RealmIQ: SESSIONS with Svetlana Makarova

  In this episode of RealmIQ Sessions, host Curt Doty interviews Svetlana Makarova, a leader in AI solutions. Svetlana shares insights from her journey, starting with a background in biochemistry and psychology, and evolving into a career focused on AI-driven digital solutions. The discussion centers around her experience in healthcare, where she has developed AI solutions that improve healthcare delivery, including expertise-finding tools and clinical decision support systems. They also explore broader AI adoption in industries beyond healthcare, the importance of AI governance, the role of large language models (LLMs), and strategies for scaling AI within enterprises. Svetlana emphasizes the importance of empowering teams through workshops and practical engagement with AI tools.

Topics Discussed:

1. Svetlana's Background: From biochemistry and psychology to AI leadership.
2. AI in Healthcare: Expertise-finding tools and clinical decision systems.
3. Scaling AI Solutions: Challenges and strategies for large enterprises.
4. AI Adoption in Business: Overcoming resistance and fostering internal AI advocates.
5. Data Governance and Privacy: Safeguarding patient data with AI in healthcare.
6. Proof of Concept (POC) for AI: Demonstrating AI's potential with low-cost, scalable solutions.
7. No-Code Solutions and the Future of AI: The importance of no-code tools for businesses.

Pull Quotes:

• Svetlana Makarova:
  • "All AI systems are really pattern finders. They find patterns in data, and that’s what they do."
  • "The unfamiliar is uncomfortable. People hear 'AI' and think 'black box,' but it’s about breaking that down."
  • "You bring people along on this journey so they understand how it works and become empowered to apply AI themselves."
• Curt Doty:
  • "AI is a revolutionary technology that can benefit business, but you have to break that wall of unfamiliarity."
  • "Building AI heroes within an organization helps transform the business from within."
  • "We are moving towards a no-code world, making things easier and increasing the speed of innovation."

Transcript

SVETLANA MAKAROVA

Hi, I'm Curt Doty with RealmIQ. This is our podcast, RealmIQ Sessions, where we talk about everything AI with AI leaders from around the world. Please give us a follow and or subscribe. Today's guest is Svetlana Makarova.  For over decades, Svetlana has been at the forefront of developing digital solutions across a range of industries, always with a focus on harnessing new technologies.

She's now focused on AI solutions, But also empowering leaders to unravel AI's complexities and discover its profound impact. Welcome Svetlana. So let's get into it. Let's, you know, elaborate, please, on your background, your solutions, and how you are working in the world of AI.  Yeah, I don't know how far you'd like me to go, but I'll just kind of start with, like, some themes about my background.

So, I don't have a traditional tech background is probably most people who end up in tech are my, but my background is in biochemistry and psychology. I just figured that I want you to kind of, I, I'm the type of person that likes to understand things to, like, the.  I want to say like molecular level, which is where biochemistry came in and then psychology.

We need to understand how people work. And, you know, and I think those 2 things, those 2 types of skill set kind of carry through my career because I am a problem solver. I like to understand and take things apart into different components to really understand it. Because if you understand. How things work, you can build things with them.

So that's kind of been my fascination and a thread throughout my life. And then I got my MBA, so it kind of enabled me to understand business at a higher level, understand kind of how things operate across big enterprises. But also, I, I'm one, again, I, I love to understand how things work. I am trained in UX visual design.

Data analytics. So, all of the components of product, basically, for me to be able to lead teams or lead kind of product initiatives more successfully. So, I've always been that way. And I think it just more recently I've been, you know, Fascinated even more. And I think the world of AI kind of came as I was in a way kind of forced into it.

But that was like that opened up a new world to me because I was like, this is a new area that I don't understand. And it's never ending. And I think that's what motivates me to continue and learn more because the moment that you think that you understand something. Like here is something like a new evolution of AI happening in this place.

So, there's always something new to learn and new skill to gain. And that kind of motivates me. But yeah, no, I think that, you know, I've been doing AI for the last, I want to say three, three years or so. And I've been focusing on leading teams, developing quite complex solutions within healthcare. And then I think, as you mentioned in the introduction, I've started to kind of raise my voice about and share my knowledge on LinkedIn primarily.

And that kind of grew a following and following that, that had. Folks’ kind of leading, you know leaning in asking for guidance, advice on solutioning, scaling solutions, how to use them in their enterprise, how to scale products, maybe that they've built to an MVP stage, but then they have difficulties actually building a full-fledged solution, change management, everything that comes across, you know, I've done at an enterprise level.

And I think if I can do it within healthcare, I think anything else, any other industry becomes. Significantly easier.  Yeah, well, fascinating that you're in healthcare. I was pre-med before I went to art school, but so I, I share a similar kind of right brain left brain equality in terms of interest and thinking.

And so, in healthcare, please tell me.  Like what you've been doing with A. I. and health care because I think that's a fascinating aspect of A. I. because there’s some altruistic goals there for, you know, people's or patients’ outcomes in terms of health care. So, elaborate on that a little more.  Yeah, and I want to emphasize.

So, I at an organization that has embraced a I quite early. So that's why they've been doing AI for quite some time. And I focus and kind of the initiatives that I support have an enterprise wide. View, but there are specialties, distinct specialties within the, the organizations that have their own AI initiatives that are very specific to their practice.

For example, if you're in radiology, you've seen all the cool stuff that's happening in AI for specific to, again, to that, that specialty, to surgery. We don't focus maybe as much on that because that is quite niched to specific specialties. So, the, the projects that I undertake have this enterprise lens and that's why scale is important.

So, anything that we build cannot just work for 1 or 2 specialties. It actually has to have this broad, broader focus. So, some of the initiatives, and I think 1 that I can talk about, and I've actually presented this at the American telehealth conference earlier this year.  2024 was one of the earlier projects we developed was an expertise finding tool that enabled us to locate different areas of expertise that someone needs to consult someone on.

So, for example, what actually happens and, and the nature of the organization that I, I work within is multidisciplinary care is very crucial. So, when you're in the moment of providing care, You have a question that you have to ask, but when you have such a huge organization, you have, I think we have over 4, 000 physicians and let's say. 

I'm not, I'm going to overestimate, but like, maybe 10, 000 you know, advanced practice providers, like, who do you ask?  Like, how can I find the right person to ask if they're not maybe my colleague, or if I don't have a network of people to, to realize. So, we've developed this AI system. To look across data across the enterprise, and we looked across lots of different data sources to create a profile for providers to understand.

Okay. In the moment of need, who can we reach out to? And these, and that was really crucial because previous to that, everything was done manually and because physicians don't have time. Even to do a lot of their administrative work, this is the last thing that they're going to worry about is figuring out how to, you know provide their areas of expertise manually to some tool.

That is going to be rule based. Type of recommendation system so we knew that we wanted to kind of take a low hanging fruit solution. We know we knew it was needed. It is highly used within the organization. Even the manually derived. It just wasn't getting updated on a regular basis. And just to give you a perspective.

The system that the, the organization was using that was manually derived physicians haven't updated sometimes for like, years. So, on an annual basis, I think they were updating the system, like, 2 to 3%, but we know the. Practice evolves skill sets of all. And so, it often happened that when that interdisciplinary connection with what happened, they would be reaching out to the wrong person or patients would be scheduled to the wrong provider.

And so that's there are multiple kind of needs that we were aware of. Addressing so it was not directly impacting healthcare in that point. So, when you were lower risk, but when you were high value to the organization, because one of the core principles of the organization that I work on is this. And they call it magic.

It's the, it's the magic of, of the organization where, where I work is that multidisciplinary care. So, when you were core objective of the organization, when you were high value. So that just to give you a sense as to how What types of solutions, then we some of the other tools, just to give you a perspective within healthcare, you know clinical decision tools.

We look at information. So how do we democratize knowledge that. Of data or knowledge of the providers, I don't know if you've ever kind of come across, or if you had a condition and you're like, oh, I have these symptoms and then you go on Google and 1 of the top results come up is. You know, again, the organization that that authors this content and, you know, how do you democratize that?

So, like, if someone has a need for a specific question, how do you take the knowledge that the organization has inherent in its EHR and its data, and then you help surface those insights back to the provider? So, it's truly putting the organizational. Data to work and then really democratizing that knowledge.

So that's kind of some of the ideas between clinical decision tools, but there's other things that are quite complex in nature that I've kind of helped lead the efforts in, but hopefully can kind of get you a taste for the types of solutions. I've been kind of behind. Yeah, no, it's, it's fascinating. And when you think about healthcare and consumers wanting information, it’s impossible to get in to see a doctor, you know.

You know, take six months sometimes to get an appointment because that's a very stressed system. Obviously emergency situations are different. But. The fact that you're providing a service and information both to providers as well as patients, potential patients or people in need of information and democratizing that, I think that's very noble.

And again I mentioned the altruistic goals of medicine, and I think you know, this is a category I believe is where AI can thrive because it's less encumbered by, Some of the challenges that are happening in Hollywood with deep fakes and, you know, all this other type of negativity around AI when you're, you're proving really valid outcomes and benefits to using this research in terms of databases, right?

Because.  We are the data people at the data, right? And healthcare sits on a huge mountain of data. There are certain laws around the protection of that, the HIPAA laws.  So how are you safeguarding those types of Aspects of privacy as it relates to patients, patients, privacy and tell me how that works a little in terms of LLMs and data sets. 

Yeah, and I'll kind of preface that some of the solutions that I focus on have a lens to that. And then kind of the answer depends, but the solutions that I focus on the overarching objective is for us to alleviate the administrative burden of the providers. So, people who are actually providing care, so how do we deliver solutions so that they can do less of the repetitive tasks and then focus on delivering care.

Patient care and so our solutions or we have other teams that kind of focus on that on the patient kind of tools. We focus on kind of the providers and the reason why I say it depends. So, for. Let's say for that expertise finding solution, we didn't have to have patient information in the actual data set to train our models.

So, we could de identify, we could basically get the type of data we were using at very specific data points that could not, would not have, Even value to us from a perspective for us to even include, because in that perspective, we worry about what the what type of skill sets the provider has not specific to what types of patients they are treating.

You can get that. So, for example. You don't even need age because again, the types of, um, maybe credentials that the provider holds, you know, they can be a pediatric physician or an adult physician. That's sufficient for us to understand, Hey, this is a pediatric, like focused, if you have a pediatric type of question.

Here are the providers that you need to focus so there are use cases or solutions where you do have to be reasonable for, like, do I really need data? And then so you do identify, and you just use the, the data that is crucial for you to develop that solution. The other example that I provided to you for the clinical decision tools, when the providers are actually delivering care to the patient, you want to integrate some of these solutions within the chart, you want to contextualize it, then you want to embed it into the workflow.

So, there are times when PHI is actually important. So, and again, the only people who are seeing this data are the providers who are providing care and the way that I think you are Safeguarding it and then there's multiple, I want to say, like, levels of control that you kind of have to pass before you can even display that data.

And so, there's rule-based controls. There's data governance. There is, you know, for you to integrate it within, you know, let's say the HR system for you to even get to that patient screen. You have to have the right levels. Of control within that HR record. So, again, there's multiple levels of security for you to get to that level.

So, let's say, if you're not authorized to see a certain piece of data. Then you won't display anything in the tool because again, there's multiple validation systems that you have to kind of pass in order for us to even surface that data back to you.  That's fantastic. So, what a wonderful playground to figure out AI governance and, and now you're branching into business, not just healthcare.

So, what are you seeing in terms of the ease of transferring your knowledge and, and discipline into areas of business, which have been slower to embrace it possibly than, than healthcare.  Yeah. And that's a, it's a.  Passionate topic of mine, I'm writing more and more on this topic on the blog slash newsletter.

So, I do think that adoption. So, and I say this a lot, even in my LinkedIn posts, if you go back, I think that unfamiliar is uncomfortable. So, organizations. That I've spoken to, I go to networking events, and they speak to a lot of business leaders. And though the time, the moment you say AI, people say black box, because again, it's unfamiliar and then it's uncomfortable.

So, people kind of stay away. And one example that I like to provide is, you know, when you go to a conference and let's say, you know, maybe your audience has some experience going to conferences or these big events, right? First of all, when you walk into a big crowd, you don't gravitate and like to You know, groups of people and just kind of start the conversation.

Some people do like, I've been called out on that before. They'll be like, I'm the type of person that will actually jump into any conversation, but most people would not. And the reason why is like, those folks are not familiar. And then you have all of these ideas and preconceptions for like what they might think of you.

And, and so typically we have the tendencies to stay away from things that are unfamiliar. And we gravitate towards booths, or we contact folks who are the same networking or event or the conference. Like, Hey, where are you? Like, let's, let's kind of hang together. Because we want to kind of gravitate towards like these familiar things.

And I think when you kind of break down the concepts and I think what, what a lot of the courses books that I've read, and I've, I've done a lot of that, they don't do AI justice is breaking down the concepts to the core of what they are. Which is kind of what I love to understand. And I like to create patterns and, and truly understand how things work.

All AI systems are really pattern finders. They find patterns in data and that's what they do. And so, when you start building on the foundations and you kind of give examples for what the AI, and I speak about it in terms of capabilities.  What can AI do, and it gives you a perspective, okay, I need data, I need this capability and okay, I get, I get a sense.

So, once you open up that blog box and you understand all AI is, is just data and capabilities, it becomes much easier to apply it to your business objectives because sometimes in the way that I kind of am. I don't just do AI. Sometimes when I come and I solution a particular problem or an objective for the organization, sometimes, and I've called out like even in different workshops you know, sometimes I'm like, that doesn't need AI.

So, all your kind of doing during workshops, when you're looking at specific processes, when you're looking to enhance products or revamp old systems, you're just basically trying to understand, well, what do I need To do in order to for this product to work for this product to scale. And you align that capability to, does this, is this a rule-based type of thing?

Can we do that with a rule-based automation or is there sufficient amount of data and potential for learning that. That actually would benefit from an AI system. So, once you kind of understand some of these capabilities, you've built the foundation, you're like, okay, I see what AI is. All it's doing is really expediting a lot of the things that you were able to do with rule-based systems significantly further.

And there's additional benefits, as I mentioned, with rule or I'm sorry, learning mechanisms, networking effects that you get from, from these systems, but not. Every product actually needs it. So, you have to be kind of critical when you're, you're developing these systems or when you're creating the strategies for these systems, you have to be like, okay, I need this.

For example, I would love for us, the objective could be I would love for us to recommend better. Products to our existing consumers, I would like for them to check out with not only 1 product but provide a relevant recommendation to them at checkout so that they can add it on. Well, I don't think that's a that's a rule-based task just because you know, Just thinking about the number of customers you have to satisfy and you kind of have to go one by one and figure out well, what kind of let's say products they've purchased in the past, what kind of things or what's their price point.

There's just so many different factors. You have to then convert into rule based. At scale, that doesn't work. So, I think that there's telltale signs when your kind of developing this workflow and you're looking at a specific task you want the system to do. And you're like, okay, well, how, how would it reasonably work?

And if scale is an issue, like, we'll never get it done with, like, by defining rules. That's like a sure. Sign that you need AI in it because that is kind of the whole benefit and premise of AI system is Scalability aspect of it learning and then being able to do that at scale like at higher volumes.

Yeah No, it's awesome. And I appreciate the analytical mind that you have You talk about this Was kind of a freeze paralysis and a barrier that even when you're at a networking event, people can't gravitate towards other, other people. Unless they have something in common, or there's some type of affinity understanding when, when that's really a barrier, because you need to be talking to those people who don't know about the benefits of AI and break the ice.

I believe that there is a thaw in that happening right now, you know, after a year and a half or more of generative AI discussions and experiments.  I think people are more curious now than ever because these companies are going to have to deal with it. The EU act is getting implemented in Europe.

We'll see how the United States follows.  But how are you going to break that, that wall, that divide and increase AI adoption? I know you do a lot of work. You do a lot of talking and writing. I do the same, but sometimes it's not enough. And you know, what's the guidance there for yourself as well as other people who are very interested in the evangelism of, you know, what is a revolutionary technology that can definitely benefit business. 

Yeah. And I think just, just probably the listeners won't know, but before we started this podcast, I asked like, who is your target audience? Right. And because I want to understand them. So, I don't get into like the geeky, the very, very geeky definitions of AI that we kind of keep at higher level, just so that you kind of target your advice or target the information that I delivered to the right audience at the right level.

And same thing goes for when you are in these discussions about AI adoption or AI in general, you kind of have to understand your audience. And then there's different types of people. So, if you're a worker and you're trying to, and your objective is to, you've heard about AI, you've heard a lot of the stories, people optimizing their workflows with AI, it's game changing.

My advice for them would be different than like the types of things that I would suggest for them to do, but a lot of it comes down to practice. And then when you talk about businesses and kind of this is where my strength is, is understanding the business side of AI and how people want, could use it to increase value for their customers or for their business.

And Basically, optimize their operational efficiencies. So, they, there's 3 types of companies that McKinsey had kind of segmented. And I think your recommendations would depend on sometimes intermix between, but you have kind of the folks who are what McKinsey calls takers.  You have makers no, hold on.

So, they have takers, shapers and makers. I'm sorry. So, the takers are folks or smaller companies, typically like a small boutique shop. Let's say they don't have a big budget to adopt the AI solutions, but they see value, or maybe they've educated themselves or seen enough in the space about AI.

They're like, Oh, I, I think that's It could be valuable into organization. I just don't know what use cases in my organization to apply this to. But kind of, again, my guidance to them would be knowing the size of the company, the budget that they have, they’re probably not going to invest into like cloud infrastructure.

So, you are going to be looking at a suite of subscription tools that will enable the operations within their organization. So, you look strategically across what kind of verticals do they have? Not an, you know, if you're a smaller boutique, you probably have maybe two or three parts of your like organization, maybe marketing, legal sales, and that's kind of what you're focused on.

So, what subscription tools can you recommend to basically augment their workflows to improve their products within those basically three pieces. But you have, again, like your mind is set on recommending tools that are existing in the market, because again, there's subscription costs, but it's Lesser costs to run because they don't need big, you know, a huge number of licenses or people to actually, like, operate these AI systems when you're kind of in that shaper segment, which is the 2nd segment people these organizations tend to have some understanding about. 

what AI can do. Maybe they've dabbled with AI, they've kind of experimented, they've built proof of concepts even. So maybe they've built these no code agents. They've demonstrated some signal that it's going to work and there's value in it. And so those types of companies tend to take an existing tool, and they customize them.

So, it could be like a small to medium segment type of customers who have a, you know, a little bit more budget. They are more willing to invest in some. Infrastructure to house those models, whether it's again, like maintained by a development agency or they bring or kind of have their own cloud environment.

And the reason why cloud is important is because a lot of these AI tools that you're customizing, they live within. That cloud providers kind of like kind of space, like call it sandbox or whatever, like of different AI tools, which makes it easier to take existing models and then customize them.

But that also, again, requires you go development side. Like development agency side, then you hire these folks to actually customize these solutions for you. Or you hire experts to actually do that on your behalf. But again, like the, the premise there is that you're taking existing like Gemini, GPT 4, GPT 3.

5, and you're adding layers on top of it, and you're customizing that solution and then let's say you integrate it into different workflows so that you, you deliver the seamless experience. So that's kind of the shaper market. And the way that I kind of think about the. Makers are people who probably are not going to go there, although I've, we've made solutions within my organization as well from scratch.

So, if there are no existing models or solutions that kind of, you can even customize, you kind of have to build your own solution from scratch. So open AI, that's what they did. They build a transformer LLM type of system, and they build it completely from scratch. So, when you, in that kind of, again, then that's kind of, that's Very expensive.

These are the big enterprises, the IBMs, the Googles, the Microsoft’s, like they play in that space because that is a very costly thing to do. So most, most companies actually fall. I want to say like small to medium sized businesses fall into that cheaper segment because they see the value. And the value, the true value of AI comes from customization and actually putting your own data to work.

And you're actually building these systems more securely because you are kind of have you. Cloud provider and all of these security systems around that you can add rule-based controls. You have more control over your data that you're enabling that you wouldn't be able to do with subscription tools.

So, again, and you're kind of saying, like, well, Okay, that's all nice and great, but how can you help these organizations like do that again? Similar thing. You're going to go across your organization. You're going to look for use cases, but now that you know, hey, they do have an infrastructure. They do have it teams.

They do have development. They might have gaps in their skill sets, but that's okay. Let's identify what objectives you've set.  For your organization, let's look across the organization where you can deliver on those objectives. Maybe it's sales. Maybe it's marketing. Let's focus on those 2 parts of the organization and let's look at what currently doesn't work.

So, you're looking at bottlenecks and pain points. Maybe old systems and I've done that before where There have been systems built in non-scalable ways and for, for, or to scale, like sometimes it just makes sense to like. Completely shut down an effort to be able to build something quicker and more scalable.

So, you're really evaluating the product mixes the, the operational workflows. And then you're saying, okay, where can we make an impact? To deliver on those objectives. And then again, you kind of go how I've, I've mentioned that you give them use cases for AI. But you bring them along on that journey.

Like it's not just, you know, me as a consultant going back and analyzing, I can do that. But the key here and the key to change management with AI is bringing folks along on this journey so that they're in it, they understand how it works, they understand what goes into it. And then they become much more empowered and enabled in the future to understand, Oh, I see.

I mean, we could probably use AI to do that. Now that we understand what it's capable of. And again, that kind of comes down to practice again. You do these enough times, you do these, the POCs enough times that I think I could do that. Like, we need that. We have this need, you know, we've identified, and I think that's a really good use case for AI, and you can even go like a level further and say, like, we could use an LLM for that. 

I wouldn't solution 1st with the right technology for multiple different reasons. But as long as you understand, like, Hey, that's a capability that I can perform and it's totally useful here. You would again provide those recommendations in a different way. But again, with the lens that you're probably going to customize something in that workflow to bring that whole full-scale benefit to that part of the organization.

Yeah, you speak of empowerment. I think it's a powerful word.  And the idea that an outside consultant like yourself doesn’t know necessarily the exact processes or workflows that the employees or business leaders are using within the organization. So, to throw it back on them to analyze their own workflows and define with your suggestions on where AI could help in certain steps.

And if you're replacing some of those steps, that's great. You know, if it's two or three steps that you're replacing with AI, you've increased their productivity by 30%. And then sharing those results cross divisions, right?  When you talk about POCs, proof of concepts even internally you, I imagine you are building AI.

Yeah. Yeah.  AI heroes within an organization to then become advocates, right? To help really transform that business. And it has to happen from within and you're facilitating that. I see. And I think that's a fantastic approach. I call it, you know, building an AI operation. You know, how do you transform existing corporate business structure?

You know, at what point do you enter? Is it the C suite? Is it HR? Is it marketing sales? I mean, I believe you need to start somewhere and then grow through positive proof of concepts that have affected workflows, shown efficiencies, and then spread that through evangelism within the organization. 

And it catches fire. Right. I mean, that's how businesses transform. It's not necessarily from a top-down mentality.  I just don't believe that a CEO is going to be the leader in that charge because he is not involved in, he or she, I should say, is not involved into that, in that day to day and may not know all those steps within various departments.

Under their purview, and so it's really working with the middle management, right? And it's not necessarily bottom up, but I think starting in the middle is where to go. Is that, is that what you have found? Is that your entry point? You know, and how, how have you strategically kind of penetrated that that bubble of corporate structure to influence it? 

Yeah, and I think in a different phase, I would call them of kind of your discovery process. You're going to involve different stakeholders, and I'll tell you why. So first you want to understand the business objectives. You're not just going to again, toss LLM, like the question, where are you going to put it? 

Okay. LLM is great. You start with the business objectives, and you need to think about, To talk to the key staples holders of that organization who tend to be the C suite or managers of a smaller agency, you know, whoever, so folks who make decisions or who set those objectives or goals for the organization.

So, you can kind of understand and everyone can kind of agree. Okay. This is what our focus is. And this is where I think strategically where we want to go in the next year, 2 years, 5 years. When you're actually doing the work that I was kind of outlining, understanding, okay, well, how, which part of the business can actually deliver on that.

So, you're going and saying marketing, well, the CEOs and the C suites and the folks understand those processes that are higher level. They don't know the intricacies of that process. So, they won't be the right people to ask to give you like the detailed workflow or how that product works. You have to go to the SMEs.

You have to ask the folks who are actually, they're doing that work or the managers of to be able to look at either that product or that service or that kind of workflow. To be able to understand, okay, here are the different steps. Okay. If we were to improve it, and then you do a workshop with them to, again, they're the SMEs, they're the experts.

I'm just kind of the guide to like help them get to a certain state because no one can do it better. And they have. Also, they bring the lens of whether it's going to work in production, right? I can recommend, I can come back behind the scenes, work for on something and come back with a recommendation a week later.

And what am I going to get? They'll be like, that's not, it's never going to work. You don't understand our situation. So, it's better to build and enhance this together. With while doing education and while empowering them so you kind of create for the lack of better word requirements for what you want that product service, whatever to do.

And then when you go into the phase of proof-of-concept building, you are working with the tech and experts. So, you don't need at that point. You've kind of captured their requirements. You understand what the intent is. You understand how you could actually deliver value back to the organization. You do some R.

Y.  Estimates and then you know, you get approval from basically leadership to move, move that forward to the proof-of-concept stage. You build the proof of concept ideally in the shortest amount of time possible, because all you're looking for that stage is not built for scale is built to.

Give you a signal whether what you're building is going to work is going to it has even the promise or like some indication that it's going to deliver on that value. So, it’s very low expectations, but and that's why you have that kind of have to optimize for speed because the time you're, you're wasting on proof of concept is also time.

You could be investing back in, in focusing on something else. So, I think empowerment and bringing your folks along on, on that journey is going to be really key. And the types of folks you're involving different during different steps of that process really vary, but they bring different. Value or different information to the process.

So hopefully it kind of ends up flowing and working in the end. And then you decide following proof of concept. Is this something that we want to invest further in? And we want to actually scale. Do we want to put to put in all of the. Production level systems in place in order to really scale it.

And sometimes for proof of concept, sometimes you're like, well, where am I going to get data? It's not going to; I can't get that quickly. You need to jump hoops, get fake data. So, the name is escaping you, but like synthetic data.  So maybe you could purchase it from like a public data set, but it directionally, or at least mimic the type of data you would be replacing it with in production.

But just think outside of the box, figure out like the objective for that proof of concept is to just demonstrate that. It's going to go into work or has some promise of working if you were to scale that solution. So that's how you de risk that investment. Because once you get into that customization workflow, that's when things get expensive, not during the proof of concept.

And that's why I think it's key to optimize for speed and do it to the less Least fidelity possible. And sometimes we've built products with code. Like all we could basically demonstrate that the end of like that proof of concept was just code and basically developer saying like, imagine that this has a UI for this, and then this is the input, and then it will show us a part of the code where the input is.

And then here's the output. Cool. Sufficient. As long as it's again, demonstrating that some promise that that type of a solution will work when it's scaled.  But we are moving towards a no code world, right? Oh, yeah, yeah, making things easier. Yeah, the need for speed breaking out of that black box. I really appreciate the themes that you've been discussing.

And I could tell you're, you are very passionate and not only passionate, but well educated and well spoken. So, thank you so much for your insights. I want to give you a moment before we have to close here.  To plug anything that you're doing any, any destination in which people can find you and consult with you, happy to support any of your efforts.

So, plug away. Yes, thank you. So, you can find me on LinkedIn. That's my primary platform for now. My website's being built, but the company that I've founded is called Spark Change where I basically guide and advise companies on how to scale their initiatives. We do workshops. Basically, following that same process that I've kind of outlined here and I think as the website is being built, I'm also trying to, again, author content that I can deliver a lot of guides ROI calculations, like how do you actually do each part of this process, some of the best practices and frameworks.

So sign up to my newsletter if you want to hear more. As I'm kind of building that kind of library of tools and so once that's ready, I will be making an announcement. So if you sign up to my newsletter, you'll be one of the first ones to know.  Awesome. Well, thanks Svetlana. I follow you. I want to continue that and love to have you back on the show and to everyone listening or watching.

Thanks for tuning in and catch more of our Realm IQ sessions on your favorite podcast platforms like Spotify and YouTube. So please follow and smash that subscribe button. Thanks so much. Thank you.