Introduction: Why AI Adoption Needs a Method

Now it's going to be the theoretical talk. I don't usually like a lot of the theoretical, so we will do also a little bit of practice. But my purpose is to share with you what has been our learning on how companies, more specifically SMBs, adopt AI. Not only how they do it, more precisely how they should do it.

and we have worked with tons of companies and when we work in one specific way the impact the time that they save and the results are much better as well as the engage of the people that sometimes is something the people the companies don't consider and is that right now the AI is going to be

used by humans and if you don't do it properly maybe the human doesn't work work well with the AI.

A Five-Level Framework for AI Adoption

So we have this first picture that let you know you have to be aware of five levels. First of all is the data governance, later is the culture and people, as I was saying before, the responsibility, ethics, privacy, all these, available resources and technology and infrastructure.

Why Complexity Varies by Company Type

Some companies have this part easier and others are really really complex. It's not the same being a consultant of a small company that you have your Google Drive or SharePoint, or being an automobile supplier that you have to be connected to all your IoT.

So this is more focused on this first scenario that you don't have a large component of infrastructure. Okay,

The Recommended AI Adoption Steps

Okay and this is about the steps it should be followed to make this adoption.

Step 1: Assess Where the Team Really Is

First of all we can see that this know where the team is. So I don't know if many of you are working a company, if you are the techie of the company or you are not as techie and that's why you are here. But what usually happens is that the company or the executives suppose that they are all using AI or that they are not using AI at all and it's not any of those.

Just for also knowing who I am speaking so I can adapt a bit more, do you think you are the tech guy or tech woman of your company or your enterprise or you are not? Are you the one that are? are. Okay, and the ones that are considered are not as tech. Okay.

Step 2: Train People (Including Privacy and Security)

So, later we have to go to train humans. That is basically our people, the ones that are not technical, and the ones that are technical, doesn't know specifically how to use AI, how to use it properly, and how to use it with good privacy and security.

And I want to share something that we have seen a lot and it's that that tech person of the company say no what I don't need to go through that training process I already know the how to use AI and maybe there are sessions or there are spaces or there are parts

that he or she get bored but I can let you know that some people I think they are really taking they are taking advantage of AI they were exposing almost all the database of their company, as Aitor was saying before, if you use

like a month ago, MCP servers that were public for everyone. We have found that cases.

Step 3: Create an AI Team

Okay, later you have to go to a form to create an AI team.

This is something that from SMBs to huge companies, all of them have an AI team. That is,

okay, we have a bit of our schedule focused on thinking what we should do, what use cases, and how is the people responding to use AI.

Later, one of my

Step 4: Identify Use Cases and Start with Quick Wins

favorites that is identifying use cases. 1Many companies fail because they go to a really high time -consuming use case and low impact because it looked good. It implemented with AI, but maybe it's not what you should be working on.

The big majority of AI use cases we work with companies in the first two or three weeks is the easiest things you can think about. Emails, meeting summaries, maybe reporting, but it's not something as fancy as we can have here with MCPs and so on.

We have to work in a quick win and a prototype That is not going to change our life completely, it's going to be a small impact, but we are going to be improving it bit by bit.

Now we are going to enter a little bit more on all of this, but I think it's going to be more rich for you, not just speaking about it, but going to do it. so we will also do a demo on an assessment about that because I think

the titles even we can go deeper if we want to train you on how to do an assessment on this methodology it's not necessary for all of you for the

Going Deeper on the Five Pillars

Pillar: Available Resources (What You Already Have)

majority of you so all of them have four levels the one that is usually most listener most understood less understood is the available resources and is okay

Okay, all the companies have a lot of information, but they are not working with this information. For example, I don't know if you have worked in any company that has a playbook for training or onboarding a new person in the team. Many companies have that, and they don't give it to the AI.

Pillar: Technology and Infrastructure

So, what do we have that we can already use to use AI properly? Later, technology and infrastructure. structure.

What I have to tell you, this is not the same to a large company with a lot of complex IT systems than SMB that only use, let's say, Serpo in Google Drive and maybe one or two tools more. So for a small SMB it's going to be easier to go to the fourth level because they don't have that complex level.

Pillar: Culture and People (Change Management)

Culture and people, that's one that is usually ignored because they think, okay, yeah, but but they will use AI. It's going to be easy. The reality is that if you don't want the cultural change, the cultural shock, it could be the life or death of the AI strategy

because the people is going to say, yeah, we are using AI, we are going to, but maybe they are not creating AI use cases themselves. Maybe they are not improving what you have or maybe they are just running the AI, going out or they don't directly run the AI.

Pillar: Data Governance, Ethics, Privacy, and AI Act Risk

1Data governance, we are in Spain, we have the AI Act and RIA in Spain and if we don't cover that we may be in a big trouble. One part is of data, that our data remain private and another one is about

use cases risk that this we have to talk about before, that you cannot for example For example, you say I upload 100 curriculums to Chatipedi and say, okay, tell me which one is better and which one I should hire, because it's the highest risk of the AI Act. And it's responsible that it's also close to this part of the AI Act. But I want to, as I said before, go a little bit deeper.

Practical Assessment Demo: Measuring AI Usage

If any of you want to scan it, this is going to go to a website that is going to help us detect our individual AI use and our company AI use. So, later if you want I can save it again.

We have four companies and four people. I'm going to focus all these four companies.

Let's speak in English. I'm going to be using one tool that is called Super Whisper for not typing.

Yeah, I think we have time. Okay.

Live Volunteer Walkthrough (Case: Paul)

Ok, so if there is any volunteer we can do it on their case, if not I will do it on Zapping. Any volunteer? 5, 4, 3, 2, 1, come here.

What's your name? Paul. Paul? Paul. Paul.

Ok, so first of all tell us a little bit about your company. We have an AI company that makes agents and also makes information to companies. We make use cases, detecting pains to help people. So if you are an AI company probably you are going to be using, I expect, using AI properly.

Okay, now, creating agents, we will start with the first pillars, resource available. So you basically need to answer these first questions. Beyond your core technical team, how many people are on your company? Eight. Eight people. Okay.

So beyond the technical team, how do you identify specific employees with not technical roles using AI you just say about the first question and answer yeah yes what we do with each people in our team is like to see how is the process they make right now no like normal way like five years ago without AI and then think and change

the position to help in all the things that are operative to put the tools we need for example we get a new sales person in the team so she had a tool for meetings the one you said before no to pro to profile the clients the sector and then to call them and like just focus on the client and not losing time making like a prospection so we could say that all

your team is even they are not technical using ai yeah i want to share with you that uh we work three three months and we duplicate the capacity like McKinsey said that is 25 % but if you are like AI driven and helping people you can in Monday we we

duplicate the tasks we may yeah so we are running out of time so no no no no come here but go straight today and to answering the question so this is not about the answer.

And later we have to answer, do you have any defined annual budget for your AI usage? For the AI tools, AI information, AI training, consultancy? More or less is 15 % of the budget. Okay, great.

The rest is people. And if you want to specify if you have any partnership, but I think with that we We have plenty for, I suggest you to, we have more questions if you want to stay. Oh, yeah.

So we can, well, we are running out of time, so basically I suggest you to do, thank you, Paul, to do that in comfy with time, but this basically goes you through all the steps you have to consider.

The only thing missing here that I suggest you to do it or with this, this is a prototype that you can access later, but it will be available this suggest that I'm going to tell you on it.

All the people of the company should go through this test or a similar one for identifying specifically okay are you using AI? So I have showed you that there is one for companies and one for people.

So all the people should go to the people and the executive team should go to the company. This basically will deliver as the example of how good or bad they are using AI.

Example Output: How to Read the Assessment Results

I'm going to show you an example. This is bytecode in lovable, just for new information, in case you ask.

And we can come here and see how it could be be shown the example of how good or bad are you using AI. So this basically

tell you okay on what top of you say it's AI are you and okay what you should be focusing on and what is your best point of working so this the the core value is when you have a team report of 20 people 50 people 100 people that have have given all this information to the company.

Wrap-Up and Q&A

So now, let's go to the Q &A.

And I really suggest you for getting all the value to this talk to do it at home, because it's really long for doing it in 15 minutes.