Why 20,000 AI Licences Delivered Nothing, with Alvaro Melendez

How this transcript was made

Transcribed by machine from the published audio, then corrected by hand. It is not a verbatim record: the changes below were made, and are counted.

This episode

75 speaker turns, 191 changes.

• opening voiceover replaced with the authored script (match 94%) and attributed to Max; 8 words of episode audio that shared the turn were kept
• closing CTA identified as CTA 2 - Subscribe (v4) (match 98%), replaced with the authored script and attributed to Max; the 10 words spoken before it stayed with Alvaro Melendez rather than being swallowed by the substitution

What was changed in these transcripts, and what was not

Generated: 2026-09-20 from 2026-09-20-named/, itself transcribed from the audio that is live on the feed. Every change below is mechanical and counted. No sentence was reworded, no claim altered, nothing removed for being awkward.

The editorial decision

Owner's choice, 2026-09-20: light clean. "Um", "uh" and stammered word repeats ("and, and") are removed, because nobody intends to say them. "You know", "I mean", "kind of" and "sort of" are KEPT - they run to roughly 2,300 instances across the catalogue and stripping them would make the page read as somebody else.

Counts across the whole set

Change	Count
stammered repeat removed	667
um / uh removed	654
space before punctuation	241
split contraction	132
guest turn named (Joe Casabona)	101
organization -> organisation	71
diarisation fragment returned to the host	62
ai -> AI	57
empty turn dropped (was filler only)	36
synthetic outro restored from script	21
organizations -> organisations	18
hyphen followed by space	14
realize -> realise	14
A.I. -> AI	10
recognized -> recognised	9
gray -> grey	9
summarize -> summarise	9
color -> colour	9
behavior -> behaviour	8
synthetic intro restored from script	8
analyze -> analyse	7
behavioral -> behavioural	7
recognize -> recognise	6
percent -> %	5
personalized -> personalised	5
skeptical -> sceptical	5
analyzing -> analysing	5
center -> centre	5
optimization -> optimisation	4
realized -> realised	4
favorite -> favourite	4
realizing -> realising	4
canceled -> cancelled	4
recognizing -> recognising	4
announcer turn relabelled from Guest	4
summarizing -> summarising	4
labor -> labour	4
behaviors -> behaviours	4
colors -> colours	4
organizational -> organisational	4
customized -> customised	3
analog -> analogue	3
optimize -> optimise	3
honor -> honour	2
optimizing -> optimising	2
summarizes -> summarises	2
personalize -> personalise	1
prioritize -> prioritise	1
prioritized -> prioritised	1
organize -> organise	1
rumor -> rumour	1
organized -> organised	1
capitalize -> capitalise	1
centered -> centred	1
organizes -> organises	1
emphasize -> emphasise	1
split name repaired (Schaefer)	1
humor -> humour	1
traveling -> travelling	1
defense -> defence	1
maximize -> maximise	1
Organizations -> Organisations	1
apologize -> apologise	1
digitized -> digitised	1
monetize -> monetise	1
analyzed -> analysed	1
narrator opener re-attributed to Max, words left as transcribed	1
theater -> theatre	1
specialized -> specialised	1
split name repaired (Wadsworth)	1
aluminum -> aluminium	1
specializes -> specialises	1
skeptic -> sceptic	1
customize -> customise	1
personalization -> personalisation	1
optimized -> optimised	1
visualize -> visualise	1
spoken domain -> humanspark.ai	0

What was deliberately left alone

• "Program". A computer program keeps that spelling in UK English; only a broadcast or a scheme is a "programme", and only a reader can tell which was meant.
• "License" and "practice". Both have a noun/verb spelling split in UK English that a rule cannot decide.
• Capitalised words mid-sentence that would otherwise have been respelled: 4 of them, most often part of a name. Left as spoken.
• Retired show and brand names. Spoken by the host and guests. A transcript that removes what was said is not a transcript.
• "Co-pilot". Sometimes Microsoft's product, sometimes the metaphor. A rule would corrupt the honest ones.

The synthetic segments

The opening voiceover on the archive episodes and the closing call to action are read by Max, the show's AI narrator, from a script written in advance. Transcribing them means guessing at words already on file, and the guesses were wrong in places - one opener came back as "an AI that boned him up" where the script says "phoned him up". Those passages carry the authored text instead, and are attributed to Max rather than to the guest: the diariser groups Max with the non-host speaker, which would otherwise put a real person's name on words they never said.

MaxYou're listening to Something About AI. Alvaro Melendez has the sharpest line in the archive for anyone rolling out licences at scale: you did not buy twenty thousand licences,

you hired twenty thousand very smart employees, so what are they going to do? He is just as blunt about the alternative, which he calls hiring Albert Einstein to serve coffee.

Two notes for the end. Alastair asks for beta readers, and that closed long ago. And he names his book by an early working title; it was published as Use AI, Stay Human.

Here's the conversation.

Alastair McDermottAI is everywhere in the news. A

Lot of businesses still doubt that it can deliver real returns.

There's a lot of flashy demos, there's a lot of big promises, but a lot of people are still struggling to find practical ways to use AI and see a return on investment.

And there's a lot of skepticism about AI's value, particularly when it comes to justifying its cost. And we'll give an example of that in our conversation today.

So what I want to do today with my guest is to cut through

Alvaro MelendezThe hype

Alastair McDermottAnd explore how businesses can actually use AI for real growth.

Talking about real examples and discussing concrete steps that you can actually take to get measurable results from AI and not just have a very expensive experiment.

And today my guest is Alvaro Melendez, founder of Krant Creativity and Technology. Alvaro, you're very welcome to

Alvaro MelendezThe show. Thank you very much for having me. I'm excited about being here.

Alastair McDermottWell, as we talked about in the pre-show, you know, this is something that is getting a lot of a lot of press right now

I think there's a lot of skepticism about the value of AI. And I personally, I think it's crazy.

Because see, looking at it from a technologist point of view, the technology is, it's increasing in power and capabilities so quickly, like it's insane what's coming out every week.

And yet we're seeing, like we just discussed the story from Chevron where they have, you know, 20 or 30,000 licenses and they're paying 30 or $40 per license per user per month

And they're not seeing a return on investment. Like, how do you think about this?

Alvaro MelendezYeah, I think it's ridiculous. I think it's basically, you must have your use case very wrong to don't, to, you know, to not see the return on investment.

It's, when you think about this, like even your regular ChatGPT, license, like the pro license is 20 bucks a month, right? 20 bucks.

if you don't find a way to save 20 bucks to make this worth, you're probably just, you know, not understanding what this is about. And that's probably also not your fault.

it's probably that the companies themselves are not doing a great job and explaining how this works and giving practical use cases. So yeah, they don't have the time for this.

They're developing it.

Alastair McDermottYeah. So, well, maybe that's our job then, but the way I, the way I see it, they must be doing like the companies who are developing these,

these products, these AI systems like Microsoft and Google and open AI. they must be doing a spectacularly bad job of showing these enterprise corporates how to use this because, you know,

if you're, if you're going to charge, you know, I don't know what exactly what the run rate is, but it's in the millions of dollars, for a corporation that size.

it is a significant investment and I can see why people are sceptical on the other hand to get a return on investment for a single user. it's, it seems to me to be so easy.

Like all you got to do is spend a couple of hours using it and you see, wow, there's so many different ways I can use this.

So we're looking at this, like making this, fit, like it, is it a serious lack of education or, you know, are they not showing people how to use these tools or do we just not have enough,

you know, do we not have enough, knowledge of real world use cases for this yet? Like where's the problem?

Alvaro MelendezYeah, I think, I think there's multiple problems. One of course is the lack of education for sure. Right. Like you get that thrown on your desk, like, Hey, here's your license. Use it.

And I mean, you're probably pretty busy with the job you have to do.

So you would have to find your own way through there to understand the technology, then understand the use case, then test it, then know if it, if you're allowed to do it or not.

So there's a lot of unknowns there that need some guidance. I think that's one problem, but I think also the other problem is just, how people think about AI in this moment.

And so, you know, many people say that all the time. I think Ethan Malik is one of them that says this all the time.

And I say that all the time is like, if you're thinking about AI as a tool, you're just massively under, cutting this thing because, AI is not a tool.

Like I like to think of AI, you know, I come from the creative world. So I, in Spanish, we have this world for a dupla. So it's like your thinking partner.

You know, when you work together, you're two brains. And most of the time in creative, you have two brains. One is more copy writing. One is more graphic and then both compliment each other.

And then they come up with the great ideas. So I think of AI much more as a dupla. So it's a thinking brain that can do thousands of things.

But if you ask it to write emails, well, then it's, it's going to write emails and that's it. Right. Someone else said, I don't remember who this was, but it's like having an internship.

I'm sorry, having an intern and say, you know, you have Albert Einstein as your intern, but you're asking them to serve coffee.

Well, Albert Einstein certainly can serve coffee, probably not the best coffee, but he can do that.

But he could also change, you know, the way you think about technology, but you're not asking it to do it. So it won't.

So again, I think if you, if you going back to the Chevron example, and this morning I was listening to the AI marketing show with Paul Ritzer, I think you're familiar with them as well.

And they were discussing the Microsoft, I don't remember the name of the report that came out, but they were basically proving how well co-pilot works.

And in some cases, and the focus was put on like, you know, people are reading 10 or spending 10% less time on reading emails. You know, what is that?

Like, that's probably the worst use case you can think of, but I'm pushing that. So, I mean, if people hear that and they might think, okay, co-pilot, you know, summarises emails.

That's great. For me, that's like, again, that's asking Albert Einstein to serve coffee. That's yeah. Yeah.

Alastair McDermottSo yeah, absolutely. And, I did hear that episode of that podcast, great podcast, to listen to as

Alvaro MelendezWell.

Alastair McDermottAnd it's, it's insane that we're looking at things like, you know, a 5% productivity saving in email when you could be looking at,

like some of the things that I've done with it are I've taken the text of all of my books and uploaded them. And I've been able to use that to pull out insights.

I've been able to use that to create new content, create marketing content. I've been able to take the transcripts of many podcasts.

I have 175 episodes of my other podcast until I've got about 250 podcast transcripts that I can use to basically data mine for insights and advice about different topics.

And if I wanted to, I could turn that into the outline of probably several books. if I was so inclined, there's so many different ways.

And that's just with the very simple things that I'm doing with it.

like I find it hard to believe that, you know, that there aren't people in companies like Chevron who can find ways to turn this into,

you know, millions of dollars worth of productivity and creativity.

Alvaro MelendezAbsolutely. Which I think it's shocking.

It's also a great opportunity, you know, because this opens like, I mean, if they are not doing it, the companies that are doing it, I mean, they're growing so fast.

They're changing so quickly the way they think about business that it's going to be hard to catch them once, you know,

the other folks understand what this was capable of, to give you a couple of examples in the, so grant, we help agencies and businesses adopt AI. Okay.

So we do a lot of training, but then we also apply our own ink lantern technology to help them.

If that's a use case, they would benefit from now, things that we see there is, for example, if you look at agencies, right. And creativity, right.

There's very profound human value of, you know, imagining stuff and turning that into a story and that story, becoming a brand and that brand, driving millions of growth.

And, you know, that's what humans excel at. And that's what makes us different from all other species on the planet. But now AI is here and suddenly AI can do that as well.

So it's a very, if you think about it, the more you think about it, the more, interesting and also a little bit scary it becomes,

but, going back to the use case in advertising, what we're doing is, you know, you take the job of a creative director,

for example, and then you break the job down and like all the tasks that they do. And then some of them you might be able to do with AI better. And some of them you don't.

So it's not replacing the job, but it's certainly replacing some of the tasks that a job has to do.

So the intention of that is, okay, let the human do what the human does best and let the machine do what the machine does best, but then put them together.

Now, the interesting thing about this is, you mentioned the Chevron case, which, which became kind of famous recently, right?

Say you buy 20,000 or whatever it is licenses, and you think of, okay, now, you know, the people have one more tool that they can use. That's one mindset.

What if you reframe the same thing and say, hey, I just hired 20,000 additional employees, right? What are they going to be doing? So did you hire 20,000 additional employees to do email?

Hmm. I don't think nobody would do that. Right?

So if you think, hey, I have suddenly this thing, the way I approach AI adoption is AI can, I mean, AI can do many things, but you can use it to do the stuff that you're doing quicker.

More efficient, more efficient, maybe be more effective. That for sure. That will reduce your cost. Yes.

but you could also try to use it to do whatever you're doing just better or do stuff that you were not able to do before. Right? Like examples.

Again, you say, how did, is it okay, Alastair, if I keep talking about agencies or you want me to? Yeah, sure. Sure. Sure. Yeah.

so agencies, for example, they would have a way to present to a client a certain idea, right?

And that idea would be presented in text form, maybe with a couple of pictures with maybe, you know, stock photography, something like that. Right?

Now you can do the same, but you don't have to use text anymore. You can have actors like AI actors reading it.

You can have, images generated in my journey that are much more, accurate to your idea, to capture your idea.

And then you can also turn that into video with runway or, you know, the other tools that are out there. So your output is much better now.

And the client is going to see, you know, a much more sophisticated, realistic execution.

And that might win you a lot of deals that you would have lost because maybe the person looking at your stuff didn't have the imagination that you have to understand what this can become.

Right? That's why, when this, company is doing movies and doing series, they don't just tell you the story, but they do a pilot, right? Because you need to see it and then you'll love it.

So same goes for advertising, but just as an example, right?

Alastair McDermottYeah. So if we have all these businesses who have effectively, like everybody has their own personal Einstein or, you know,

and by the way, maybe that's not too much of a, of a, of a leap because right now the, the current frontier models are kind of in around PhD level in terms of intelligence.

they still make some very basic mistakes on some things that you wouldn't expect them to make mistakes on, but on, for the most part,

they're like, they're able to pass, you know, they're able to pass bar exams in every state. They're able to pass, you know, medical, exam inations that they're able to do.

so it's like having, you know, an intern with a PhD in your pocket.

but again, you have to give them directions and if you ask, if you ask them to make coffee, then yeah, like that's out, like asking Einstein to make coffee.

But the issue here is, people, employees who have been, you know, told, Hey, you know, you've now been enabled for Gemini or co-pilot, you know, go use it.

they have no idea of what those capabilities are and they also don't know some of the pitfalls and some of the, some of the things that you kind of, you learn along the way.

As you start using these systems, you learn that, Hey, every so often you need to start a new chat probably more frequently than you might realise because otherwise you can start to see

problems with the context window. And, it's starting to forget things and make stuff up.

and also then there's this problem of, Ethan Malik, I think referenced this recently, type one versus type two errors where type one in,

and this comes from the world's statistics, but type one errors are when we, see something is missing. And it's like, Oh, it's missing something that should be there.

And we're really used to seeing, Oh, there's something missing from this. What we're not used to as humans is seeing something there that shouldn't be there.

something that looks really plausible and it looks like it, yes, it should be in that list, but it's actually totally made up.

And this is what got that lawyer in, in New York who, who, filed some legal documents on a case and, was thrown out of court because, in a very embarrassing way for him and his law firm.

because it made up all of these case references that didn't exist, but it made them up in a very plausible way because these large language models are really good at generating language that

looks plausible. so there's lots of nuance to using these systems.

And so it's like having these really smart interns available to us, but you do have to feel like you have to tell people how to use them and tell them and warn them about those kinds of

errors.

Alvaro MelendezA hundred percent. That's why I like thinking of it more or less like, you know, you say you bring 20,000 people into your organisation.

Are you not going to tell nobody like what these people are supposed to do? Like, are they just going to sit there? Like, you know, you wouldn't do that. Like if you think of that.

So I think that's why it's so important to not think of AI as a tool and think of it as like, yeah, this is a new type of intelligence and it is a very good intelligence.

Yes, it can, you know, it is not great at everything. That's why the use cases are so important. So, for our listeners, right? I would say, think of it, of it like this.

You can hire as many people as you want for 20 bucks a month or something. And they're really smart. So now you have to think, what do they, what do I want them to do in my organisation?

How will this new team, let's call the AI team, be infused into all of my processes and how will it interact with my human staff and how am I going to make the most for my people?

Because they're going to be happy that they can spend more time doing what they do best and what they love.

And then AI can take some of the burden of the stuff that they're maybe not as much enjoying, right? And then how do they work together? That doesn't happen automatically.

It needs direction. It needs leadership commitment. So this needs to come from, you know, from the owners, from the directors, from whoever is making the decisions.

They need to understand it. And the problem that we keep seeing is those folks are the ones that have less time to try stuff out. And therefore, they know the least.

So they're making decisions about something without knowing what they're talking about. So that is very dangerous. But let's say we go for what should you do? Well, get those folks together.

Get someone or do it yourself if you're in the capability. Train them on AI. Explain them what is a frontier model. What is an LLM? What does it do? Let them see it.

Hopefully, let them experience it. Because learning through doing is so much better.

So once you get someone to, hey, I do this all the time and I worked on this with David Levitt and Jonathan Briggs. It was amazing.

We asked the team, like, hey, give it this database about the Titanic, you know, about the Titanic people that, how do you call them?

Like, the people that were on the Titanic, a real data set that is very famous and has been used for many things. You give them that and say, now, give that data set to ChatGPT.

Now, ask it questions about the data. And we're talking senior leadership, right?

And suddenly, they start, oh, my God, like, ChatGPT is giving me, like, correlations and it's giving me predictions about if I would have survived or not,

depending on my age or my gender, things like that. It opens their brain. And then you tell them, you know what?

Now, why don't we create, like, a game where you give it, like, your age, your gender, and then it tells you if you survive or don't.

And then it can make a joke about it or it can, you know, read the future like an astrologist.

So, you start playing with that and they do it and they see, oh, my God, this thing can program a game. And I have never done that and now I can.

And now let's turn this into a song and now let's turn this into a movie.

And, you know, once you do that in 30, 40 minutes, they will never go back to thinking AI is only good for email once they see that, once they try that.

And then the use cases unlock because once you understand it, but that's number one, you need to understand. And you need your folks, like your leadership team to understand it.

Then you put them together and then you say, now let's talk how we're going to adopt this thing. So, let's break down our jobs. What is important?

What do we think this can be, AI can do this? How would it look like? Who gives the input? Who gives the output? Who revises it?

You know, as you said, AI, if you want to do like numbers and you need to rely on that, maybe AI is not great for that.

Or you're going to have to have a very smart, detail-oriented person supervising that to avoid those type of mistakes. So, maybe that's not the best use case.

But maybe it's great at helping you know, think of business strategies in innovating in new products. Maybe it's great at helping you with research, with creativity.

Maybe it can help you with customer service. Anyway, you will figure that out for your own business once you understand what it can do.

And then once you have that, you have to say, not everyone is going to adopt this. Not everyone is going to want to invest the time that is required to get a decent usability out of this.

So, you need to select those people. You need to handpick them and then form something like what Paul Retzer calls the AI Council, which I like as well.

So, you create a group of AI experts that are internal because this is super important. As a business, you can't rely on third-party vendors to be your AI guide.

That needs to be a core to your business. It needs to be internal. And then they can work with vendors, of course.

But these people need to be inside the organisation, coming from different roles, understanding the business as a whole,

and then finding ways to grow with the use cases because this goes so fast that something that wasn't possible yesterday maybe is possible tomorrow, probably.

So, you need to have that AI Council being connected to innovation, to the world.

And then bringing your use cases forward in a way that is applicable, in a way that is doable in the short term. And this can be very big or very small, depending on your company.

But I think anyone, if your company is one employee, you will, I mean, the ROI on AI for you is going to be great.

And I have not seen, like, I have not been in one AI workshop with a company where they immediately figure out, like, very high-value use cases that are 100% doable, right?

So, once you do that, and then I will stop, but once you have your people understanding what AI is, identifying the use cases, putting them on a roadmap, very important, ethics and legal.

You want to bring that in immediately because there's stuff that maybe can be done, but maybe shouldn't be done or shouldn't be done in that way.

So, you need to know that to avoid any future problems that you might fall into just because you don't know. So, that is really important.

Alastair McDermottYeah, there's a lot to dig into there. A couple of things that you said, I mean, it seems to me this is a culture issue, and it needs, like most cultural issues, it needs to be embedded.

Alvaro MelendezAnd I'm talking about corporate culture.

Alastair McDermottIt needs to come from the top in terms of if leadership are not embracing AI, then it's going to be very easy for the organisation as a whole to be sceptical and not to make the most of it.

One of the other things that just struck me is, I think that for a lot of people who were encountering the good types of AI that we're using,

and when I say that, a lot of people experimented with the free versions of ChatGPT and experimented with the free version of ChatGPT for many months was version 3.5.

And it was actually very limited compared to what the current state-of-the-art models can do.

And so, if you experimented with that, I can understand why you would think that it's not a very powerful system and it's not very smart.

But it is, you know, in technology terms, that's like looking at cars that were built, you know, in the 1900s and comparing them to cars that are on the road today because it's moving that

fast. Like, it's insane, the level of development here.

And then the other thing is, if people's first encountering, the place that people first encounter these is in some of the tools that they use all the time. Like, for example, in email.

So, I have it set up.

Alvaro MelendezI use Google Workspace.

Alastair McDermottAnd so, I have Gemini built into my tools and documents.

But if that was the first place where I encountered AI, I really wouldn't think that it was very powerful because, like, what it's saying to me is, hey, summarise this email

Alvaro MelendezThread.

Alastair McDermottLet me help you do a search better. Let me help you write a reply. And those are all fine things to do. And, you know, like, they do help, you know, save a minute or two here and there.

And the search feature is certainly a lot better than the original built-in search feature.

But, again, those are super limited compared to what the current capabilities of the current state-of-the-art models

And, again, those state-of-the-art models, the ones that are commercially available that people log into and use daily,

there are even more state-of-the-art models that are available in, for example, Google's AI Studio, where they make the latest versions of the models available that people aren't even seeing

on a, you know, on a day-to-day basis. So, I think the issue here is looking how we perceive, like, as business leaders, how we perceive what the capabilities are.

Your perception of what the capabilities are may be way behind where they actually are. And the same for users. And that's going to drive skepticism.

And that's the problem that I see is because people aren't, you know, as deep in it as you and I are.

You know, they're not, you know, they're not playing around with these systems all the time and saying, oh, my God, it can do this, it can do this, it can do this.

And so I think that because they're encountering it in that very limited way at the start, that's driving some of this skepticism.

And I think that's, you know, that's unfortunate because that may be limiting how people are approaching and using

Alvaro MelendezIt.

Alastair McDermottAnd that's why, you know, your example of playing around with things like the Titanic database, that's a really interesting way to play around with.

I think when we play around with real data, and like I've done it myself with, like I mentioned previously, the podcast transcripts, which is a mountain of data that I

Alvaro MelendezCan use.

Alastair McDermottThere's lots and lots of things that you can do when you feed data to the machine that you won't see if you just start with a simple prompt.

Alvaro MelendezYeah, 100%. And I think, again, like, unfortunately, like, we can't look at Microsoft OpenAI, Google, Anthropic.

We can't look at them to train us and to guide us on the use cases because they're busy building it. And I think, I mean, if you, oh, we need to talk about this, man.

You saw the Sidney McLaughlin Gemini video that Google put out for the Olympics.

You saw that, like, where it's a girl that is a fan of Sid ney, and then the father wants to help the girl to write a letter to Sidney,

like a fan letter, because she's, you know, like the biggest star. And then he uses Gemini to write the letter.

And then you're like, yeah, that's exactly the use case that you don't want to do. That's exactly what you don't want to do. And that's Google saying it, right? Yeah.

So I think there's, like, a huge disconnect for some reason, I don't know why, between the Google or all of these, like, tech innovators and then the brand and the storytelling folks.

Like, I think they're doing a terrible job in showing what this can really do. I don't know why, and, you know, I haven't figured that out.

But, again, I think maybe a good example is, like, you're a business leader. Say you can hire Einstein for $20. Like, what would you have them do?

And then if you answer that question, probably you can use AI for that. And then you can ask, okay, and maybe if I can hire 10 of them, what would I have them do?

And then let's say I can hire, you know, mathematicians or maybe I can, you know, a new creative. Or maybe I can hire a big strategist. All of them are present in the current AI.

And as I said, it moves so fast that I think even for people like you and me that are in this all the time, it's impossible to keep up with everything that happens all the time.

Because this moves too fast. So I think, again, like, the leadership I would concentrate on, get the fundamentals right. Get the fundamentals right. Understand what this is capable of.

Maybe not today, but what this is capable as a concept. And then you can think about the applications, whatever is capable, the AI is capable today. Yeah.

Alastair McDermottAnd, you know, you walked us through a very good framework or approach to that earlier on.

And it is as simple as give people access to the tools, but also show them how to use the tools and get them to play around with it in a way where it's,

you know, it's actually showcasing the capabilities in a real way. Not just, you know, here's, like, summarise this email

Alvaro MelendezThread.

Alastair McDermottBecause I think you're, you know, that's getting Einstein to make you coffee.

Alvaro MelendezAnd that's a great analogy.

Alastair McDermottThe way I see it, and one of the things people are fearful of is AI coming in and taking jobs. And that is a real, like, that is a real concern, particularly for certain types of jobs.

But I was talking to somebody recently, and they gave me the analogy of somebody in the Arctic with a sled team of huskies or the dogs that are pulling.

And the AI is the dog sled, and the human is the person

Alvaro MelendezDriving.

Alastair McDermottAnd that's, like, we can use it to do so much more if we drive it right, if we ask it the right things. And I think that right now we're only scratching the surface of that.

Alvaro MelendezWe haven't figured it out.

Alastair McDermottAnd that's where the use cases that you're talking about

Alvaro MelendezCome in.

Alastair McDermottBut they're all, typically, they're all going to be business-specific. Now, there are some kind of general use cases that will work for everybody.

But typically, they're going to be business-specific. But you can only figure those out once you get people on your team starting to use these tools.

Like, you have to spend hours using them. But, you know, that commitment of time, it's not hours using them in isolation. It's hours using them working on what you're working on.

So you're still making progress with your work. It's just you're probably doing the task that you're doing. You're probably doing it a bit better and a bit quicker.

But I think that for us, we're scratching the surface of a technology or a group of technologies that can do so much for us if we play around with them.

And if we say, okay, maybe the skepticism isn't correct. You know, maybe we're just like in the Gartner hype cycle.

Maybe we're just in the period where we're on the downslope and the disillusionment.

But when you look at things like, you know, OpenAI's advanced voice technology, like, it's insanely good what it

Alvaro MelendezCan do.

Alastair McDermottAnd that's now getting rolled out to people. And I know people were concerned, you know, is this just a faked up demo? But no, people are using it now.

And the capabilities of that are insanely good. So, and that's just like just one piece of like AI is a group of all of these technologies that we put together.

And that's just one of those pieces of technology. But the capabilities that will be on like available to the general public in a year, 18 months, like who knows what that will have.

Alvaro MelendezYeah. You know, another tip that is maybe useful for our listeners is you are the leader in your company and your company is whatever size, it's 10 people or 100 or 1,000 or 10,000.

That doesn't matter. Now think this through for a moment. Think about the problems you solve for your customers. So that's why you exist.

And now think there's someone out there anywhere and that person is an AI expert. And they have the idea to solve the same problem that you're solving. So, but they're only one.

So how are they going to do it? So they're going to use technology to do whatever they can with one person. Now think what if they were like two and then three and then 10.

How quickly could they replace you as a company with much more power because they're using AI? And that's happening, right? That's happening everywhere.

Now, my tip here is don't wait for that to happen. Do that internally. So find one, two, three, four, five, whatever size of company you are. Find a few people that like this.

They are there. You just have to ask. They like it. They want to learn more about it and then give them the task.

Like, hey, your task as my AI counsel is to basically impl ode the company from inside.

Like, create everything that we do AI driven so that our competitor, AI driven, comes from inside and becomes then outside. So there's very little risk.

I would say there's zero risk to it, right?

Just take these people and give them some time and give them the permission, but also the task to do this in a responsible, ethical way, of course.

And then let them experiment and give them time to come up with the use cases, support them and test it. There's nothing you can lose while doing this.

I 100% think that this is going to be, it already is, but it's going to be such a transformation that we can't even imagine with everything that's coming.

And the companies that don't get in early really are going to struggle much. And we don't want to see that happen.

Alastair McDermottSo let's just wrap back around. So what we're saying is there's a lot of skepticism about the return on investment of AI.

And in particular, a lot of that is coming because people are being given AI tools but aren't being educated how to use them properly, aren't being shown how to experiment.

There aren't kind of AI, like cross-functional AI teams, like AI councils, like you talked about.

And that if you actually spend time on putting those things in place and educating leadership and showing leadership how to use these tools, that can set a culture in place.

At least that was my wording.

I don't know if that was the, don't want to put my words in your mouth, but that was kind of what we were saying is you're creating this culture where people understand the capabilities of

the current systems at the very least and start to get into that mindset of playing around with and having permission to play around with different AI tools.

And I know even that, like I saw a post on a discussion forum where a woman was accosted by her manager. Her manager said, hey, are you using AI in a kind of a very

Alvaro MelendezNegative way?

Alastair McDermottAnd she said, no, I'm not using it, I swear. You know, and it's like, if that's the approach that people are taking in your organisation, you've got problems.

You know, if like, it's got, it's got to be, you know, it's got to be better than that.

Encouraging people to use it and play around with them and figuring out how to use it and setting guidelines for

Alvaro MelendezPeople.

Alastair McDermottLike we talked about the training people, but you mentioned ethics and the legal side of things as well. And that's stuff I've covered on the podcast before.

But it's really important to understand, you know, like there could potentially, like there are so many different AI systems out there.

There could be biases and there are biases built into them.

You know, if you're making, you know, financial decisions, like whether to give somebody a mortgage, like there could be biases built into the data set.

Like, for example, if they live in a particular street address or if they have a particular name that the system could be biasing. And, you know, there's all of that kind of stuff built in.

So we need to be really aware of those and counter those, proactively counter those biases when we're using these systems as well. So there is a lot to learn.

And I think that's, you know, that that's a challenge.

Like this isn't this isn't something where, you know, you do what Chevron did, which was just enable it for 20,000 users and say, oh, like it's going to be fine.

You know, we're just going to pay 20,000 by $30 a month. And we're going to get return on investment immediately without training anybody, without showing them how to use

Alvaro MelendezThese tools.

Alastair McDermottYeah, no, it's like it's so exciting because there's so much that we can do with it. But at the same time, like I hear these stories and it's it's it's discouraging as well.

Alvaro MelendezYeah, but, you know, I think, again, like I see that as an opportunity because like culture changes slowly.

So I think as a company, what I would say is don't get frustrated if you're not able to train your 100 employees. Some of them won't like it and won't adopt it. Right.

You're you're in a much better place than most if you're able to convince 10 to do it, then start with those 10 and buy 10 licenses.

And, you know, this is a real case with with a client of ours. There are 300 people company, more or less. And we started an AI council with them and we started the training and all of that.

So we've trained everyone, but they're not all using it. Now, we have an AI council of 10 people and they are using it all the time and they're coming up with use cases.

They are delivering the use cases to the rest of the team. The rest of the team is not using it yet. Right. They don't feel comfortable with it yet.

Now, what we have seen is so the AI council is all in there. They're becoming experts real quick. Now, a lot of people asked, hey, I would like to join the AI council.

Now, you can't make the AI council huge because then it loses ability to act. So what we did is, hey, let's create like support groups like we call them like that.

So it's like chats, like group chats where anyone can join. And then what we discussed there is applications of AI in their business. Right. Someone comes in and say, hey, you know what?

I just tested the new Midjourney version 6.1 and I did this prompt and look at my result. And they post that in the chat.

And then someone can come in and say, hey, did you guys see that they just yesterday released this new thing and that? And then, you know, people get interested slowly.

But once they start feeling comfortable talking about it and lose the, you know, most people are afraid of it because they feel like it's too technical. It's too difficult.

So they don't want to do it. Once they start losing it, adoption happens very natural. Because as you said, like once you try it, once you get results out of it, there's no going back.

Like there's no way I could work without AI today. You know, there's no way. I do so much more with it that it will be very difficult for me to go back.

And that happens to everyone eventually once they start getting in. I have a very fun story as well. Like my father-in-law was visiting me yesterday, like last week.

And he does valuations, right? So he needs to analyse a lot of data. And he was like, hey, Alberto, I want to learn to code in R. And I'm like, what? You want to learn to code R?

And he was like, yeah, buy me. Please help me buy a book on Amazon for that. And I was like, oh, my God, yeah, I'm going to buy it. But, you know, he's no programmer. He's never done it.

So it was going to be very challenging. So I said, why don't you try ChatGPT? And I installed the free version on his phone.

And I was like, ask it questions about R and ask it to train you on R. And then ask it if it will help you solve the use cases that you have for your valuation company.

And he started doing that. And he was like blown away. He was like, oh, my God, like, I don't know if I need the book, you know, now.

Because this guy is basically coaching me, helping me get it. I think it's a good idea to get the book anyway. But he was so fast. And he was like, the book had like exercises with data.

And he was feeding the data into the tool and then doing the calculations.

And then he realised like, yeah, this is, it's difficult to describe the level of brain power it gives you in a matter of minutes. Yeah.

Alastair McDermottI'm still astounded at what it can do and the capabilities. And like I said, I fed a lot of data into these systems. And, you know, it's hungry for data.

And the more that you can give it, the more that it can do. And I was teaching a workshop recently in, so I'm in the west coast of Ireland.

And we have sections of kind of more rural regions where Irish is the first language. For most of Ireland, English is the first language. But for certain places, Irish is the first language.

And so I was, I was helping the organisation. I was teaching this AI workshop introduction to AI. But I was able to use the voice chat to translate because I don't speak Irish very well.

Alvaro MelendezI'm pretty weak, to be honest. It's not my strongest.

Alastair McDermottSo I used it to translate in real time, translate my introduction into Irish. And the translation, so this was the kind of the basic version of the chat. And it did that chat in real time.

And it had terrible pronunciation.

Alvaro MelendezIt was really bad.

Alastair McDermottBut then I turned on the projector and I said, I know the pronunciation was terrible. But look at the written Irish. And it was absolutely perfect. And it had done that in real time.

Like, and that's like, this isn't a purpose-built translation tool. And Irish is a very small language. It's not, it's not spoken by a lot of people, you know.

So like, it's, it's insane that it's insane, the capabilities, things like that it can do. So I can't even imagine what it's going to be able to do in, you know, five, six years time.

So last thing I want to ask you about is can you just tell me a little bit more about your own business and how you're, how you're using it and what you're doing with AI

Alvaro MelendezFor your clients? Yeah, sure. So we started the business in 2019. So before like the Chad, GPT, Gen AI revolution.

So we were working on a lot of machine learning, a lot of cloud computing, and then helping companies make sense of their data and, you know, those type of things.

Recently, we've specialised in two things. So we help with AI adoption, as I just said, from training to AI council creation, to use case development, to tools, if they need them.

So the whole thing, that's one of our areas of expertise. And the other one is Ink Lantern. And Ink Lantern is a AI driven brand building platform.

So basically what it does is it helps you measure your brand love, compare it against your competitors as a company, and then give you insights from the best brands in the world.

So think of it as a team of very smart strategists, creat ives, that would be analysing what you're doing and what everyone else is doing.

And then they would be telling you like, hey, we could do this, we could do that. This seems to be a very good idea because the data supports that this is a good idea to work with.

So basically measuring brand love, but then also giving you the information you need to grow it.

And what we have discovered since working on this is that technology alone for many users is still difficult to use.

So we provide like the service of, you know, we as humans go in with you, hold your hand, go through the data, explain it, you know, answer questions, have ideas.

And then, so we help them make the most of the data. That's what we do.

Alastair McDermottYou know, and this, it just goes back to, for me, there are so many places where we still need a human in the loop, even with all of these tools available. We still need the person.

And it's going to be like that. There are going to be a whole class of jobs where that's not going to be the case. And unfortunately, many of those jobs are going to go away.

But there are so many jobs where that's not going to happen

I don't see an AI being a diving instructor anytime soon, for example, you know, but that's, that's an extreme

Alvaro MelendezExample.

Alastair McDermottBut, you know, like you just talked about, you know, we still need experts to explain and talk to people.

Alvaro MelendezThere's, there's a, there's a component of human energy that is just not replaceable by AI for now, I guess. But so we're an AI company.

So we work with AI all the time and we work remote and that, you know, when we have an in-person workshop, it's so different from when you have humans connecting and this,

that what happens there is magical to some extent. And then AI can help turn that into something more scalable and bigger, but it doesn't replace that.

So I think, again, like I'm a positive, I decide to actively be a positive thinking person.

And I believe that AI will help us do a lot of stuff that we don't like doing, like expense reports, who likes doing that, right? So let AI handle that or taxes. I don't know.

I don't like that, do that for me, but then I don't want AI to write the letter of my kid. I don't want AI to write the letter to Santa. I don't want AI to bake me my birthday cake.

I don't want AI to, you know, go walk the dog. I don't want AI to read my son the bedtime stories, right? All those things I want to do. AI can do it.

Yeah, of course, but I don't want it to do it because that's what I enjoy. So I hope the same goes for our professional lives.

Alastair McDermottYeah, I'm with you 100%. Alvaro, where can people find you if they want to learn more about you and InkLantern?

Alvaro MelendezYeah, so LinkedIn is probably the best place to contact me directly. That's the platform I'm most active on.

And then InkLantern.io, like InkLantern.io, there you can see information about the product. And you can actually try it out there. We have a demo that is available.

But also feel free to just reach out directly to me over LinkedIn and, you know, we can have a chat.

And I like having these conversations, doing the connections with people, even if they don't become customers, that doesn't matter.

There's always so much I learned from just, you know, hearing out what people have to say.

Alastair McDermottYeah, I agree with you. And I'll have a call with pretty much anybody on any topic because I think it's always useful to talk to people.

One thing I want to mention before we go, I have a new book coming out called the AI Powered

Alvaro MelendezThought Leader.

Alastair McDermottAnd right now I'm looking for beta readers. So if you're interested in getting an early copy of that and being a beta reader, give me

Alvaro MelendezSome feedback.

Alastair McDermottI'd really appreciate you reaching out. And you can find links to that in the show notes or you can just reach out to me on LinkedIn. So that's the AI Powered Thought Leader.

And that's based on the season one of this podcast, which was called the AI Powered Thought

Alvaro MelendezLeader.

Alastair McDermottSo, yeah. So that's it for today. Alvaro Melendez, thank you so much for coming on and sharing your knowledge with us.

Alvaro MelendezThank you, Alisa. Have a great day and enjoy AI.

MaxThanks for listening to Something About AI.

If you are new here, there is a back catalogue worth your time and a lot more on the way, so hit subscribe wherever you are listening and you will not miss it.

And if you want to go deeper, Alastair keeps a growing library of free resources over at humanspark.ai.