Value Driven Data Science: Boost your impact. Earn what you’re worth. Rewrite your career algorithm.

[00:00:00] Dr Genevieve Hayes: Hello, and welcome to Value Driven Data Science, the podcast that helps data scientists transform their technical expertise into tangible business value, career autonomy, and financial reward. I'm Dr. Genevieve Hayes, and today I'm joined by Heidi Araya. Heidi is the CEO and chief AI consultant of BrightLogic, an AI automation agency that specializes in delivering people first solutions that unlock the potential of small to medium sized businesses.
[00:00:34] She is also a patented inventor. An international keynote speaker and the author of two upcoming books, one on process improvement for small businesses and the other on career and personal reinvention. In this episode, we'll uncover practical strategies for implementing high return AI solutions in resource constrained organizations, helping you to create value for businesses traditionally considered too small for AI.
[00:01:05] So get ready to boost your impact. Earn what you're worth and rewrite your career algorithm. Heidi, welcome to the show.
[00:01:14] Heidi Araya: Thank you so much. I'm so pleased to be here today.
[00:01:16] Dr Genevieve Hayes: Whenever you hear about organizations integrating AI into their operations, it's usually the biggest businesses with equally large amounts of cash. Yet, despite how it appears, these businesses make up only a small proportion of the economy. In Australia, for example, 99. 8 percent of businesses employ fewer than 200 staff.
[00:01:40] However, to date, these businesses have missed out on many of the benefits of the AI revolution. And one of the reasons is because many data scientists just don't know how to create business value in organisations with limited resources and technical constraints. But it can be done, and for data scientists working outside the big business bubble, knowing how to do so can open the door to a world of possibilities for your career.
[00:02:09] And Heidi, you're a living example of the truth of this statement. So to begin with I'm curious to know, given so many data scientists do struggle to see how their skills can be used to create value in smaller organisations, how did you come to see the potential value AI could bring to these organisations that everyone else was missing?
[00:02:31] Heidi Araya: Yeah. Interesting. So my whole career has been in enterprise, so that's important to know because I haven't worked with small businesses at all. And originally when I took my program at MIT, I thought this is where I would end up again. I would be leading some sort of data science initiative or machine learning initiative inside a large organization.
[00:02:49] But by the time I finished the program, I thought to myself, this is going to be just like the other change and process improvement initiatives that I've experienced in my career where The data wasn't going to be there and yet there's a rush. So we need to get this AI thing out. So let's just do the thing.
[00:03:05] And I thought, I just don't want to do that anymore. At the same time I had been engaging with. My local community and small business owners and kind of talking to them about their needs and people weren't even adopting AI, not even chat GPT. And I thought, these are the people that I need to work with, not only because they really need me, no one is serving them, but also they're not going to let anybody go as a result of AI, because they're just struggling.
[00:03:31] And so this was kind of my thought process around how I recognize AI's potential was just listening. I was able to then. And this is where I came in, because I was very interested in how to leverage my background in process improvement and value stream mapping that I had in large organizations and my customer focus, because the process improvement initiatives always around.
[00:03:51] There's no need to improve a process that there's no customer at the end, or why does this process exist? So I was very focused on the outcome that the business was interested in and I originally was focused on operational process improvements, which I thought, you know, we can do a lot of those and we can automate a lot of those.
[00:04:11] Interestingly enough, though, as I was listening, they wanted help with the sales and marketing process or some AI educational or some something very concrete. That was not an operational process, even though I told them. I can give you back 10 hours in your week. If we improve this operational process, so you can focus on marketing and sales.
[00:04:31] No, no, they just wanted to focus more on marketing and sales up front. So I guess my point to recognizing as potential was my own journey my background and process improvement. Listening to what business owners actually needed and their problem statements that they came to me with. Can I do this?
[00:04:46] Can I do that? And I really was focused on pragmatic versus pie in the sky idea. So, what can I help them with today? They only have limited funds. What is the sort of minimum proof of concept that I can get them to sign on, whether it's some research or some quick AI receptionist
[00:05:04] Dr Genevieve Hayes: So just to clarify, when you're talking about these small businesses, how big are the businesses in your client base?
[00:05:10] Heidi Araya: 1 to 100 people.
[00:05:14] Dr Genevieve Hayes: And what industries are we talking about?
[00:05:17] Heidi Araya: So roofers, cruise company, window replacement company, of the smaller businesses, the solopreneurs, their authors, their attorneys. I have a voice over IP company. I've got a handyman that's got seven employees. So those are the kind of businesses that I've been supporting.
[00:05:35] Dr Genevieve Hayes: So it sounds like you've got quite a diverse range of people, but also it's a very diverse range of levels of sophistication of those businesses.
[00:05:44] Heidi Araya: Yes. Yes, it is. And interestingly, the attorneys they have just one specific use case. The attorneys have podcast episodes and lots of intellectual property of their own. And people would go to them and call them up and say Hey, random question. And they said, but that answer is on my website.
[00:06:02] You just go to the website. And so they were able to direct people to the website instead of spending their day answering questions that could have just been answered by the chat bot. So for them, that actually solves a big business problem. They were sophisticated enough to find me and reach out to me.
[00:06:19] Other times people come to me with a problem like, oh, I'm spending all my time answering the same questions over and over again. I say, oh, have you considered an AI chatbot? But these already knew they they wanted an AI chatbot. The roofing companies. Well, I had 1 that said when we close and on the weekends, we're losing a lot of phone calls. That's decreasing our Google ranking. So they were losing 160 to 170 phone calls per month and the guy had heard a podcast episode that I was on and said, I think we need this AI receptionist to answer the call.
[00:06:50] And so they can book roofing. Inspection appointments and that way our Google ranking will stay high. We answer calls right away and then 47 percent of their calls came in after hours. So it's like some of these use cases are very simple and niche, but they solve an actual big business problem for them.
[00:07:08] Dr Genevieve Hayes: One thing a lot of people miss, when you talk about small to medium sized businesses, the images that often springs to mind is that of, your little restaurants and your dry cleaners that you find down at the local shops. There are actually some quite sophisticated businesses that also happen to be relatively small.
[00:07:26] For example, a few years back I worked for a government insurer that had Just under 200 employees, which means that the ABS would class them as being a medium sized organization.
[00:07:38] Of course because of the relatively small size, that did create challenges when it came to doing data science. Which leads me into the next question. What are some of the specific challenges that you've found that small businesses encounter when implementing AI and data science solutions, larger organizations might not?
[00:08:02] Heidi Araya: Well, larger organizations are free to spend a ton of money to have their AI solution, whatever it might be. So let's. Manufacturer data, let's get the data somehow, but smaller businesses are just less focused on data. They're just literally trying to survive and they don't have anyone looking at that.
[00:08:19] They're marketing persons. If they even have 1 is not looking at that kind of stuff. They're just trying to get by on the day to day. They are not very sophisticated as far as technology. Of course, like you noted, they don't have a lot of data and, a lot of times instead of enterprise platforms that have lots and lots of integration options and API's to look into, they often have very unsophisticated systems or spreadsheets they're using.
[00:08:41] And I think lastly, they're just risk averse because what they have been doing so far has worked for them and they don't have a lot of money to invest also. That's why when I'm pitching some solution to a smaller business. I don't even start with AI. So well, it's not an actual business. It's a nonprofit, but I had a guy reach out to me saying we are struggling to reply to the people who want to donate.
[00:09:06] And then there's a particular cadence that you should be reaching out to people and sharing this nonprofit. Here's the great stuff that we did. You're supposed to be doing that 7 times before you ask them for more money. So you're nurturing your donors. He said, we're just not doing that. We can't do that.
[00:09:21] This is not a new problem for nonprofits. I've spoken with a lot of them over the past year. And he said, well, I think we need AI. And then I told him, actually, if you're not even reaching out again to them. You don't even need a I just start sending automated emails and nurture sequence. Why complicate things with a I then, depending on the response that you get, then you can add a I to this and kind of superpower.
[00:09:45] But so that's another thing that I would sort of pitch to people who want to get started with smaller businesses. They don't have lots of money to spend. They're very risk averse, solve an immediate problem and then just say, okay, maybe you don't need AI now, but we can super charge you later. So I have those solutions as well.
[00:10:01] Dr Genevieve Hayes: One of the problems I found when I was working for that smaller government organization is that a lot of the enterprise options for software were pitched at massive organizations the size of a big bank. So it just wasn't economical for us to get the enterprise version
[00:10:19] It basically ruled out a lot of solutions that we would have liked to have had. And not to mention even if we could have afforded the enterprise version, we couldn't afford the support. and we had 200 people. If you only have 20 people or less, that's going to be magnified even further.
[00:10:37] How do you deal with that?
[00:10:38] Heidi Araya: So over the past few years, well, 2 years, especially there have been platforms now built that. We're not built 2 years ago. If you're not aware of the platforms, I'll give you 1 example. So in October of 2023, I was very excited about these chat bots. And I said, let's build a chat bot platform because I need the secure platform.
[00:10:57] And so I work with a couple of people. We build it in Microsoft Azure with secure storage blob storage. And we were doing all this stuff and it was a very rudimentary platform, but it suited this person's needs because he didn't want to put his books anywhere publicly this. He's like, it's my IP.
[00:11:13] So my 28 books. And of course, we need a big storage for the 28 books. Shortly thereafter, though, the 3rd party platforms. Became more robust, they started adding features and so my team couldn't keep up. With the new features and these products that were coming out and had whole teams dedicated to them that are now adding on, compliance features and all these security features.
[00:11:34] So I do build. Custom software is needed and I'm building some right now for some clients, but if they don't need it, then search for a 3rd party platform. Perhaps the problem has been solved in another way before or an automation, or perhaps if it's not, if the ROI isn't there, is that really the problem to go after?
[00:11:53] And maybe it's not that problem that you want to go after right now.
[00:11:56] Dr Genevieve Hayes: So you've alluded to a number of solutions that you've provided to your clients, AI chatbots aI receptionists, things like that. So it sounds like there are a number of use cases that come up again and again in your work.
[00:12:10] Which are the ones that you find deliver the greatest ROI to your small business clients?
[00:12:16] Heidi Araya: I find that something as simple as an AI receptionist who can answer the calls when they're not there. 1 example is a handyman that I had with 7 employees. He was missing calls. And, of course, when people call a handyman,, they want to answer right away, or they want to know that you're working on it.
[00:12:32] So I actually went through and did a process mapping activity. And discover that he could actually automate 75 percent of his process. Without going into too much detail here. We identified the highest ROI for him was going to be those phone calls that he was missing because he was missing like 15 per day.
[00:12:52] And the AI receptionist could handle the call gracefully, ask the customer their name, the city they were in, the phone number, what the project was about, and then also ask them to send photos to an email address or text them to this place. So, by the time he looked in the evenings, between 6 and 9 PM, he would have the photo.
[00:13:09] He would already know what city they were in. So he could give an estimate for the job and already kind of pre dedicate the person who would go there. So that tends to deliver high ROI because. We need to answer those calls right away, because 95 percent of our leads are going to go elsewhere. If you don't respond back in 5 minutes.
[00:13:25] And the chat bot, it's a simple solution. But from the chat bot perspective, I've had people create a I personas of themselves. The guy was the author with 28 books podcast hosts and any number, even internal customer support documents. So many of those. It's such an accessible solution to say, speak with AI Heidi, if you only have here and so my clients love those, any kind of chatbot that's built on information or can be trained on the use case that you want.
[00:13:59] I had called just before this . chatbot trained to empower women and all the research that's out there that shows about what women are bringing to the workplace. So those are like really simple solutions, but I guess those are the 2 that stand out. And the 3rd one would be, although.
[00:14:13] It's not for super small companies, but a slightly bigger, just because it takes lots of work would be automating some part of the sales process. So if you have a decent sized business in the U S it'd be like one to 5 million in revenue, which is fitting in a small business area and you have some sales people and they're struggling to follow up with
[00:14:34] the customers and the lead. So, for example, I go to networking events and I get leads. I get customers. I have to add them to my CRM and I, oh, my gosh, I mean, I have stacks. I can't follow up with everyone who has expressed interest and that list just keeps growing. So, for companies who have a database of leads.
[00:14:52] Automating the nurture sequence to do outreach is super valuable because no human could do that. I have had a cruise line that had 26, 000 leads that no human could reach out and the guy says, I can't make people call them and anyway, they don't answer and then the volume of calls for AI.
[00:15:11] So AI can make 1800 outbound calls per minute and a human can make what like 100 in a day. Even if they capture just 5 percent of the valid leads, we calculated it'd be 5 million in revenue coming in. So you know, just fascinating. And you wouldn't want to do that. You wouldn't want to react all those leads, right?
[00:15:27] Imagine if we did all the calls in one day and then the guy was getting overloaded and he couldn't handle the calls coming in to book the cruises. But those kinds of things that take humans a lot of time where it's impossible, or you just. Don't have time to do it. I guess those are the high ROI opportunities.
[00:15:43] Look for those things that people can't do because of time or energy or I'm answering the same calls all the time and answering all the same questions.
[00:15:52] Dr Genevieve Hayes: Do you have a process for identifying which of these high ROI AI implementations is going to be best for a particular small organization?
[00:16:01] Heidi Araya: Yes, I do. So based on my years in process improvement, I start with asking them, what problems are you experiencing? So typical person might come to say, Hey, I hear you, the AI lady. So how can AI help me? And I just say, well, I don't know. Tell me what problems are you experiencing or challenges that you're experiencing in your business today?
[00:16:19] Or maybe you have a goal you'd like to reach, like scale your business from 800 K to a million or something like they do. And then start there. If they have nothing, then I don't know, AI only exists and the data exists, , to solve a business problem. So I'm not here to just pitch you on adopting AI without a purpose.
[00:16:40] So assuming that they have a problem that they're aware of. I then start with a process mapping activity where I go through end to end what the process looks like today. If they don't have a process yet today, then I say you need to get it first because I can't automate a process. It doesn't exist or can't add a process.
[00:16:59] So I go through this process mapping activity, or I help them establish that process. I talk about the pain points and where I could offer Opportunities and maybe it's not a, I, maybe it's just an automation set some clear objectives and goals for that. And then try to run a proof of concept because remember, these are risk averse folks.
[00:17:20] So you say, well, let's just start with this. Your AI receptionist answers the calls after hours, and then you see how it goes and we'll examine the results and we can iterate on them. But that's my approach. Really?
[00:17:30] Dr Genevieve Hayes: Yeah, and it's pretty much the standard data science approach just applied to smaller organizations. But I'm guessing given the size of these organizations, they're not going to have a lot of data that you can train models and things like that on. So. How do you deal with these budget and resource constraints and how do they affect your approach to delivering the solutions for smaller clients?
[00:17:54] Heidi Araya: I deliver something really stripped down as needed. I'm not trying to sell them bells and whistles, just like the nonprofit. And I said, look, you're a nonprofit. I can't sell you my super duper fancy AI sales solution and marketing solution, because it's going to be a lot of money for you.
[00:18:09] And then you're just going to walk away, but I don't think you need that. So I talk to them at the level where I'm really just trying to help them solve the next problem. And like I said, in this particular case with a nonprofit, I said. You don't actually need AI, but I can help you have someone build the automation for you and I'll turn it over to you.
[00:18:27] You can run it in your platform. So I have to be very pragmatic. Number one, I'm not trying to sell them something that they can't afford because they won't. What that does though, is establish trust in the relationship because I'm listening to them and I'm not trying to oversell them on something.
[00:18:40] Very few people and consultants want to talk to these businesses because there's not a lot of big money in it, but that's okay. I'm not going to be the Accenture or someone going in and I think it's a missed opportunity because some of those small businesses will become bigger businesses because of AI and the things that we're doing to help them.
[00:18:58] So I look at it a longer term approach.
[00:19:01] Dr Genevieve Hayes: But and I think I already know the answer to this are you actually coding things from scratch or are you making, taking advantage of pre built solutions?
[00:19:11] Heidi Araya: Yes, I'm taking advantage of prebuilt solutions. I'm not coding anything from scratch. Yeah, I have clients who need some coding, but I have an offshore team to do the more robust coding stuff when needed. But no, most of the time I'm leveraging. Yeah. Yeah. The robust third party platforms that exist out there today.
[00:19:28] Dr Genevieve Hayes: Yeah, because I'm guessing that for an organization this size it just wouldn't be economical to build custom solutions.
[00:19:35] Heidi Araya: Probably not. I
[00:19:36] Dr Genevieve Hayes: And I've noticed there's a particular piece of software that I'm using at the moment and it's in development still. So I get these emails about all the latest bells and whistles they've added.
[00:19:48] And from what I can gather, they are making use of a lot of prebuilt solutions. For example, they've just added AI functionality and they've actually asked and said, you know, we're using Google Gemini for AI. So they haven't. They've gone out and built their own LLM. They've just got a pre built solution.
[00:20:10] Heidi Araya: see. Yes. Okay. Exactly. So I will leverage open AI or anthropic or whatever in my automations. But yes, I'm certainly not building anything like that. And for the solutions where, like, you alluded to small companies don't have data. And how do I do that? Sometimes I have to work with them to get the data that they need, which is some pre upfront work.
[00:20:31] So I do consulting to come up with the information they need. So 1 example is. I have a company that wants to automate part of their sales process, but they are just hired their first salesperson. I said, well, we can use AI to onboard your salesperson more quickly. But once you have an established process, then we can automate the nurture sequence and all this sequence.
[00:20:50] But until you have that process where the person knows, Okay. How to do outreach, who's your ideal client profile? Like, there's all this upfront work we need to do. So we can train the AI on at least this kind of information. So I help them get the information they need. And then we're leveraging 3rd party automation platforms to make this things happen.
[00:21:08] Dr Genevieve Hayes: So to make this concrete for our listeners, could you share a specific example of how you've helped a small business to achieve significant improvements using AI?
[00:21:19] Heidi Araya: Yeah. So there is a business lending company. And they had lots of leads in their database that nobody was reaching out to. So people had contacted them. And in fact, with those leads and AI doing some outreach and asking if those, those people, the owners of the business were still interested in those loans, people called back and that campaign generated 30 million of extra business for that customer.
[00:21:50] In a very rapid amount of time, such that they had to hire more salespeople to close more of the calls. So you see, it's not always about letting people go or downsizing, but really this AI powered campaign help them sell more. So that's just 1 example.
[00:22:04] Dr Genevieve Hayes: So, even though we're dealing with small businesses here, the returns can be quite a large number.
[00:22:11] Heidi Araya: tremendous. Even the cruise company, although we didn't end up closing on the cruise company. They had 5 million worth of leads in their database. And the company was making under a million dollars a year. So you could see that would have been a significant improvement.
[00:22:27] Had they invested, say, 10, 000 dollars, they could have gotten 5, 000, 000 and scaled from 800 up that was 1 of their goals was to scale from 800 K to 1, 000, 000 in this year. And then in the end, they just they weren't ready for the AI solution. So those are tremendous, tremendous opportunities.
[00:22:43] And. The handyman got a bunch of calls every day on his AI receptionist that he was then able to keep his people busier and then said, I don't want you to AI everything because if you did, then I would have too many jobs. And I don't have the people to handle the job. So first I have to hire somebody and then I will add on more AI.
[00:23:01] Dr Genevieve Hayes: So what I'm hearing here is the data science process for dealing with small businesses is pretty much the same as when you're dealing with larger enterprises, except you're relying a lot more on pre built solutions than you might with a bigger organization. So any data scientist who knows how to work with a big organization can potentially apply those skills to a small business.
[00:23:26] And even though these organizations might seem small, the ROI is probably in percentage terms far greater than it might be in a bigger business. So it is worth making the investment in order. to help these organizations thrive.
[00:23:45] Heidi Araya: Definitely. It's definitely worth it. And some of these, you couldn't even hire enough people to make say, those phone calls, right. To do the work, you couldn't even invest that much or hire those many people to do the thing. So the scale at which we're able to operate today and the cost at which you can deliver those solutions, if a business is able to invest even just a little money, it could be a huge ROI for them.
[00:24:06] Dr Genevieve Hayes: So, what's the single most important change our listeners could make tomorrow to accelerate their data science impact and results when working with smaller businesses?
[00:24:17] Heidi Araya: I think it is listen to your actual potential business owners and see what they want. It took me lots and lots of phone calls, lots and lots of listening. What words are they using for the troubles that they're experiencing? Because. What my customers are telling me here in the U. S. is going to be different than what they might be telling you and your customers.
[00:24:36] So yeah, I would just say, start by doing listening tours and start gathering information on what the pain points they're experiencing. And don't approach it with, hey, I'm a data scientist. I'm going to offer you some solution. Just start listening to what the pain points are first, and then maybe seeing how your expertise can help them.
[00:24:53] Dr Genevieve Hayes: Yeah, that's very good advice. Don't start with the sale in mind, start with the intention to help.
[00:24:59] Heidi Araya: Yeah. What problem are they experiencing? Or a challenge or goal they're trying to reach this year.
[00:25:04] Dr Genevieve Hayes: So, for listeners who want to get into, in contact with you, what can they do?
[00:25:09] Heidi Araya: You can find me on LinkedIn. Heidi Araya. You can check out my website at bright logic dot. A. I. Those are the places that you can typically find me. You can send me a message on LinkedIn. I'll respond right away.
[00:25:21] Dr Genevieve Hayes: Okay. And there you have it. Another value packed episode to help turn your data skills into serious clout, cash, and career freedom. If you enjoyed this episode, why not make it a double? Next week, catch Heidi's Value Boost, a five minute episode where she shares one powerful tip for getting real results real fast.
[00:25:44] Make sure you're subscribed so you don't miss it. Thanks for joining me today, Heidi.
[00:25:49] Heidi Araya: You're very welcome so much. Thanks for inviting me on the podcast.
[00:25:53] Dr Genevieve Hayes: And for those in the audience, thanks for listening. I'm Dr. Genevieve Hayes, and this has been Value Driven Data Science.