Leveraging AI to Transform Business: Key Insights from Michael Lorberbaum on 

AI and generative AI are reshaping the way businesses operate, unlocking potential across industries. In a recent episode of *The Kelly Wendlandt Podcast*, Kelly dives deep into the world of AI with special guest Michael Lorberbaum, a seasoned expert on generative AI and its practical applications. Together, they explore the real-world impact of AI, focusing on how companies can extract real value from these technologies, from improving marketing content to revolutionizing finance departments.

 

This episode highlights key tools, practical applications, and real-world case studies, giving listeners a roadmap to make AI work for their businesses. Whether you’re curious about AI’s capabilities or looking for tips on implementation, Lorberbaum offers a wealth of knowledge to help professionals move from theory to action.

 

Where to Start: Key AI Tools You Should Know

 

In today’s rapidly evolving tech landscape, it’s easy to get caught up in AI hype. But as Lorberbaum explains, separating the hype from real-world applications is the first step to success.

 

“People want help understanding what’s hype and what’s real,” says Lorberbaum. While buzzwords like “ChatGPT” and “Anthropic’s Claude” are tossed around, these tools aren’t just tech for tech’s sake. They offer genuine value—if used correctly. ChatGPT, OpenAI’s most recognizable product, and Claude are two major players in the AI space, and businesses can subscribe to these tools for as little as $20 per month to start experimenting.

 

However, according to Lorberbaum, the real game-changer is Retrieval Augmented Generation (RAG). This fancy-sounding tool essentially allows companies to load their own content into AI, enabling it to generate highly customized responses. Want customer service agents to answer queries faster? Load your product manuals, and the AI will serve up solutions faster than you can say, “customer satisfaction.”

Generative AI Beyond Marketing: Practical Applications You Didn’t Know Existed

Think AI is only for marketers cranking out content? Think again. Lorberbaum points out that nearly every department in a business can benefit from AI, especially those dealing with mountains of data. Whether it’s automating workflows, improving demand forecasting in manufacturing, or enhancing customer service, AI tools can free up valuable time and resources.

 

As Lorberbaum explains, even companies with messy data can make strides: “Generative AI can handle unstructured data.” This means that, even if your data lakes are more like data swamps, AI can wade through the muck and start making sense of it. But before companies get too ahead of themselves, Lorberbaum stresses the importance of setting clear goals. “You want to be able to tell the AI which of these [data sources] they should rely on with the greatest degree of confidence.”

 

AI isn’t a magic wand—it won’t clean your data for you—but it’s a powerful assistant that can help automate repetitive, time-consuming tasks. For example, finance departments can benefit from AI tools by automating routine calculations, reducing human error, and even forecasting financial trends. Manufacturing, too, can leverage AI for everything from maintenance manuals in the field to streamlining supply chain communication.

Starting Your AI Journey: The First Step to Business Automation

So, how do you actually get started with AI? Lorberbaum emphasizes a two-pronged approach: start with small experiments and always keep ROI in mind. It’s easy to get lost in the AI rabbit hole, but businesses should focus on solving specific problems first. Whether it’s reducing admin work or improving customer service, start with a problem-centric approach. This will ensure you’re not just adopting AI for the sake of it, but actually adding measurable value to your organization.

 

Lorberbaum suggests thinking of AI as a junior employee. “You want to tell the system which of these [data points] to rely on. Treat it like a new hire—you have to train it,” he says. AI tools won’t take over human jobs, but they will make routine tasks quicker and easier, allowing employees to focus on more strategic initiatives.

Unlocking ROI with AI Tools

One of the standout takeaways from the podcast is the potential ROI from using AI effectively. “A good starting point is expecting to gain 10-20% more bandwidth,” Lorberbaum states. By reducing time spent on administrative tasks, AI can help companies free up significant resources. He gives the example of demand forecasting in manufacturing—one of the most data-intensive business functions. AI can analyze historical sales data, project future trends, and streamline ordering processes, leading to reduced costs and more accurate decision-making.

 

This efficiency extends to content creation as well. Imagine being able to instantly translate marketing videos into multiple languages or summarize complex reports at the click of a button. AI tools like Notebook LM allow businesses to do just that. Lorberbaum envisions a future where AI can help turn reports into podcasts, complete with conversational summaries—perfect for busy execs on the go.

AI Generated Transcript

Kelly Wendlandt (00:00.519)

One and we are live with Michael Lorberbaum. Michael, how are you this morning?

 

Michael Lorberbaum (00:05.259)

doing great. It is wonderful to connect with you.

 

Kelly Wendlandt (00:08.802)

How is Central, you’re in the cities in Minnesota, right? Is that right?

 

Michael Lorberbaum (00:13.899)

I am in the cities. We’ve got some wonderful weather this morning. So for early October, it’s a nice day outside.

 

Kelly Wendlandt (00:21.004)

Doesn’t get any better than this actually. It’s gonna be a high of 81. It’s gonna be too warm for me, but it’s gonna be very nice.

 

Michael Lorberbaum (00:27.657)

I wholeheartedly agree.

 

Kelly Wendlandt (00:29.398)

Yeah. So, Michael, you are an expert in AI. You work with companies to help them understand the technology and the technologies that are out there related to AI. you also help with actual practical use cases. How do companies use the tools? Do I have that right?

 

Michael Lorberbaum (00:53.155)

That’s exactly right. And helping companies think about how to really get value from a lot of these generative AI things that are all over the news. And so people want help understanding what’s hype, what’s real, and can I really get an ROI from something as simple as chat GPT?

 

Kelly Wendlandt (01:13.654)

Yeah, so let’s start with that because I do have people around me that talk about chat GPT and they talk about rag and they talk about some of the other tools. So what are the primary tools that are being used in business that companies are trying to leverage to to make their companies better? Whatever better is.

 

Michael Lorberbaum (01:17.08)

Yeah.

 

Michael Lorberbaum (01:35.929)

Great question. So the primary tools these days are really OpenAI’s ChatGPT and Anthropix Claude. And as a consumer, any of us can really buy our own subscription to any of those and get familiar with it. They’re $20 a month. You mentioned RAG, Retrieval Augmented Generation, and that’s nothing more complicated than loading some content into the system.

 

so that it can use it as additional context to help respond to you. And so if you are a author and you want an AI to speak the way you do or to write the way you do, you can load up your content and it will write in your style. Or maybe as a company, you have a bunch of technical content that your customer service agents use. You can load that up.

 

And then that AI tool can respond in real time with technical support for either eating their existing support agents or giving advice directly to customers.

 

Kelly Wendlandt (02:47.808)

And so that’s part of the reg. That would be under the reg or the anthropic or the, yeah.

 

Michael Lorberbaum (02:53.463)

That’s under the rack. So let me give you a little bit of an example and.

 

When I think about real use cases, I think about it in three flavors, so you can get a lot of this off the shelf. You can go to Anthropics website. You can go to open AI and just buy a subscription and get answers. The next level of sophistication under data integration is using things like rag where you’re giving it context based on your own information. It’s a little bit harder to get started from an enterprise approach.

 

but you can definitely do it. They’ve actually gotten a lot better quality responses using things like RAG. And then some of the biggest Fortune 500 companies and others are building their own custom models. And that’s an even higher level of complexity and sophistication. But what’s cool is even in the early phases of off the shelf tools, you can get value in almost every business function. So whether a marketer is trying to

 

improve their own content or someone in IT is trying to get help debugging a program. It’s all possible with some of these off the shelf solutions.

 

Kelly Wendlandt (04:13.16)

What I see data next to it and that’s something that at Logisol, my company, we are seeing a lot of activity in our data group and we have a lot of companies that say, I met with the board and they want to know what we’re doing with AI. And I’m still in the meetings I have talking about how do we get our data clean enough that we get information that’s useful to our executives.

 

And so how big a lift is it for people? You know, are people ready for AI or how in your mind is it concurrent or do people need to get their data straight? It’s, know, at some basic level before AI can be meaningful to them.

 

Michael Lorberbaum (04:57.825)

generally recommend testing because what makes generative AI so much more powerful than the types of AI we had before is it can handle messiness. It can handle unstructured data, so I would recommend loading the data as you have it and trying to evaluate if you get quality output back. If you get into some of those more sophisticated use cases in that middle.

 

You can actually put workflow steps in that clean up some of the messiness that exists. So as an example, if you load up all prior customer support interactions, it may have a conversation about, you know, a customer’s pet in there. You can take various steps to remove that type of conversation so it never gets surfaced back as a part of an answer. And so there’s

 

different things that you can do to start to automate some of this data cleaning with generative AI tools.

 

Kelly Wendlandt (06:07.254)

When it comes to like, you know, groups like finance, for example, and there’s, there’s multiple databases and data warehouses, data lakes that have been created over the years. And at this point, there’s this, in theory, the same piece of data in 10 different systems, but that piece of data is actually not the same value in all 10, 10 databases. And so there’s discussions around which system is.

 

the system of record and you might have executives that have their own pet data lake that they think is the right one. Does AI help in that data cleansing or can it referee which database is the actual data source of record or not? Is it still the humans having to fight it out in a boardroom?

 

Michael Lorberbaum (06:56.537)

It’s still the humans fighting it out in the boardroom. I recommend that you tell the system which of these sources is your source of truth. Which of these things should be the primary source for a given data element? Yes, this messiness certainly exists. It exists across many enterprises. But to help improve the accuracy of what you get out, you want to be able to tell the program or

 

Think of it like a junior employee or intern. You want to be able to tell this person or AI which of these they should rely on with the greatest degree of confidence.

 

Kelly Wendlandt (07:36.554)

Okay, yeah, and we have a process and we’re going through that where we take people through determining and agreeing on, know, it’s the alignment, the whole alignment process prior to getting to this technology, this technology phase. So that all makes sense. So assuming the company gets to that point, they’ve decided, okay, here’s the data of record that we are going to go with. I see marketing groups right now using this technology to create content, work.

 

Michael Lorberbaum (07:46.531)

That’s right.

 

Kelly Wendlandt (08:05.406)

some customer service and those kind of things. Are there other areas where you say AI can come in and make a difference? Specifically like a finance group or in manufacturing, know, like the heart of some of these businesses. Where does AI fit in?

 

Michael Lorberbaum (08:22.905)

100%. And so what I’ve seen is almost every business function has complicated work that could be automated or things that are data intensive, computationally burdensome, and you can use AI to help across the board. So I think you mentioned manufacturing as an example, you could, as a simple case, load up all of the maintenance manuals.

 

onto a tool so people have access to all the technical details in the field wherever they need. There are people who are now doing things with supply chain where they’re automating different communication steps so you don’t need a human to close the loop between systems. And then the more sophisticated companies are improving their own demand forecasting tools to influence their own ordering. So.

 

You can kind of feel the progression of sophistication, but even in those early phases, lots and lots of companies are getting value. One estimate I saw earlier this week is if someone is really using AI to automate their work, they could likely gain something on the order of 10 to 20 % more bandwidth because they’re removing some of the admin work from their day to day, and that would be a.

 

typical business professional. So obviously for some functions it may be more, some jobs it may be less, but 10 to 20 % is a good going in assumption as you’re starting experiments.

 

Kelly Wendlandt (10:02.453)

How does a company, if they want to address, they know they have an admin at that issue. They know that they have their people are doing too much work that’s not critical to what they consider to be the heart of the business, generation, customer service, whatever it is. Where does a company start? Is it assessing? it just bringing a tool, buying a tool and bringing in, where do you suggest people start?

 

Michael Lorberbaum (10:26.709)

Love the question and it fed me almost immediately to this and that is I think of two different exploration exercises and you can use a workflow like this to get some of the people doing the job to brainstorm possible ways these tools could be used. And so you can start with a problem centric approach. I have a problem. really want to solve it and you go through brainstorming solutions evaluating how much value it creates.

 

the cost and then running some experiments or something that starts with capabilities and doing something very similar. What is really fun is you start with these six very common AI capabilities and when you start there, the ideation really starts. Getting into some fun areas and spaces so.

 

We’ve all likely seen chat GPT generate images and respond to questions. That’s what people have expected, but now I think within the last month or so we’re starting to see some real advances in audio, whether it’s the conversational AI with chat GPT where you can have what feels like a real time conversation or what came out from Google of notebook LM.

 

So you can load a bunch of data and get a recording and engaging conversation of a couple people talking about whatever you loaded. So imagine you get a whole pile of reports at the end of a month that, you know, takes a long time to sort through. You could load that into something called Notebook LM and it will make a podcast for you of two people.

 

conversationally talking about the highlights, the key insights from that data set. But I use this, this is one of the more powerful tools for people to brainstorm ways that the capabilities of generative AI can add value to their business. And just coming up with both simple and complex applications for these various capabilities.

 

Kelly Wendlandt (12:48.002)

Very cool. And for those of you who are listening to the audio version of this podcast, have, we have some slides up and it’s, it’s, Michael has potential applications, simple, complex, and you go through an exercise of, of sitting at a room and talking through details and figuring out then how do we plug AI into, to, you know, help with whatever the problem is.

 

Michael Lorberbaum (13:11.993)

And so for the listeners I would just highlight some of the real fun capabilities that have come out involve creating content. And when you think of content, it’s everything from text to video, but also code. You can enhance that same content into as many languages as you could imagine. So think about instant translation of your customer videos, whether it be marketing or self help.

 

translated into someone’s native language. You can summarize key insights, review large historical context of data. You can do so with technical or non technical information, but also create connections and insights across very disparate datasets. So if you load it as an example, all of the historical. Marketing summaries from the past 12 months.

 

and asked it where are there substantial variances that we should explore as opportunities, it can do that and do so very quickly, more efficiently than a human could.

 

Kelly Wendlandt (14:22.23)

Just based on what you’re saying, I’m traveling in the next six months to a number of different countries. And what you’re telling me sparks a lot of curiosity around if I could use hearing aids that AI in, you know, it probably won’t be in six months, but I bet in two years, AI will be able to translate Japanese for me as they’re speaking to me real time or close to real time. And I can, I can probably respond. We’re probably not too far away from that, which is pretty cool.

 

Michael Lorberbaum (14:49.081)

We’re not too far. I would say in six months you might get close. So if you’re a Google user, Google has an audio translate already that is not quite real time, but you can listen in a foreign language and it will translate. Another hint for those interested, you can even take a picture of a menu at a restaurant and it will instantly translate the menu into your native language making.

 

ordering a lot more pleasant experience.

 

Kelly Wendlandt (15:22.466)

That’s awesome. Michael Lorberbaum, thank you so much for your time. This has been fascinating. For everyone else out there, thanks for listening and you are watching the Vodcast.

 

Michael Lorberbaum (15:34.457)

Thanks, Kelly.

 

Ready to Take Your Business to the Next Level? It’s Time to Embrace AI

AI isn’t just a buzzword—it’s a tool that’s reshaping how we work, communicate, and grow our businesses. As Lorberbaum points out, whether you’re in marketing, finance, manufacturing, or customer service, there are practical applications for AI that can streamline workflows and increase ROI. The key is starting small, experimenting with off-the-shelf tools, and finding ways to tailor them to your unique business needs.

 

Now’s the time to start your own AI journey. Subscribe to *The Kelly Wendlandt Podcast* for more expert insights from industry leaders. You don’t want to miss out on the future of work—make sure your business is ready.