What Businesses Get Wrong About AI (and How to Fix It)

[: 00:00:00

[00:00:26] I am Betsy Cerulo, your host, and welcome to my guest today, Gabriel Legend, CEO of Visionaire it. Today we are talking about how to wisely use ai. From a business standpoint and understand what risks may present themselves. So pull up a chair, enjoy your favorite beverage, and let's get started. Gabriel, welcome.

[: 00:00:51

[00:00:56] Gavriel Legynd: me.

[: 00:00:56

[00:01:20] So I know that we're really only gonna scratch the, uh, tip of the iceberg today.

[: 00:01:25

[00:01:39] Right. You know, one of the biggest things that I see leaders trying to do everything at once. They wanna do all the things. And I guess because AI is the new shiny thing, they wanna figure it out. Right. And I, and I get that right, but. They get very excited about AI and then all of a sudden wanna implement it across the whole [00:02:00] entire organization.

[: 00:02:01

[00:02:24] To see what happens, your successes, right, and then measure the impact, not just the hype, you know? Then it iterate and improve upon where you started. Once you've got all that dialed in, then you're gonna move to the next area. We actually normally start most of our clients out with focusing in on their marketing workflows because it's a great place to learn without breaking anything critical within the organization.

[: 00:02:54

[00:03:18] But here's the thing that's really important, is you've got to design with the end in mind. We, we tell everybody that design that's mapped this whole entire thing out, and design with the end in mind from the start. So before you start implementing anything, really think through the data sources. What integrations are you going to need to integrate?

[: 00:03:40

[00:04:07] Is AI gonna take over my job and all of that? We don't wanna do that. So the other piece of this is making sure, now that you have your team on board, they understand that we are moving forward, that you now have a person. That is going to be your champion in all of this. You gotta find somebody in the organization who's naturally curious, understands the technology, understands our game plan, where we're going, 'cause that's gonna help the whole team really get on board as well.

[: 00:04:39

[00:05:05] Betsy Cerulo: you know, from, um, from my organization.

[: 00:05:08

[00:05:38] So I, you know, I certainly hear a lot of. Feedback, pushback about is it gonna replace my job now, you know, in the, in the work of human resources, it's human resources. So you can't throw everything out and think that AI is going to be this, [00:06:00] um, you know, this shiny new car that's gonna save everything.

[: 00:06:04

[00:06:22] Gavriel Legynd: So again, it goes back to people feeling reassured that from a leadership perspective, that you're giving them peace and confidence and that.

[: 00:06:37

[00:07:02] And it sounds like within your organization, you understand that you're in the game of relationships. Relationships, absolutely. With your, with your employees, your team members relationships. With your customers and then your employees, relationships with customers, and creating this ecosystem of understanding that we all need each other, but we also need to leverage technology so that we can go faster.

[: 00:07:28

[00:07:49] Betsy Cerulo: And, you know, that's a good point when you said AI as a copilot, because from what I see internally with, uh, different, different programs that we have, [00:08:00] AI is an add-on.

[: 00:08:03

[00:08:26] Gavriel Legynd: right?

[: 00:08:27

[00:08:50] The workforce is changing, um, as it should, you know, it's, it's innovation. Um, and, and again, I think it's a real opportunity [00:09:00] for all of us to. Continue to grow and expand our, our companies. And there's so many wonderful tools out here. I don't tend to get into talking a whole lot about specific tools, but there's a lot of things, there's a lot of shiny things that are available to us, and so I, I think it's a, a great time for us to start investing in our people by providing them with the best co-pilots.

[: 00:09:27

[00:09:32] Betsy Cerulo: Yep. I couldn't agree with you more. So, when adopting ai, what security risk or compliance pitfalls should businesses be most aware of?

[: 00:09:43

[00:09:53] Everyone's trying to move faster again, doing more with less, and they're just not thinking through [00:10:00] what these AI tools actually mean in their business. So I bet you if you walked through some offices right now, you'd find people, unfortunately copying customer data into chat, GPD. To get help with emails, we're throwing financial projections to analyze the data and to get quick summaries.

[: 00:10:21

[00:10:47] Your proprietary information is basically walking out the door and you're opening yourself up to data breaches. Your brand reputation is on the line. Your competitors. You know, [00:11:00] have access to some of this information, but most companies have absolutely no idea what's actually happening with their people and how they're interacting with these tools.

[: 00:11:09

[00:11:30] Right? But I'm talking about using something that, uh, can see when someone's about to share sensitive data and stop it in real time, automatically redacting what's. What should be protected and giving your security team actual visibility into what's happening across all of these different AI tools that are available at the browser level, browser level, and even um, custom built, right?

[: 00:11:58

[00:12:32] Um, and, and, and that's it.

[: 00:12:35

[00:13:10] So is, is there any kind of data that shows. What's, you know, a percentage of what's accurate and what might not not be accurate?

[: 00:13:22

[00:13:35] They're, they're, they're training on a lot of this, the, the data. Right? And so the thing is, is yes. AI is going to hallucinate, perhaps it is going to tell you what you want to hear in you understanding more about how to use these tools. It is about learning how to best prompt, right? So as you're building out a [00:14:00] prompt and you're asking it some type of question, um, that you know you need to have facts behind it.

[: 00:14:09

[00:14:35] Right, but now it's aggregating that data a lot easier for you so that as you do the prompt, it's giving you the, the output, but it's also giving you the source. You can go check it and say, yes, no, this makes sense, or perhaps it took it out of context. Those things do happen.

[: 00:14:53

[00:15:01] You know, it came up really quick. Let me go over to this website that I know is as trusted as it can be to see if it's the same. So I would see maybe what looked good, what, what didn't feel right. Um, but you know, then I wanna ask the question 'cause we see. We see lately that there's lots of, um, what's the best way to put it?

[: 00:15:26

[00:15:52] Does AI pick that up or is it extracting it, even if it's a website? Uh, [00:16:00] what you would think is a, is a legitimate website or reliable? Does it pick but the data's not accurate? Does it pick, still pick it up if the accurate, if the data is inaccurate, of course.

[: 00:16:13

[00:16:26] I think there's a ton of, there's a ton of websites that, again, again, we, we think the data is correct, but it, it isn't. And so these models, Chad, CT, quad, et cetera, are going out here and searching for. Data and then now you have to wonder, well, where did it really come from? Again, validating that source because if it rep references a website that you know is not reputable, then of course now it's, it's not right.

[: 00:16:58

[00:17:10] Betsy Cerulo: Mm-hmm.

[: 00:17:11

[00:17:21] Um, you know, I don't know where should I send my college? My, my, my student to college that wants to major in biology. It must be a top 10 school, whatever, right? You understand where I'm going with, but if you're, you're asking it to go out there and search and understand more about, uh, what school is better than another school and things like that.

[: 00:17:41

[00:18:02] Who knows right now that model, sorry. Now chat. GPT is gonna pull back that information and now you can go and verify and again, check the information to validate it.

[: 00:18:12

[00:18:28] Is there also a way that teachers are able to find out, is this, this sounds good, it's it, you know, it looks like a good paper, but is, is there, um. I guess are there, is there software out there or something that, lets say a teacher know if their student is basically plagiarizing and pulling it right off a chat, GPT or, or, or another resource?

[: 00:18:58

[00:19:53] Again, it is a tool I think that as, as part of children or, or [00:20:00] young adults growing up in today's society and us wanting to make sure that they're prepared for the workforce of the future, that they understand how to use these tools and that it is not used in a way that they, um, are not. Adding their own energy, their own voice, their own thoughts to it.

[: 00:20:24

[00:21:04] Betsy Cerulo: Well, you know, that's, that's a lead in to my next question, you know, in business. So how can companies, and you've alluded to some of it, mm-hmm. How can companies train their teams to use AI responsibly and confidently without fear of replacement?

[: 00:21:19

[00:21:34] So what I always tell leaders is to think about, um, AI is again being their their copilot, right? It's your partner. That's it. It's not about replacing anyone, it's about making sure everybody's able to. Move the, the needle forward and improve productivity. The companies that are really winning right now are the ones that are using AI to drive sales, marketing, cutting down on mistakes, speeding up [00:22:00] things, but the humans are still running the show.

[: 00:22:03

[00:22:26] So now it really does shift their mindsets. Instead of getting people scared, they're actually excited. So when you bring your team to the conversation early and often, where they're able to see where AI can actually improve their job satisfaction. So that's, that is possible. Mm-hmm. So perhaps it takes over that boring thing that they didn't wanna do, or the repetitive stuff that they, they they just didn't like.

[: 00:22:57

[00:23:19] And that's the stuff that humans do well. And so we should be encouraging our team to continue to do those things. But the biggest thing here is, is just being honest about where you're trying to accomplish, what you're trying to accomplish with your team. Getting their buy-in and commitment early, and if you tell people you wanna train them up instead of replacing, we're training you.

[: 00:23:38

[00:24:03] And so nobody's shuffling and, you know, slow walking. It's like, no, that's just really not going to, to help us move the needle.

[: 00:24:11

[00:24:33] If we are giving over proprietary information into that tool, it then becomes. Public information, correct?

[: 00:24:43

[00:24:44] Betsy Cerulo: Okay. So it's best when it's, if you are wanting to frame something that's more general, that's where it's probably a better tool. So you're not giving away the secret sauce.

[: 00:24:58

[00:25:04] Um, there are several different. Platforms that you're able to use where you have essentially a closed system, right? So your data, your information is not part of the training model on any other platform, right? But you're able to make the best use of the platform, again, to refine text or to say it in a better way, tho those things, right?

[: 00:25:32

[00:26:06] So again, there's, there's, there's other, uh, ways in which you can leverage a, as an example, Microsoft Copilot, right? Mm-hmm. Yep. You can use that, right? So there's, there's different ways in order to do that where now you're in a closed system and your data is protected.

[: 00:26:23

[00:26:34] Okay, well there's gotta be a way to protect the data. So when you use the term a closed system, that makes it more, uh, understandable, where it's almost like, um, what's the word I want? It's almost like a firewall for your, your data.

[: 00:26:51

[00:27:06] I was about to say something, um, that wouldn't make that much sense, but essentially if you said, I'm going to make sure that copilot. Is available and accessible to my team so that they don't have to use chat GBT or things like that. 'cause there's a a, a short, short window. Short deadline, right? Because again, people gotta get it done and obviously chat, GBT is one of those ways, or coll, et cetera, is one of those ways.

[: 00:27:37

[00:27:48] Betsy Cerulo: Mm-hmm.

[: 00:27:49

[00:28:00] Yes. We want you to use AI a thousand percent. 'cause now from a proposal perspective, we can roll out more. Better, faster if they're not allowed to use chat GBT, but still or any other tool. Um, but they still have the same deadline. The odds are that people are gonna still try to use and meet the deadline using these other tools, right?

[: 00:28:26

[00:28:51] So now you're putting guardrails on even at the browser level, how they're operating. But again, there's so many different options out here and, and, and I, and I try to [00:29:00] tell, uh, leaders that are trying to figure it out. That there is a way that you can do this. You can be both innovative and secure at the same time, um, and continue to provide your team members with the tools and access to things, to, to really show up in a totally different way.

[: 00:29:20

[00:29:21] Betsy Cerulo: advising a CEO today, perhaps me, what, what phased roadmap would you recommend for adopting AI while designing? With the end in mind.

[: 00:29:36

[00:29:54] We just talked about copilot, but Salesforce has AI built in. Um, your [00:30:00] marketing automation platform probably already has some AI features that you really haven't even touched. So why not start there where you already are and figure out. You know, how can we better utilize these features within our current products that we're, again, already paying for?

[: 00:30:18

[00:30:44] But. Now you have to think about how can we get these tools to talk to each other. Mm-hmm. Maybe your customer service AI agent starts talking and feeding, uh, insights into maybe your sales team or your marketing [00:31:00] bot is helping inform product development as an example. But essentially you're building these, this data infrastructure and developing an expertise also within the, the company.

[: 00:31:12

[00:31:38] So now you have an agent, um, that is running 20 or can run 24 7. Imagine having, uh, your whole entire on onboard customer onboarding process be. Automated and coordinate different things between different teams, or they can even make outbound phone calls mm-hmm. As example to, to [00:32:00] customers, customer satisfaction, et cetera.

[: 00:32:02

[00:32:31] Walk.

[: 00:32:32

[00:32:51] Gavriel Legynd: Absolutely. So that is what we call workshopping.

[: 00:32:55

[00:33:24] I'd say, okay, great. How many proposals do you want to. Create, have your team create in a month. If you say, Hey, I want it to be five. Okay, great. So now we've got the end. The end is five in 30 days. We also look at other areas, and again, that's just for proposals, but we look at other areas in the business where maybe there's some inefficiencies and come up with a, a, a, a roadmap, but also a workshop.

[: 00:33:54

[00:34:12] That if we integrate AI into one of these areas, we're gonna start to kind of see how Wow. Um, there we've, we're, we're really seeing more efficiency happening within the team. Our team is way more productive again, in looking at the data, but we have to design with the end in mind. We have to leverage data acro all along the way because we can't improve what we don't measure.

[: 00:34:38

[00:35:11] Right. That's a huge piece of this, right? Assessing risk, looking at the security impacts. Is there vulnerabilities just because you wanna buy a piece of technology because it's mm-hmm. The, you know, the, the new shiny thing does not always mean that we should, and it does not mean that it is going to be the most, uh, viable.

[: 00:35:32

[00:35:36] Betsy Cerulo: Now, is there a certain size company or certain industry that, that your company typically focuses on?

[: 00:35:46

[00:35:52] But historically we've supported a lot of folks in the architectural engineering, construction, manufacturing, [00:36:00] uh, more the industrial kinds of companies. But again, over the last, I'd say two years, we've had a ton of different types of companies we've been supporting, um, proposal writing companies, which is interesting.

[: 00:36:11

[00:36:40] And how can we use and leverage these tools, um, to, to do that?

[: 00:36:47

[00:37:07] Gavriel Legynd: Yeah, so what we're, we're really most proud of is just the fact that we have.

[: 00:37:13

[00:37:37] And so because. The world has, has, has been changing so much, especially over, you know, since 2025, um, that they're trying to figure those things out. There's tons of businesses, unfortunately, that have gone out of business or, or lost a significant, significant amount of, of contracts. And [00:38:00] so us being part of their te their go-to team that helps them not just.

[: 00:38:08

[00:38:35] Bringing us in, we can help solve a lot of those problems, and we understand where you're going. We understand how to really secure your data and be part of your growth.

[: 00:38:47

[00:39:11] Whether or not we like the changes. As leaders, you have to keep looking down the road because if you just say, well, we did it that way, you know what? What we did a year ago, it does not matter. Now, as far as customer service, integrity, all the foundational things that, that we have built our companies on, yes, those all stay intact and we enhance them.

[: 00:39:35

[00:39:52] Gavriel Legynd: I agree with you, and it goes back to something I was saying about you're paying for some of these tools already.

[: 00:39:59

[00:40:22] And, and, and, and that's not the best way for us to, again, stay cutting edge, to stay nimble. Mm-hmm. Uh, and, and support our, our employees and continue to grow our, our company.

[: 00:40:36

[00:40:57] And you've made a really good case how [00:41:00] you don't have to be. Afraid of replacing, you have being replaced. You have to understand the tools to see how you can even leverage your role into more of a position of strength. And I think the more information that we all have on technology and how to use it, right versus racing to get it done, I think we're gonna get people's attention more when we stay on that path of how to improve.

[: 00:41:29

[00:41:32] Betsy Cerulo: Oh, Gabriel, absolutely. I just really appreciate what you've given us today and appreciate the conversation and as always, your work is, is so filled with integrity that it, it's, you are the kind of co-pilot I would recommend for my own company and to, and to our listeners out there.

[: 00:41:51

[00:41:55] Gavriel Legynd: Thank you so much, Betsy. It's been a pleasure to be here, uh, with you. Deep bow of [00:42:00] gratitude and I look forward to, uh, supporting you and or your listeners in the very near future.

[: 00:42:07

[00:42:15] Seek out Gabriel and Gabriel, how should people get in touch with you, should they be interested in your support?

[: 00:42:21

[00:42:35] Betsy Cerulo: Wonderful. Thank you so much, and I wish for all of our listeners much success and more wisdom as you thrive and grow your businesses.

[: 00:42:46

[00:42:47] Gavriel Legynd: Thank you.

[: 00:42:48