I'm really happy to be here and thank you for having me. I'll take a few minutes of your attention before lunch and I hope I will be as interested as your smartphones and laptops. So let's take a look at more strategic and more organizational part of how probably to launch or maybe start your AI startup within your enterprise. So probably we'll have today a few things to discuss. So of course where we are right now in terms of AI revolution or evolution at this moment. So challenges and also strategies and some case studies. And we'll sum up everything about, so let's start, where are we now?
Right now? So as you can see, so the first point, which is point A. So where are we exactly right now that the only very, very few companies and enterprises are using and deployed your, their AI applications and also the AI in the their company.
So almost one fifth of enterprises are started to playing around and probably piloting the deployment. So that means that they started to working on something that probably might be in the future deployed on their production and start working on on AI and supporting this. And most of them unfortunately doesn't, don't, don't use it. So the point B.
So probably the how I can see also based on some statistics and some information from different sources, we will have probably expected. One third is expecting innovation and growth while deploying AI into their enterprises and also improved. So that that's very, very important. So improved efficiency and productivity of enterprises, people, teams, departments, et cetera. And it of course it'll reduce costs. So as you can also see that the last point that in six years, so we'll be 10 x times that we can, we have right now this market.
So that means that we have like 3.6 billion at this moment or 3 36. And then in 10 years we'll have 10 times bigger. So who we are, so at, I'm founder of this term, so I'm running a company for seven years right now. So we build custom AI and LM based software. So probably we help com our enterprises and tech companies and startups to drive their ROI by three elements. So hyper automation, enhanced decision and per hyper personalization. So our in a very nutshell, so our, our mission is to help companies to adopt AI and to make people focus on what matters the most.
So probably on more strategic and their lives as well. So we are number one, a land based company in Poland and we also recognize backlash on the last year as a top AI company. So I'm personally an and a entrepreneur.
I launched a few startups before. I also advise for pre-seed and seed startups as a strategic partner. I also recognize by linking as a top air voice, I work globally with enterprise and startups with different budgets, started with a few thousands of euros up to millions in, in their budget.
So that's, that's also important because I would like to underline that everything that you'll see is probably based on our hands-on experience. It's not just a chat GBT generated presentation, it was based on our years of experience and the back of of course with the engineers who are working on this on a daily basis. There are a few challenges of course while implementing ai. The first one is the very broad well what we can see based on, on probably our conversation and also our contact with customers that they are, they don't know what they don't know.
So that means that they are, don't really know what is the capabilities of large language models at this moment and AI in general. So the second one is of course it's ethics and privacy, which is new, something new of course it needs to be more specific by by this and regulations that are also going very fast and it's changed and probably you need to stay up to date with those changes and technologies, which also requires implementing or changing your implementation strategy data. Of course everything that probably new currency or someone new that is a currency of of now and future.
We need to have a lot of data and it should be prepared in some way to make it usable for ai. So resources allocation, so budget team and also computation for learning AI in the future. Scalability. So having in mind that to, to first launch the first project in a scalability in the future.
So probably it should be scalable from day one and also fast pace. I think it's also imp impact the, the launching and also the, the AI companies in general. The environment changes so fast. So there are also a few specific challenges within the enterprises. So the first one is alignment with the strategy.
So probably, of course enterprises have their own strategies with, which might not be aligned with AI implementation at this moment, but some revolution comes in, et cetera.
So change management, most or few employees that we can observe that are really resist and may resist from implementing AI or even trying to pilot with ai integration with the existing systems, which is also very challenging, meaning that IMP implementation something new into, or really it might be old school system or something in different technologies with different com, different departments on on also different time zones, et cetera.
It might be very really challenging not only from the technical perspective and cybersecurity. So probably this is a new target for cyber attacks.
So we, we need to, to think about this also and probably how to do this, how to foster the startup within the enterprise. There are three elements that I can see, of course we can talk for hours.
I, I guess, but I, I found that I, I wanted to focus on three most important elements at the very beginning and very first phases. So the first one is a leverage your strength. The second one is implement really smart. And the third one is use external experts. Let's start with the first, first one. So if you are an enterprise or running or a part of enterprise, so as a every single company on earth, you have your own strength and weaknesses, right? So just simply SWOT exercise so you can analyze what is your strength for most of enterprises that we were speaking is resources.
Of course bigger companies have bigger budgets and bigger access to their resources. It's a data availability. So probably, hopefully those companies collect their data during like years and have their warehouse of big bunch of data that might be usable or used in the future because of resources and better, bigger budgets or better access to resources. They might also experiment with the AI solutions and of course talents partners to, to involve.
So probably having in mind that budget of course and also different established brands on the market might also engage or involve better talents and partners comparing to like startups or smaller companies and of course infrastructure already in place. So it might be very, very quite easy to implement for these kind of companies.
So talk, having in mind those a few example elements. Of course we, we can go deeper into each one. So the first one is data.
I i I truly believe that it's a really currency of of, of now of present and the on the future. So the better data and the more data we have, the better per performance we can achieve in the future because the better results we can achieve in terms of training and also wide, wide data elements. The second one is the resources. If your company has better resources. So that means that you are able to engage better par partners and talents.
So just go to the market, look at what you're really need and then get back and understand your, I will take later about this. So experimentation as I mentioned. So sometimes for different use cases you might need to experiment and possibly you might have a better results comparing to a standard use cases without experimentation. The second one is implement, implement smart, what does it mean in in practice?
So we can see that from the project that we are engaged and we are involved. So we can see that there are two perspective very important.
So the first one is a human of course and the second one is a process. So having in mind that we have two probably filters to think about before implementation and also having right expectations regarding the project and regarding the first phase. So in the first phase, so we can probably separate or divide into three phases regarding the human filter. So the first one is the AI should support humans. The second one is human should support ai. And the third one is, which is hyper automation.
Of course everyone is want to to have this one right now, but in in real world it's like just a case For next few years AI supports AI and probably humans' role here is just to orchestrate and play around to configuring both AI to for hyper automation.
The second filter is a process. So from the perspective of process, very important to also divide this into three stages. The first one is improve existing process with ai. So probably not trying to make a revolution, rather make evolution of your existing process. The second stage is redesigning your process around ai.
So once you are you confirmed and you achieved what you wanted the first stage, you can move to the next stage which is redesigning the process around more ai. And the next one probably set new boundaries.
That's, that means that we really don't know what we don't know. We can see and observe that there are a lot of new business models are right now regarding this revolution. So that's the the last stage of course as a, every strategy should be measured, right? So we need to define what is the success point and of how we will measure this.
So having this in mind, it's much better to benchmark in the future. Thus the first stage, second and third stage is achieved and also measured in a good way.
And do, did we get the results we wanted to get? Very simple probably rule or maybe technique, but unfortunately most of customers and probably most of companies that we're observing, they're not following this. So starting with the coherent action. So of course starting with the brainstorm idea, prioritize this. So some example criteria might, might be here. So it's a feasibility probably how feasible or not this solution is at this moment regarding the technical advancement et cetera. So value for company, for process, for your business or department, et cetera.
So how big or what kind of value it'll give you once you implement AI into your process. The sec, the second one is effort. So effort means that how much money budget or how much time you'll spend on it.
Time is also the second part of this effort and strategic alignment. So does it align with your strategy overall with your enterprise strategy or not? And then only after this few steps. So try to build your team or probably get another resources on board which will be very crucial for your next phases.
And only then kick off the process of building something and build a pilot and measure it, make a retrospective, collect your feedback from all stakeholders and get back to your development process. That's also, we also observe very, very interesting element which is reusability. So keep in mind that building something from from scratch, it's very probably cost. So probably you'll spend a lot of money and a lot of time to establish this process and to develop something.
And after this, so we observed it, a few companies reuse the same technology into within their enterprise or even within their departments, which is really great for them because they pay one.
And then just for implementation to different spaces, different processes and and departments. I would rather advise to use external experts as I mentioned before. So the one of the challenges is we don't know what we don't know.
So to having, having those experts on board so we can probably broaden our knowledge and understanding and awareness of at least capabilities what, what is feasible, what is not feasible, et cetera. So probably to make it in a good way. So there are two, two simple steps. The first one is understand your needs. So probably assess your starting point. So if your company or your enterprise already have this kind of team. So there is no need to engage another team if they will be able to dedicate some resources for your new pilot projects. So it's okay, so that's fine. Let's start with this.
For some reason companies don't really want to engage for experimental projects or r and d projects at this stage.
So they would like to focus on their core business and instead engage someone else who would like who, who will develop this for for them. So analyze your strategic requirements, which is might be time to market sometimes and also environments. So as an example, three of your biggest competitors might already implement this in their enterprises. They reduce their cost like twice and probably now it's not nice to have but it's a must have to implement this.
So, and that this will impact also your time to market strategy and understanding of of bigger environment. Once you'll understand this and once you understand your needs, so start with identifying the scope of work. So probably what should be done in the first stage, you may do this on your own or you can also consult this with companies who may assist you with the development process or some strategic companies, et cetera.
So once you identify your scope of work, so try to make a data-driven decision, which means that engage your potential vendors or potential partners in the discussion and try to scorecard them later evaluate they take their capabilities and also look at their portfolio. This is a must have because a lot of companies who are working that, that are working in just simply software, they are of course put the star on their flag that they're AI company but that's not true.
So you do, you really need to evaluate them from the AI perspective. You are going to work with them in this context. And then the best really, really best strategy that we can see at very profitable and have has bigger chance for ROI is negotiate your contract terms to the model pay as you go. So probably you don't pay for something that you really need, don't need to pay.
So that means that you pay only for hours solutions that you get and start with the very baby steps to build trust. So start with a simple one project.
Once you'll build your trust, you can move forward with another project and engage, engage it more. As you can see here is three categories in-house team, software development team and AI development team. So probably in a very short summary. So it's quite difficult to be on track and also to monitor every, everything in AI industries. It's evolving very fast.
The expertise is needed might be on a high level and it won't be cost efficient for to to involve your in-house team if you don't have such a dedicated team for this project or a software development company, which is just focusing on few areas at least 3, 5, 4 10, et cetera. And as you may know probably is the more you focus on something one the better results are right?
There are a few use, there are a few case studies that I found very, very interesting.
The first one is Bloomberg, as you may know that they built their own GPT which is alternative to Chad GPT, but they focused only on their specific niche, which is financial niche, right? And they built this fi 50 billion parameters, large language model. They analyzed their strength and probably opportunities. The opportunity was collect all their data and utilize this in a financial niche and probably they did it and they utilized this in a very smart way and as the, as the result. So they got outperforms over other L LMS such as charge GPT even that time.
So they built an opportunity for opening new markets, entering new niches and building new products, offering and revenue streams. They increased or improved their operational efficiencies and they built a very unique competitive advantage.
The second project which is we work directly within this project. So it was a company that has more than in 15 countries, their sales teams.
The challenge was that they probably had the direct dependencies or connections between their sales teams and probably the capacity, meaning that the more sales guys they have, the more sales revenue they can generate. So they wouldn't, they would like to make it more scalable and probably de re or reconnect or deconnect this connection between headcount and capacity of their revenue. They got a thousands of inquiries every day for every department.
And what we did for them that we built a custom conversational AI that was automated, all inbound email communication and probably utilizing of course their existing data and their product catalogs. Meaning that the only one thing needed to do with this application is that to open your interface, look at priorities on your emails you got from different potential customers, which already scored and prioritized for you as a sales guy.
And the next generate response and send this inquiry or probably this proposal to our customers.
So probably they don't really waste time on looking to the product catalogs, which has like thousands of position. They focus just on answering initial responses or getting, sending initial responses. They focus on account management instead of just working with the shallow tasks, et cetera. And by this they increased their revenue overall and it reduced a lot of manual work by like around 64% I guess. And they achieved ROI in like two monthly. So once this solution was launched, so they got this ROI in two monthly.
So that's, that's all from my side. If you would like to keep in touch with on LinkedIn with me. So here is a tag. You can just simply scan this and me on LinkedIn. We also wrote a book which is the LLM book, which might give you a more and broad understanding of what LLM capabilities are, what kind of technology it has and how to implement this and how to start working with this. I'll be today for a whole day and also as you can see in going to a cruise for the first part.
So if you'd like to talk to me, I'll really happy to discuss with you your challenges and your understanding or, or even just share my expertise. Thank you very much.
Thank you. Thank you very much. We really appreciate you sharing your expertise here.
So yes, please if you have more questions, come to anoni. Thank you. Thank you.
