Beyond the Hype: How Small Businesses Can Actually Leverage AI

This is I think AI sometimes gets labeled as taking our jobs, but it's no more than any other new technology.

It it should have a personal touch your company always, but AI help respond quickly in the hours that you are resting, taking time with family.

So what sort of tools are people using or what are you recommending? What are the some of the big wins you've had with

Um I think too much um attention goes to content creation and uh things like okay lead generation uh AI by voice, but I think it's more important to use the AI in daily tasks that are not directly visible to your client.

AI is it hype or is it help? Good day and welcome to the show.

I'm your host Pete Cooper and our topic is AI hype or help. Today's guest is Ellen Enkhardt. She's with a company called Sensei, her own company, where she goes and helps companies streamline their workflows. In other words, how businesses operate through the use of AI and other digitization tools. So this is going to be quite interesting because we're going to get to the core of how AI is helping businesses today. So let's get into it. Hi, welcome and welcome to the show. Uh we know each other through a mutual networking group. I'm going to talk a little bit about that. Um but our theme as you know is about AI and you working in this field as well.

So for the audience, can you tell us a little bit about your business? Um so we get some background.

Yeah, um I help small and medium companies in Spain and other European countries with their um lead generation and also to structure their internal processes. It can be administration, finance, to make their work easier.

Yes. So, lead generation, if anybody's not in the market in the marketing field, what is lead generation?

Um, lead generation, to yeah, say it in an easy way, is acquiring new clients.

Yes. A new client is a lead, right? So,

Yeah. Mhm. And so, there's I guess there's digital online techniques for that.

I wonder what the world did before the internet with lead generation. I guess they would stand out there, you know, with a billboard.

Yeah, or they were just calling at the door for B2C, and take their car and visit the the business. Uh

Take their car and visit the business, right? Going through the front door. Can I speak to so-and-so? And

And a lot of cold calling also, of course. Taking the telephone. Things young people do not like to do very much, you know. Yeah.

Ellen, so where where are you located now? And and what's your personal background? I mean, where do you come from?

Yeah.

I recently moved to the Netherlands in January, and before I've been living for almost 20 years in Madrid. And before I started working in Paris, and I'm originally from Brussels in Belgium.

Okay. So, you've sort of done a excursion and come back.

Yeah, yeah. Up south, and then back north.

Yeah. I mean, so the Netherlands and Brussels, are they one and the same place? I always forget country versus town. Are they Which is which?

Other countries, but the towns are quite near, yeah. Neighboring countries.

Yeah, right. Well, I can tell from your lovely accent that you're somewhere in Europe Uh and I guess it's sort of across the channel from the UK.

I have been there that part of the world myself. I was in The Hague with a friend visiting a friend gosh 20 years ago now. It was really fun and of course there are huge bike culture, right? Everyone rides bikes around.

Yeah, true.

Yeah.

It's very important.

Yeah. So we touched on

like you have a problem.

If you don't have a bike you have a problem?

Yeah, and if you can't ride it it's even a bigger one.

All right. So how are you with bike riding then?

I'm okay. Yeah, and my family is too.

Yeah, yeah. So and then we so we talked about lead generation and then the other thing that you do is helping streamline businesses, right?

I mean that that I think that was something like So tell us a bit more about that.

Yeah, so a lot of small companies in Spain they're working as they did 20, 30, 40 years ago because it's family businesses and it's almost like people come in have to wait and I look at how they are working and we try to structure things and um yeah, make sure that they don't lose a lot of time on attending people that people don't have to wait and in a first place we use the typical digitalization methods as a booking page and afterwards if things are more streamlined, if we see that they can receive all their email messages, WhatsApp, Telegram, whatever in one place then we will look how we can apply AI agents to uh make it easier for uh their personal to work smoother.

Yeah, we're going to talk a lot about AI as a theme. So, before we get to AI

uh do you have any examples that uh that you could share without necessarily specifics about who or or when or why numbers, but can you give us an example of where it it it worked before AI cuz we're obviously going to get into that.

Yeah. Uh could you repeat the question because I'm a bit lost here.

Oh, okay. So, I didn't want to sort of get straight into AI.

Yeah, I see.

The the thing that you do applies without AI, right? Because you go in there and you help companies.

I I was trying to give you could give us a bit of an example just to illustrate it without you know, you don't need to name name names, right? But just give us a more detailed account of what sort of things you know, what you've done.

a test case or something.

Yeah.

For a real client I I was doing.

Yeah. So, the first thing I do is an audit and um that's talking a lot with uh the CEO of the small companies and with the uh personal and I uh look at how their uh day-to-day work looks like and uh what things could be done faster or um also yeah, where um certain processes are not streamlined and and have to be like modernized.

So, I'm thinking of Henry Ford was uh the guy that made, you know, made production very uh very much a thing when they're building cars because before then every car manufacturer made everything totally bespoke, right? They would they would build the car and then they would build another car and then and they would start from scratch pretty much every time.

But he's decided that hey, we make everything exactly the same and did time and motion studies and and had the you know, the moving conveyor belt with the cars on it and people would come in with specialist skills and add the wheels and add the panels and so he was able to make I think he's he's I don't know whether this is true or not, but the the urban myth is that he told people that yes, you can get the Model T Ford in any color as long as it's black. So I just standardized it, right? Um I mean that's a classic 20th century kind of example of of that, but that I guess that that streamlining approach is still very much alive today, right?

I mean, things are always changing, businesses are forming doing different things. They tend to do things bespoke to start with and then they they get more and more efficient as they well, hypothetically, they get more and more efficient as they scale and different companies come in and try and find things better. And in order to compete, they need to be efficient as well.

Mhm.

Yeah, I think that the good thing about doing an audit and looking what should be changed in the company is very helpful because I'm someone who's outside their day-to-day business and and it's always good to get the outside view because people are working a lot, doing a lot of hours in those small companies and they don't realize that they're like turning around in a wheel and that they cannot grow when they don't have their processes yeah, streamlined.

Yeah, exactly. That they're stuck in the they're sort of just going around and doing the same things and I guess working harder, not necessarily working smarter. Like, how can we trim this down this process down?

So, mapping out the process that that's currently there, I guess is part of it. Um and then um suggesting, you know, streamlining methods and techniques that they may not be aware of. I mean, it's it's particularly true with AI emerging so quickly now, but the people don't realize, you know, what's about what's available and what things can be done.

Yeah, that's that's important that it's not the client who says, "Okay, I want an AI agent for, for example, um incoming messages. " I think it's important that the first part of the work is done by us, the AI strategists, and that afterwards, together we look at what the possibilities are because most people who want to work with AI

Mhm.

don't realize where they can make the most gain of it. And And the same thing without AI, it's important that someone from outside has a view on your business and and look how things could be done in another, a better way.

Yeah. Well, let's talk a little bit about AI AI then. So, back to those two areas. So, where where are people missing opportunities with AI today?

Um I think too much um attention goes to content creation and uh things like, okay, lead generation, uh AI by voice, but I think it's more important to use the AI in daily tasks that are not directly visible to your client, but that help to do you the administrative work, for example.

Mhm.

Yeah, so um I mean, are you using LLMs? I mean, is that one of the key things that you're bringing in to to the workplace, which has not been there?

Um no, I think the LLMs are getting too much attention. Like everyone is using it for their own, but there are other things AI can do that people don't realize all the time. So, uh most businesses think of, "Okay, I want to work with AI, LLM. So, I am talking with a chatbot, and it's helping me to generate content. " Uh it can be for social media, but also um yeah, to create doc- -uments for clients, uh invoices. And I think that's only a smart part of what AI can do.

So, this is why for me it's really important to see their really important tasks, and and what's like the core business of the company, and and not use AI uh yeah, for to communicate with the outside world.

So, you're not in the business of sort of creating content with what you do for companies. You can let the social media uh managers and the marketers do that. You're much more on the operational side of the business, would you say?

Yeah. Yeah, I would say. Mhm.

Mhm. So, what sort of tools are people using, or what are you recommending? What are the some of the big wins you've had with with some tools that that you've found helpful?

Um what has been useful for my clients is AI chatbots where people could ask their FAQs, let's say, so that the FAQs are being answered by the AI

Mhm.

but with the input of the last 100 or 200 email and WhatsApp conversations that real people had with them. Uh another one also gathering of documents so that the administrative personnel doesn't have to uh go back and forward with the client before they can make a proposal. Um that has been useful and also uh tools to make uh price proposals

Mhm.

depending on for example in Spain um one document that the government edits with prices for uh car sales and and buys so the cost is always the same for the same model and the same uh year of uh of the car, you see?

Mhm.

And and and then we can add the company's cost to do those uh the administrative mhm s- selling and buying so yeah. Mhm.

So we were talking about this the other day. I thought it was a really good example. So this is where uh somebody wants to bring a car into a European country.

Yeah, importation, yeah.

Mhm.

to import

And say I'm let's say I'm an individual or small business and I want to move myself over to to Europe and I've got a car here in uh in America and I I to have it over there. I can ship it over there, right? But there's an a administrative burden of doing that, right? Which is quite complex. Yes.

quite complex. And then we made the algorithm ask you, for example, are you a car owner in Spain and the seller and the buyer are in Spain? Is it someone Are you buying a car from within Europe or outside of Europe? So and then it it says what documents you need for your specific situation and it also creates the proposal with the cost depending on all those variables.

Yeah, I this is

a lot of time and it's it's nicer also for the client that he doesn't have to call and to email the office who does this.

Right. I think you explained this to me the other day and I thought it was a really good example where the chat the chatbot was a much better tool for for gathering information from the client and because I think the way you explained it is they don't always have everything right there like on the desk sitting there. They have to go and find this document, get that thing, get this piece of information. And you know, what are they going to do? Are they going to they're going to ring up and say, "Oh, I've got this document.

" And somebody's not there. They're out to lunch, you know, then they come back and then they call back and you have this ping-ponging just to get one document over and you might have 20 or 30 that need to be found. So with the chat, you can Oh, I found this one. The chat will then pass that on. Uh maybe ask some more questions to clarify both ends. It kind of acts as a 24-hour monitor um to to to try and establish communication.

Yeah. And what we also did was link that information with a CRM.

Mhm.

So the administrative has all the info even if the conversation started, for example, on social media, on WhatsApp, or by email.

Mhm.

So, it's it brings everything together to work uh more smoothly. And yeah, they before uh the chatbot and the CRM, they were working on their messages that they received from Friday like 2: 00 in the afternoon because people leave early on Fridays.

Mhm.

They don't work in the weekend. And then on Monday, uh they were working till 1: 00 in the afternoon just to catch up with all the messages. And they told me that they uh did not convert those leads. So, this was a thing I thought that should have been, yeah.

Oh, so it's even to the point where because they're not responding quickly enough, they're losing business.

Yeah.

And I think AI can help small companies uh to work more agile and to lose

Mhm.

less business to the bigger companies who do have people who work in in the weekends or who have already implemented AI.

Yeah. It's interesting how I'm observing a a shift from an old an old-fashioned way of working to a new way of working. And I've witnessed this myself where uh it's it used to be the case that people went to the office and they did their 9: 00 to 5: 00 because they needed to get everybody in the office together to work together for things. So, if you want you were a customer, you would turn up there from 9: 00 to 5: 00, too.

And then you would be part of that whole system of of face-to-face meetings, uh 9: 00 to 5: 00. That's all changing now. I mean, you and I both know this very well because we work globally, so we're across time zones. And we work remotely. There is now this great difficulty with our type of work to have this kind of 9-5 face-to-face under one roof type of arrangement. It's It's also very inefficient uh because you might have to have several meetings before you get to the you know, in your case. Uh and this kind of solves that because it brings together this this intermediary to bring together

Mhm.

people who are remote, working different times. Um

Yeah.

and and sort of promoting the solution, which is ultimately what the client wants as a solution um much more quickly, much more easily, reducing all that friction. So, it makes a lot of sense to me, yeah.

Yeah. Yeah, I think people buy from people. They don't buy from AI, but they would like um stop buying from you if your AI is working badly. So, I think it's important to have both.

Mhm. Mhm.

Yeah. It It should have a personal touch, your company, always, but AI helps to uh respond quickly in the hours that you are resting, taking time with family, and yeah.

And then you can go and re-intersect that. I think that's really important, too. So, the how is AI functioning in this?

It's It's the backbone for the chatbots, right? So, there's an interface to the various LLMs, whichever one you choose, through an API that drives the chatbot. So, there's an inter So, it's acting as an interpreter. Um and then it's also working where it's um I guess privately, it's looking at the internal information within the company uh and saying, "Okay, I can put that over here as part of the CRM. I'm going to pull that from the CRM. " So, there's to me there's these two piece places where AI are being used. Is that right?

Yeah. Yeah. Yes, it is.

Okay.

That sounds good. That sounds like interesting application.

What I particularly like about it is how it's it's adding value and it's not trying to replace humans.

No, it's

It's trying to streamline um a new way of working. So, I think it's really interesting and I think it's a very good application of AI.

Yeah.

Um

And maybe we could see it also as working with AI as a competitive advantage against companies with the same size, uh same services that don't use it. So, if you're an early adopter, let's say, and you use it well,

Mhm.

then it's the other way around.

Your employees will be able to keep their job because the company will go well and and the benefits of the company will make it easier to pay the salaries at the end of the month.

Exactly. So, we rolled out a solution for a a clinic in Australia. So, it's Allied Health Physiotherapy or Physical Therapy. Um they have 40 staff and we rolled out a solution to do uh insurance reconciliations. The CEO told us that they're saving now 2, 000 dollars a month. Um which for a business like that which has very low margins, so it was really affecting their profitability um in a very positive way and making them sustainable.

And when uh she had to sort of wind down the virtual assistant that was doing this, I said, "Well, it's a shame that that person doesn't have that work. " And she said, "Yes, but you saved 40 jobs. " So, yes, I think, you know, some of these inefficiencies may, you know, have some fallout. Um it may mean that the that person who is doing that I guess routine job will be replaced. But if you don't do it, then the whole business is a jeopardy, and that's a far bigger problem. And let's face it, this is not something new. Businesses are always looking to make efficiency drives, and sometimes the a role that's there is no longer needed.

And so I think what I'm going where I'm going with this is I think AI sometimes gets labeled as taking our jobs, but it's no more than any other new technology, you know?

not. Yeah.

I think it it takes away the more boring parts of the job.

Right.

And it it it makes your job more interesting, but on the other hand, I can understand that young people have like more difficulties to start working because it would like um replace more the junior, the easier tasks.

Mhm.

But I think those things with time, it it will recover. Yeah.

Yeah. Yeah, I mean we still need new smart people coming in, young, smart, dynamic people coming into the junior roles.

It's just that they're not necessarily coming in and doing that particular task, they're doing something else.

Or or maybe they don't have to be like the first 5 years doing boring jobs before we trust them with more complex tasks.

Yes, right. That's a good point. You You brought up the word trust, uh and that's I think something I wanted to explore with you today is how much are we trusting AI with our businesses and and with with things? What's your opinion on that?

Yeah, that's an interesting thing. I have the idea that outside of Europe and the US, uh like countries in Latin America, uh for example, in Asia, they're adopting AI very fast, and and it seems like they trust it.

And in in Europe and the US, like we use it a lot, but we don't really trust it.

Mhm. Mhm.

Yeah. So so that's a bit of a contradiction. That if you have people Yeah.

That's an interesting observation. I The only witness I have to that is in Dubai and um the Middle East, particularly Dubai, um because I've visited there and how much they committed to AI. I was there over a year ago and they said they wanted to have a million prompt engineers uh through their system. I'm not sure the time frame, but it was a lot of people go through their prompt engineering program.

And I just saw something on LinkedIn where one of the the Saudi princes, not Saudi, one of the Dubai prince princes said, "We want the whole private sector to move over to AI in the next 2 years. " Um to use AI efficiently. So they're really quite aggressive with you know, their plans for AI. Yeah. It's funny, isn't it? Why do we have skepticism and they're just all in?

Yeah.

Well, I know from my personal experience having had junior staff come on straight out of college, they go straight to the tools. And sometimes they tell me sometimes they don't. It doesn't really matter cuz I encourage it anyway.

But then they use the tools and they generate something for me, a document, and it's full of problems and errors. And And I say, "Well, look, you can't You can't just go to an LLM and flick it to me. That's of no value to me because I can do that myself. You have to be You have to make sure those documents are 100% before they leave your And this is still This is still I'm still not getting that. I have to say. I keep reminding them, but I still don't get that. I mean, glaring errors that I see, you know, where the heading says 12 steps and there's only six steps actually listed.

I think the the LLM's that that are being used are not very good with things obvious things like the numbers that are misaligned, right? Like we seem to be a lot better at just oh that number doesn't match with that number.

And very important with the LLM's is what you put in is what you get out. Yeah.

Yes, exactly.

If you don't put quality in it, you get rubbish.

Yeah, yeah, yeah. So I say to them they have to read and understand that every single line of every document that they create and sign and get it signed off before it comes to me. But I'm still frustrated that it doesn't happen that still doesn't happen.

And I think this is where I think having a bit more experience in the real world gives an advantage. I think I'm seeing a shift from um I guess a classic example is a software developer comes in with straight out of college with all of this you know, great new skills that they've just acquired in college and now they're kind of given a big salary and trusted with big responsibilities. Um uh yeah, paid a lot of money especially here in the US for all that you know, unique skill and knowledge. Um I don't see that's necessarily the same anymore. I think we're starting to shift back a little bit towards people that have experience because you can see where AI is going wrong.

And then you can spot it and so people with a bit more experience but world knowledge

be able to spot it.

That's right.

If you don't have the experience of you don't or you don't think and this this is also the importance of education. Education will have to change but not completely because if if people stop thinking then the AI output will get worse. Yeah.

Yes, the AI output will get worse because they're not catching those things. I mean, I've had my LLM say to me, "Oh, good catch. " You know, well, that was pretty obvious to me. Why wasn't obvious to you?

Um I use it a lot for, you know, back and forward to brainstorm to I mean, I'm I'm all in with using it, but I don't trust anything that it creates. Everything Everything needs to to me has to have outside references and I have to understand the logic 100% and do the numbers myself. Um and it makes its fair share of mistakes, but it is still much much faster and much higher quality than I could do without it. Hence, that's why I use it.

It's an accelerator. I think AI works well uh in cases where you say, "This is the data you have to use and do not use anything else.

" So, for for some types of tasks, also in the Netherlands here, they were um at a university where I went to um yeah, talk with the students. They said, "We had a um like a a lawyer thing with with laws a typical education and uh they said like it it doesn't it's not useful anymore because the AI does the job of the secretary of justice a lot better than than the people can do it. So,

Yeah. Well, that's a bit scary, isn't it? The AI is making I think that's the That is definitely a scary future where AI is making judgments. It's already happening. I mean, there's a you know, the recruitment side of things is is happening.

You know, you have to go through people that are find trying to find work. They have to go through certain gates that are AI AI is creating for them. Yeah.

I think that can be complicated in HR, right? If you're looking for a job and you don't really know how I mean, I'm honestly I wouldn't know what types of AI they are using and how I should write my yeah, curriculum to get the job. I I wouldn't know.

No. No, I mean And and who's to say that that when they prompt up or they create an AI agent or a gateway that they've put it in there correctly. You know, that who's to say that they know how to use AI properly. I don't think they know.

So, people who could be great candidates might be getting gated out because we don't know how to use the tools properly. I mean, nobody knows how to use this the HR tools properly yet. I'm sure there's companies out there professing to do this well. Um but who else I guess there's a whole problems with the whole HR recruitment industry is it is kind of out of control and it kind of does what it's own does its own thing. Um and it's very frustrating for anybody who's actively seeking work. You know, to have to put so many applications in and put go through so many gates to find out there's actually not even a job at the end of it. They're just they're just trying to find candidates.

Yeah, I don't know if it works the same in the US, but in Europe, there are a lot of public employment companies that help you like to write your resume and and and things like that, but then they should like help you to to pass the the AI filter. In the future or or already now.

Yeah, yeah. I think the same thing happens in the investment world. Uh Uh So, you know, through through Skilling my company, we we raise raise capital for ventures. And there's a lot of uh there's a lot of people out there. They fall into several categories. There's the genuine investors, of course, the real people that are going to write the checks.

Then you have these intermediatories who can set up funds and do other things. They They're definitely gatekeepers. They won't allow you to access. They don't know have their own sets of criteria. They're not so bad, either. The The other ones are those that want to be investors and want to be treated like an investor, but really are not investors. They're the time wasters types. Um and, you know, again, there's no industry to check that. And then the final one would be those that I guess sharks, you know, they they're saying they're going to help you raise capital, but they're actually positioning themselves for some kind of a, you know, a cash payout, you know.

So many thousands of dollars, $20, 000, and we'll get your pitch deck done.

Yeah.

You know, they that I would say that's the shark sort of behavior um that happens in in the investment world, totally unregulated. Anyhow, we're getting a little bit off track, but

Yeah, yeah.

And back to the AI world, I mean, are you seeing other emerging ways that that AI is um is playing into your work?

Um yeah, probably when I was working in marketing, that a lot of com- smaller companies, before they would contract someone to do their social media, like 12 years ago, I started as a social media manager, and I'm talking with people from that industry, and they're saying it's becoming really complicated to maintain their clients because they're like, "Oh, but you use AI? " Of course I do. Why wouldn't I? I mean, it helps a lot. And then they think they can do it as well as someone with a lot of experience. And also, um, design is becoming more complex to convince that uh a real person can do it better than than AI. Yeah.

Yeah, I think that is I was speaking to someone on Tuesday, so 2 days ago, exactly about that point. In the marketing field, she was finding this a constant frustration that people would say, "Oh, yeah, I just put this and I created this logo for for us. Let's just use this. I did this in ChatGPT or whatever. " And she looks at it and she goes, "Well, yeah, but that logo is has problems here, there, and it So, she's already on the back foot, you know, with the client. Now, the client's thinking, "Well, it's fine. I'm going to use it. It's I've done it for free, and you want to charge me for that. " Can't seeing the value difference, right?

No. Yeah.

Yeah.

And I mean, how do you I mean, how if if ChatGPT's generating something for you, which is self-reinforcing, you know, you ask it you know, to do this, "Why is it good? " It'll tell you why it's good. And then you're leaving it up to, you know, the agency to say, "Well, you know, in my experience, that logo isn't the right match for these reasons. " And also just How does an AI model work? It looks out at what's in in the digital universe and averages things, right? It's just a predictor, right? It's got a bit of noise injected, but ultimately it's looking for something which is resolving to the norm, the middle, which is probably pretty close to something that's already there. Yeah.

That's one of the I guess we could touch on to this this one of the things that's emerging for me with AI is that fundamentally it's just an averaging of what is already out there. So

exists.

Yeah, so

really creative

Yeah.

you still need people.

That's right. I mean

AI is efficient but not creative. I don't find it creative. I The LLM models are like um sometimes a bit annoying. It's like

Yes.

Yeah.

There's some books that have been written about this type of future, this dystopian type future where um I I can't remember the names of the books but everybody is kind of uh going about in a very monotonous that we're all just you know, going around and doing the average of everything. No nothing outside the box and this is where AI is I think is taking us. Everything's just an averaging of everything, right?

Yeah.

But it should not be like this.

As we were talking before, you could also use it as a competitive advantage and if you have good people you have people that do their job well and AI can help them uh work more efficiently and and do the job in an better and faster way

Mhm.

Yeah, it's reinforced.

Yes. It's interesting how technology you framed it very nicely and I have experience in a completely unrelated field but you coming from that part of the world would understand this. Uh 12 years ago I was interested in e-bikes. This is bicycles with electric motors in them.

Yeah.

And over the years I've seen people be resistant to it. They would say things like you're cheating.

Like I'd ride the e-bike and they'd say you're cheating.

And then I didn't like that because that's not a fair assessment. And then I so I had my comeback after that. I said, "Well, yeah, but I can go further and faster and have more fun. "

Yeah.

Now I see people have come across to e-bikes and you know, everybody's using them and I don't get that that point of view anymore. Unless there's some real laggards out there that will never convert, but the majority of people have shifted across and they love e-bikes as much as I do. And I think we're going to have the same with AI.

And I like the way you frame it that as long as people are involved, um you know, AI's going to take us a long way forward. But a lot of those those fears or those objections are going to be resolved.

Yeah, that's what always happens with innovations, I think.

Exactly. Exactly. And we're all part of the digital transformation, aren't we? So, you know, I think there's plenty of consultants out there that will a kind of do what you do before the AI revolution and say, "Well, I'm going to help transform your organization. " And I always used the example of a guy who's trying to sell hot dogs, right? You know what a hot dog is. You know, you put the sausage in a bun. I mean, yeah.

Back in the day, he just would have his little trolley and he would make his hot dogs and he would go down to the street and he would take cash and he would give them a hot dog. Now, that doesn't even work. He has to go down there with his trolley and his hot dog and he has to have an EFTPOS machine to take money cuz people don't walk around with cash anymore. So, he's already transformed in the most simplest way to the digital world whether he likes it or not. And uh it I'm finding it increasingly difficult to find any business that is not part of the digital world that hasn't transformed. And then we're going to say I believe to see the same thing with AI.

We're going to look back and say what business is not using AI. No. We're not going to find any.

Street artists have hit already.

Yeah. So, what was that? The artists? Street artists?

that are asking for money, they saw that with the digitalization people didn't have coins with them. So, they're now everywhere you can pay by cards also.

There you go. Perfect example, yeah.

Everybody is adapting.

Yes. Well, if you don't adapt, in that case, so in the hot dog case, the if you don't have cash, you can't buy. So, if you're a business person, you have to say, "Okay, I have to meet the customer on their terms. They don't have cash. They've got a card.

I'm going to have to use a card I have to accept the card somehow. " Yeah. So, Ellen, do you have any questions for me?

Um Yeah. Maybe with Skillion, your company,

Mhm.

what has been the biggest challenge when implementing AI at one of your clients?

Uh I think founder issues are probably the biggest ones. In my world, some of the things that they want us to do are quite difficult and require quite a lot of engineering uh effort to get them to work really well and reliably. Um and they don't understand the effort, the cost associated with build building those things.

I think they think that because somebody can stand up a prototype on a weekend using one of these these tools like Claude for example, which which which we do as well. They think that that's the then the solution, but it's not the solution. There's there's a big gap between a a hacked prototype and a working reliable system.

Yeah.

Which requires monitoring, reporting, uh support structures, um all sorts of things that need to go in and that's a very hard thing for a founder to understand. They go in with the idea that hey, I've already got it working. No, not in the way you need it working. Yeah.

Sounds like

And I have that constantly as a problem. Yeah.

Yeah. Yeah.

Yeah, because also the AI agents, they say that uh Claude or uh Gemini, they answer the same questions than your chatbot. If I upload a document and you're like,

Mhm.

but we you cannot do it like this, I mean.

Yeah, yeah. Yeah. And uh well, that brings us to pretty much to the end of the time we have here. I have one final question for you, Alan. AI, is it hype or does it help?

Yeah, I think it's uh more help than hype certainly, but only when applied uh pragmatically. Yeah.

Mhm.

I see the AI uh to simplify operations and to improve efficiency and give businesses also more control on their operations and uh on their growth. Yeah.

Yeah.

Yeah, I 100% agree with you,

Yeah. Ellen. Well, this has been a real pleasure. Thank you for joining us. I will see you in our networking group. Um and I will continue our discussions as we do.

Yeah. Okay. Pleasure is all mine.

Thank you, Ellen. Bye-bye. Well, I hope you found that as interesting as I did. Ellen had some really good examples of how AI isn't just hype, but it's actually helping companies to be more efficient. She used a great example of where a chatbot is actually acting as the interface between a client and the company.

That would help streamline the onboarding process because there's a lot of back and forward between those particular clients to get the um to get the business in through the door. And the chatbot was filling in all the blanks, which occur as we move remotely and across time zones between different people. So, really filling in the gap. Now, the AI is used in two ways. Uh as a translator, which is the front end of the chatbot, but also to access the system such as the CRM in the back background. That is, wasn't just about streamlining, was actually helping them to capture leads, which otherwise they might have missed. So, a really good example of how it can be used.

Of course, we talked about LLMs, large language models like ChatGPT and Claude and Gemini and how they're helping us all be more efficient. And so, that's obviously an important part of how businesses are adapting to it. And most of us are using those tools already. Now, there's some uh limitations and some I guess some areas of concern. One of them is how much LLMs can mislead us uh into believing something's correct and it's not. Um we both agreed that experience really comes in. The human needs to be in the loop to do the checking to make sure that it's valid. Secondly, was the averaging effect.

There's not a lot of original stuff coming from uh these LLMs because they are really just taking what is statistically the most common next word in in an answer, um which is extremely helpful, um but it does limit the ability of an AI to actually do something original. So, I hope you enjoyed that. If you want to know more about Censai, Ellen's business, follow the links in the description. If you want to know more about Skillion, please do the same. I hope that helps, and thanks for watching.