What I'm gonna talk about today is how man Ray has harnessed AI power in a productive way to help companies basically hurt data in a hybrid cloud environment and achieve, or maximize the value out of cloud where everybody's selling us productivity and agility making business faster. But at the end of the day, we are getting stuck in the world of data. So just numbers. 97% of organization are using cloud services or looking at least to use cloud services. The biggest trend in the past decade case. But the, the key selling point is basically introducing functions such as a collaboration, agility, operational cost saving and so on and so forth. But it comes with challenges as you can see, personal data protection and add to that, the, what we called dog data, the data that we don't know anything about, okay, it's not just the databases, the SAP databases, the structured databases. It's all the files that have been accumulated over the years by our businesses. Not only on-prem, but also on any other systems. And this is the key problem here.
So we all know that data has been exploding. We also know that 90% of this dog data is unstructured and this accounts, again, 90% of all data in the organization, but the pace of growth is what is traveling here. It says here that up until 2003, the world has accumulated the just five exabytes of data. And basically today we create this amount every two days, which is enormous. Think about the, think about the pressure on the, on the companies. In this case, it's not just about the data. It's the flow of new systems into our environment, the cloud software tools and services that entering the core business processes are actually putting a light on how critical it is to control and govern the data across these platforms. It's much easier to collaborate now, but whereas the data now, once I collaborate on the data, it's not just about controlling the data.
It's an environment of diverse forms of the data and the scale of it. This is where AI actually provides the value. And I'll show you some examples later on AI processing unstructured data in an unsupervised manner, expose the user to a dashboard, which all the information about this dark data is being grouped into actionable categories. This is the output. Basically think about how much time AI can save in putting that data into that order. This information helps the organization to flow those categories. All those lists into protection tools, easy, of course it needs to inspect. It needs to believe the AI has done a good job. Therefore it will will provide it some analytics tools. This is the first step, but the next step would be to flow this information into protection tool. Therefore AI plays a significant role in automating this processes. And as we see it, I will give you several examples later on, but you can see several of those categories, the ability to recommend on minimizing data, which is an important task.
When you want to migrate to the cloud, you want to use cloud services. You want to use it with data that is being used. Think about purposeful processing from GDPR requirements, okay? Identify piles of PII that hasn't been used for years, you should delete it or securely migrate the data to the cloud. Find the data that is being used, but it needs to be encrypted on the way to the cloud while being shared and collaborated and so on and so forth. I can provide you several examples. So example number one, the immediate value would be to minimize the human dependency by having AI, doing its own training over your data. Once completed this training provides analytical capabilities to the end user to verify this analytics, is this really data that needs to be minimized? The second stage would be AI tracking the user feedback loop.
Yes, I accept this recommendation. No, I need to refine this recommendation. I need to filter that data by specific location and so and so forth that will actually feed back loop into the models. AI builds those models and track the data in that essence, this saves tons of hours of human and money, which is basically scares in this environment. There's no much skillset to analyze this amount of data, but it's not just being just, you know, general speaking data, but also data that has been processed in the past by end users, office 365, for instance, in this case can provide your capability to automatically label files, right, and set the policy on the files. But of course, users cannot do it over all the files. AI can capture that tagging and labeling and propagate that to similar data. Therefore again, leverage existing investment done by the organization to take it to the next level.
But you know, it's not just about software. It's not just about those capabilities. Also around the hardware stuck. It's very important in today's data center. Reality to actually free up as much CPU power to the traditional applications. AI is, is known with its capability to work very good with GPU power. You can see the improvement ratios here in basic tasks of AI, whether it be single processing or generation or clustering mechanisms, all those can be accelerated. The cost is stuck very fast. This is very important when we are going forward to adopt cloud or hybrid mode and set this capabilities again, architecture is also influencer. AI can be distributed using microservices, architecture, therefore reach to the endpoint and bring the AI to your endpoints. No matter where their endpoints are going to be, therefore help, therefore help govern the data and overcome the data governance control challenges. All those will help companies in a much more effective way to get through the pile of unstructured data and all the challenges around data privacy and cloud adoption. Thank you.
Thank you. And.
