case, we should target that and say, maybe we should change the way we deliver our experience. Maybe there’s something else going on with that client that we should unravel and dig into, to make a better experience for them. And to have happy clients.
Brian: On the flip side, we can take the influencer side say these are people maybe I should introduce to other prospects. These are people who are raving fans, like at Redtail we always say, build raving fans. I think that’s core to basically any business really, you should all be building raving fans every day. So if we can take those influencers and know who they are automatically through the machine learning, we can build groups and say let’s target them and say Hey, can you introduce to some friends or I have some people that are maybe coming over into my business and we have some event going on? Can you please join? Because I don’t want to invite a detractor, i if you invited detracted today, it’s gonna go south. There’s a good chance of that. So that’s what sentiment analysis is gonna really help us do.
Cleaning Data is Difficult
Craig: So, all the things you’re doing now, I just want to roll back a little bit. I remember speaking to you about this the first time at the Envestnet conference, like in 2018, you were you were so excited, like, this is what we’re doing, it’s gonna be really cool. But you’re also talking about things how difficult it was to clean the data because things like “401k”, is it numbers, is it a part of is a part of a phone number? is it part of a an amount? Is it really the 401k account type? Can you talk about how difficult it is to clean the data and some of the things you learned when you’re trying to clean the data?
Brian: Especially when it came to acronyms or short phrases for things in the financial services industry with spell isle, I got 401k beneficiary IRA. Like that makes sense. I’ve been tagged with that object and say, This is a financial account we’re talking about here. But if you start shortening those things, are you 401k, 401(K), 401-K. There’s a million variations on this. We get rid of all the actual characters, and we just squish them together and say 401k, but what does that mean in context? You’re right. What if I said, Hey, I need $401k to buy this house. Well, we’re not talking about my retirement account. We’re now talking about cash dollars. And so we’ve actually had to build a whole linguistic set that said, these type of phrases, if arounds other words, if it’s this it means this. So we translate it into a normalized state to say any variation of the 401k’s we just talked about, when we identify properly will tag as a 401k is what we’re talking about. I don’t really care about that. You said 401k, I care about if that’s the conversation is involving the retirement account. That’s what matters. Because that piece of data if we know you’re talking about that, I can start grouping things together. I can start looking for other pieces of data somewhere else, whether it’s documents or another connection somewhere else, or prepping email for our client, knowing we’re talking about their retirement account. That’s what’s so tricky.
Brian: And there’s a million other variations of that. People always say like, or how about your Roth. Well, that’s great, but sometimes Roth, a lot of people capitalize the R. It could be a last name. Are we talking about a Roth [IRA account], or we’re talking about Roth? That’s really confusing, right? There’s a lot of those nuances that we have to keep working on. But we’re getting better at that.
Brian: What it is, you train these models to say when I see these patterns, I most likely mean x. So if I say I say a string that says, I’m looking to invest into my IRA, okay, we’re talking about a retirement account, Individual Retirement Account investing, we get a couple of those key phrases we know we’re talking about a retirement financial account, right? Or if we say, Hey, I’m looking to withdraw money from an IRA. Okay, now we have an action associated with the withdrawal. We have one to invest, and we have a withdrawal, we can tie these together and they’ll start doing the next step of all this, which is building the data segment say, they’re looking for an action on something.
Brian: So you think about it, you go, hey, they want to withdraw from their IRA that can start a workflow. Just that one message. So that’s really where I’m trying to get this is that down the road, I want to automatically bubble it up and make an action item for the advisor or a staff and say, they just sent this email in 30 seconds ago. It says they want to withdraw from the IRA for XYZ maybe it’s automatically added the workflow, started it, put that content in there, and now it’s actionable in real time. That would be huge. I think we’re gonna see a lot of that expand into our Speak platform because texting with advisors is more and more ubiquitous in how we work for advisors, right? It’s going to become the standard that we can start talking via texts and doing things. That is the same message comes in, why can I bubble up write to the advisor in the text and say, Sure, cool. You’re talking about this account or this account and help them actualize?
Craig: Or while they’re texting sort of like we’ve used to leave memos to people like memo to my assistant, please do this, in your texting. The client can say I want to invest in my 401k, okay. We will take care of that and they will know that means to kick off an action.
Brian: Exactly. So even if it is something I still believe it should be consensual action. So it should be something that the advisor on staff says execute, right? Because suddenly, we can prep in advance like when we go back to those workflows and stuff. And you said, Hey, you have a IRA withdrawal workflow for a client, a couple of steps to verify things which account all the steps, we can start all that without consent. And then when they actually start working on workflow that becomes a consent item. So there’s some interesting twist on that too. Because I think the one of the big fears I have and I think a lot of advisors would have is machine learning AI gets more and more built into the solutions, is that we don’t want to take action without prior knowledge. I don’t want to be surprised, ever. Like I love data for give me the data early. And let me make a decision. But maybe prompt moves and gestures. That would help a lot. That would really save me a lot of time and especially the things that you know, the brain works so differently is so powerful that any computer that if you just glance at, you will already know your head already knows, oh, I want to do this. Well, if the software already prompts you with two or three choices, and one of them is going to be that choice, one click, you’re done and execute that.
Craig: it’s like I was at the Envestnet conference this past week in Charlotte, and Bill Crager was on stage and he was reminding us of Guy Kasparov, who was at a previous conference talking about the hybrid, machines and humans are way more powerful than either.
Brian: Yes. 100%. 1000%. The human brain is a masterpiece of engineering.
Craig: It’s wetware.
Brian: Yeah, totally. Wetware, there we go.
Craig: We’re just moist computers.
Brian: Yeah, it’s phenomenal and the processing power, computers can’t do it, computers just don’t have the nuance. We’re programming the nuances into it to kind of understand what humans mean. You have to get to quantum computing, really advanced stuff way beyond financial services for sure to get into that type of AI. But if you combined the two and put some intelligence like that human brain power with some automated processing, that’s a huge win still.
Craig: And that’s where we’re talking about the suggestions where it doesn’t take action, but have the AI/ML, bubble up what it thinks are the top three suggestions and let the advisor pick.
Brian: Yes, exactly. So and then enable that the job of us from a software side is to make it that that one suggestion when they hit that it’s automatic, it just does a whole bunch of other series of steps behind the scenes. That’s going to really improve efficiency and effectiveness for advisors. Because then you can support 10 times as many clients
Craig: Right. But you should be able to do that, Redtail is full of workflows. It’s a workflow engine, right?
Brian: Yes, but we don’t know which workflow to start for somebody automatically. We can’t suggest the workflow.
Craig: So I’m saying you can link them together once you can suggest then you can go right into Redtail’s predefined workflows.
Brian: Yep, exactly.
Is Any of This Live Yet?
Craig: So is any of this live yet?
Brian: Only in test cases, nope. There’s nothing live. We’ve done a few betas here and there with clients, but we haven’t actually pushed it out production and it’s not ready yet. It’s getting closer and closer. I think what we’re gonna see us doing more on the components of AI was more than natural language processing and machine learning. I think a couple things one, extraction of data from documents. So we have our new Redtail imaging coming out here next month or so. And one of the key tenants of that is to extract data out of those documents to identify fields of information that may be like for example, the use case of a building form, is you upload a client document, let’s say every application for Redtail. As you upload that you find there’s a mismatch between your CRM client data and the document data that should be popped up and flagged to you.
Brian: I mean, it could be under the account data is right on the form but that could mean that your CRM is out of date. If we can make sure that you’re aware of those type of inconsistencies, that’s where we can really apply some machine learning and do some really cool stuff. So you’re gonna see that starting to pop up, extraction of data, and then really automate more paperwork, or get rid of it, I hate paperwork. So it should all be digital, I think in my world. And if we can get rid of some of that and say, Hey, I did upload a doc that had XYZ in it, maybe a trust document or a will or something like that. Why is that not checking up and notifying other people or other systems saying cool, they’ve uploaded that, completed more of a profile. There’s so many things we can do on that front.
Brian: But right now, the way that the world is, we scan a document or take a picture on our phone or whatever, attach it to our books or records, and we have the main lead tag everything down the line, and you get anything wrong, that’s a mistake, go fix it. It’s just time consuming. So the goal is to eliminate all of that 100% and say, we know exactly what type of document this is. And it’s cool, most documents like that have a look and feel to them all. They’re always consistent or relatively consistent.
Craig: Relatively consistent is the key.
Brian: Sometimes you get the big bar I look for example here at my desk, but sometimes you scan something in a big bar and that’s a screwed up all your process the opportunity. Sometimes it just that flat out mistakes, sometimes handwritten notes, just don’t convert, you know? All those type of things happen to.
Lessons Learned
Craig: Oh, sure. So you’ve been doing this for four years, approximately. And so what are some of the biggest takeaways, what are the things you’ve learned? In these four years of going through the data trying to make it work, trial and error, the billions and billions of pieces of data you have
Brian: Massive amounts of data processed. It’s interesting because we anonymize everything. So for one, we try and bubble up like you’re dealing with 20,000 firms, and we’re trying to normalize the data across the board. So when we’re talking about financial planning, and we’re talking about budgeting for client or something, that means the same across all 20,000 firms, and that’s really, really tricky to do. We have this common sentiments of the ways people talk. There’s common patterns people talk. There’s common patterns around order of events that people generally follow in their lives. And so taking all this information with a prophecy and tagging it has been the biggest effort we’ve had to do.
Brian: So there was the linguistics part we talked about whether it’s identifying financial terms, that was a big project. Because you run it through and you realize, I got most of them. And then here’s 1000 more that just popped up when we scanned the rest of the data. And so that anonymization and that kind of normalizes it. That’s been the biggest learning curve, I think more than anything else. The model training, I believe, I’ve probably done it 100 times at this stage. I’m exaggerating a bit but we go through and we try to model the sentences and the way people talk. It’s really just tricky. We’ve even tried some big automated solutions. We’ve done the AWS, the Google Cloud, using a third party solutions. We were shocked at how disparate it was from our workload was doing so theirs is just so generalized for the world, that to really get into a financial services lingo. It requires us to do the same amount of effort and just apply it to their stuff. So we just keep doing it internally.
Craig: Wait hold on a second. Can you talk about processing and tagging? What does it mean when you say tagging? What are you tagging? What is the tag?
Brian: What we’re tagging is what a word or phrase means. You’re putting a classification on a word or phrase, right. So therefore, that 401k example, it’s still a great example, 401k is it a dollar amount or is that a financial account? What are we talking about here? If we say, Poland, are we talking about the country? If we say a Polish, are we talking about the language or the people, which? That matters. So you have to take a little more context and you would tag and say cool, these are geopolitical entities. These are corporations. These are people’s full names. Like the names have been actually another one. Another big struggle of ours identifying names, because most are pretty straightforward. But when you get into initials, or shorthand nicknames and stuff like that, you can really screw that up.
Craig: Or names that match other things.
Brian: Or doesn’t match at all. We’ve had the word my name Brian pops up as a country sometimes. It’s has to do with the way that it’s used in the sentence. So sometimes, not necessarily a country but a geopolitical entity, and be like, Oh, Brian, like an organization. No it’s not.
Craig: It also depends how they pronounce it.
Brian: My name is always misspelled, Brain, people always do it, just messed up everything. It’s, you know, you gotta really just extrapolate what the real meaning is, and not specifically the details much. The tagging that we’re talking about is us going through saying, the computer doesn’t realize what the sentence says. What’s it referring to? Is it positive or negative sentiment those type of things, means this. Now we can use it for future learning.
Craig: Interesting, so we’ve got the the normalization of data ensuring terms mean the same thing common speech patterns. What about identification of signals? So how hard is that to do and what are some things you learn when you’re trying to identify signals in all this data?
Brian: This was an interesting one. This one is something that I was really passionate and starting early on which was identified a signal basically, can you determine a future action based on some content, right? Interactions with people. So how can we identify and tell you that the client is ready to do college planning? I mean, there’s the obvious, ask the client, call and say, Hey, you ready? No. Okay, I’ll check back in a quarter. You also may have just raw data, you may have things like the kids date of birth. Okay, cool. Well, they’re about college age. It’s about time, right? Well, in all your communication, your texting, your messaging, everything else your meeting notes, is all the information actually that identifies when it should be. Right, when’s a good time to address a topic for the client?
Brian: And so the signal identification is actually really interesting, because it’s actually it’s not as hard as some of the other things we’ve talked about. It’s actually something you do based on history, like, Okay, we have some basic stats of a client relationship, husband, wife, partners, or whatever it might be. We have their date of birth, we have their net worth, we have all these factors. What are they talking about now? As what’s important, what have you been plugging into your meeting notes that may be identified as important to them? And they may mention three, four years in advance. Oh, yeah. It’d be awesome if my kid went to Oregon University, right. Okay, cool. Well, that’s still way too far. I’ll start talking about that. But you can tag that back and say Oregon University three years cool. Start doing the calculations. And the algorithm automatically identifies and says, a potential times in three years from now, so I’ll be ready to tell you tin three years. You should talk to your client about their kids called where and how much you can do all these kind of crazy cool things in advance automatically with intelligence.
Craig: Or you might catch an email where the clients talking about their child plays soccer. And you might say, well, maybe they’ll get a scholarship or maybe they’re interested in this type of so you could be looking to that.
Brian: Yeah. Or how about even simpler things? I mean, what if it’s like, in Sacramento here we have a AAA baseball team. What if I knew I had a couple of tickets? Who are the clients that actually would appreciate that? Sure. I can call my top client and offer it, but what if my top clients like, I don’t care about baseball, I like basketball. What are you talking about? What if you give it to somebody that actually was in something meaningful to them. Right? That’s a huge win for you, from your business standpoint, if you get the right persons to the right spot. So you can use information for that as well.
Craig: Like when you and I talk you always talking about we should go skiing or snowboarding. Because you know that’s what I like.
Brian: Yep, I mean, can we do next winter snowmobiling?
Craig: We could do that. Yeah. Anything cold, anything cold weather related I’m good. Good. So while we got tests, we have identification of signals. And was there anything else was there anything else I missed that you that you learned in this whole process that you wanted to share?
Brian: Think that chatbots are just too soon. I think we’re gonna see a rise of those again.
Craig: Well, I hate chat bots. Stupid chat bots. When are we gonna get smart chat bots?
Brian: They’re getting there though, right? I mean, I’ve seen some applications. It’s starting to get there where it actually can do the actual for you. I mean, even the Amazon stuff for refunds, and all that being better and better every day. I think when you’ve seen the chat bots automatically connect you intelligently to a human that can solve your problem if you need it. That’s that’d be really cool. But I think chatbots are just a tad early. Same with robo-advisors. I mean, it’s just the technology was just too early. It came out hard and strong, but it had a lot of flaws and batches. And as we learn more and more about client relationships, we can actually do insightful, meaningful chatbots that feel almost human. I mean, it’s can be kind of creepy. I remember the story, I mean, it was probably about four years ago when Facebook turned on with a chatbot that they built and they started arguing and talking in their own language. I mean, this idea is still in the back of everybody’s head, right? This is gonna take over the world but that’s never been the intent.
Craig: There’s the also the Facebook chat bot that after a day of running out in the public became racist.
Brian: Oh, my gosh, I forgot about that. But we’re not computers, obviously.
Craig: Yeah, one thing with chatbots. I want to say that there was I can’t remember which vendor had a chat bot It was either my cable company or the phone company. Sorry, it was interactive voice response. So you can tell it’s a computer and it says hold on, I’m looking it up for you and then we’ll go to I know your computer you’re not typing anything. Why are you doing that?
Brian: Just trying to build that bridge of connection with people.
Craig: It’s trying to fool me to think that it’s really a human typing on a keyboard.
Brian: Really? Not at all. Yeah, it’s just a soundbite.
Craig: You can go a little too far with that stuff. Awesome. So what can we expect in the future? Is anything on the roadmap?coming soon around AI, and machine learning, NLP that’s feeding other other new things from besides the Redtail imaging?
Brian: So Redtail imaging, we’re gonna focus on Redtail Speak as well, while looking at what we can do for that real time comps, and that’s where the Chatbot kind of comes back a little bit. I think it’s going to still be the one sided chatbot right now, where it says to the advisor to the staff, which is saying suggesting actions I can do real quick, based on the testing going on, rather than your client taxes automated response like that. Because I don’t think I know we’re not there yet. I also know the clients are necessarily there yet.
Brian: Other things I want to see are when should we do something is bringing up that informed decision making right? We’re starting to see some of our integration partners, just using our data to identify things like for example, latter life integration, that will identify people who maybe are targets for potential churn, and they can automatically create an opportunity back to the system. Well, take them to the next step will be actually connecting those dots will take one click application done, and it’s ready to go you’ve already opened it up and then attach to your client accounts and everything’s ready to go through latter life. Like that’d be really cool. Actually, that’s halfway there. It is already cool.
Brian: I think we’re also gonna see a lot of this pop up more and more risks based on the planning space. Adaptively saying, hey, while I write down markets crashing and we’re recording this right. Now it’s up today. Okay, but it’s been a rough couple of cycles, right? And so now we gotta apply that to the risk to planning and everything else auto adjust everything. In the moment, especially if you start adding to your CRM data that says, Okay, this client is kind of a nervous Nellie over here, we should be acting on it really fast. That can be a huge win as well. That’s where machine learning and AI can bubble that up and say, These are the people you really got to care about bearing a high low whatever crisis might be going on.
Craig: Yeah, and that’s what it’s all about. It’s a huge wins you want it you want to see these wins build up over time, so that advisors and clients feel comfortable with the recommendations.
Brian: And it’s nearly invisible to them, right, it should be just seamless as a part of their their life flow, whatever their workflow in their life, just like it just happens. It’s not something magical that’s to be just turned on. It’s just intuitive. And that’s I think, where it will win and finally come out, we are so close. I mean, we obviously have it with Alexa and stuff like that. We just got to get out into the actual business portion of our lives and put it into the firm’s and get that going. Once we do sky’s the limit.
Craig: We need it to be ubiquitous, intuitive and seamless.
Brian: Yes. It’s got to be.
Craig: Great. It’s got Brian we have reached the end I really appreciate your time.
Brian: What, already? Oh no, no, no, we got at least another hour.
Craig: We should do like a Joe Rogan, a three-hour interview,
Brian: A three-hour podcast? If that’s the case, but I’m gonna need some bourbon and maybe some outdoor seating.
Craig: Alright, well, schedule that.
Brian: Do it. Come join me. We’re gonna hang out on the deck and talk all things tech.
Craig: Done. I’m gonna book that. Don’t think I won’t. Cool So Brian. So I want to say everyone who wants to learn more about Redtail please go to RedtailTechnology.com and sign up if you’re an advisor and learn more about what’s going on. And I really appreciate you being here and sharing with us Brian, so much. I’m looking forward to seeing you again soon. We we saw each other at T3 two weeks ago and in person for like five minutes and hopefully we’ll bump into each other again somewhere.
Brian: Yep, will I see you at Wealth Stack.
Craig: No, not that one.
Brian: Well maybe I’ll see some of your listeners at Wealth Stack.
Craig: You’ll see people will listen to the podcast, which is hopefully everybody.
Brian: I’m excited. I’m joining a great Think Tank session on diversity and inclusion. Gonna be really awesome.
Craig: Excellent. So anyone who’s going to Wealth Stack, please check out Brian’s panel. Brian, thanks so much, man. Appreciate it.
Brian: Thanks, Craig.