Hello everyone. And welcome to this co Cole webinar on the way to becoming a cognitive enterprise. My name is Anne Bailey, and I'm an Analyst with co our Cole. If you were expecting to hear from my colleague Matthias, he unfortunately cannot be with us today. So I'm delivering this webinar on his behalf. So if you are unfamiliar with keeping a coal, this is a, a short introduction to what we do. So we deliver focus, content and services in three areas on identity and access management, cybersecurity and artificial intelligence.
And we do this in a variety of ways through reports, webinars like this one advisory projects and conferences. If you are not familiar with Casey plus, this is our new content and research platform. So this is a great way to become familiar on a huge variety of topics in a simple user friendly way. So I encourage you to check it out, especially topics like the one we are covering today,
The events coming up include some very high profile ones like the EIC in may of 2020. So I'd encourage you to definitely check out this event and others like it that are coming up in the next year.
And now for some information about this webinar in particular, you are muted centrally and we are controlling these features. And so there's no need to worry about muting or unmuting yourself throughout the webinar. We are recording this and you will have access to the recording and the slide deck, as for questions, feel free to submit a question in the go panel and I'll be able to take at and address towards of the webinar today are to discuss becoming a cognitive enterprise and really deconstruct what this means. And some key points to look for during this transition.
And of course, to address your questions. So let's get started.
So a cognitive enterprise is really made up of four components. These are listed below as adaptive, interactive, iterative, and contextual. And when taken together, the cognitive enterprise is really next generation business model, which is leveraging some converging technologies and data. So these four dimensions really bring a cognitive enterprise to the next level value.
So an adaptive aspect is that these cognitive enterprises are learning from realtime data and they're able to record incremental data updates as they occur and be learning constantly the cognitive enterprises interactive. It interacts with users with other systems, processes, devices, and cloud services, and a cognitive enterprise is iterative. It's able to make relevant recommendations, ask questions, understand context, use additional information and remember previous interactions. And lastly, a cognitive enterprise is contextual.
It's able to understand, identify and extract contextual elements, such as meaning syntax, time, location, business, area regulations, user profile process, and task and objective.
So the cognitive enterprise has a few promises which are able to be made available. And so let's take a look at these in practice, which are for practical improvements, which we can take as an example. So the first achievement to unlock that cognitive enterprises can relieve and empower your employees. So your staff is better than bots and can pro perform much more meaningful tasks.
As long as the burden of very repetitive tasks has been shifted to an automated process. By shifting that burden, you can enable your employees to become much more creative and engage in much more thoughtful tasks. So second cognitive enterprise can build innovative and meaningful customer experiences.
So shifting towards a individualized customer experience, instead of simply customized, really allows you to take your customer seriously and bring them to the next level of service. A cognitive enterprise can create new services and offerings.
So by leveraging the access to data, new technologies and new ways of interacting in an ecosystem, this can improve the quality, the amount of base data, which then opens up opportunities for new services and offerings and enterprises can understand reason and interact. So when you are in touch with your peers, this can enable meaningful conversations and go beyond the simple questions by asking for details and deliver genuine value.
So let's take a look at the big picture of the cognitive enterprises.
There are six major components that make the cognitive enterprise possible, and it should be noted that these components cannot be taken individually. They are only able to help an organization transform into a cognitive enterprise when they're taken holistically. So for example, you cannot take, you cannot have artificial intelligence without data. You cannot be a cognitive inform enterprise without cloud, as a new infrastructure platform. An APIs are the basis, the whole ecosystem, because everything is inter interdependent on each other.
So let's take a look at each of these components in a little more detail. So innovation is a big part of cognitive enterprises. This is the agility and culture change. The way that organizations work, deliver value. A cognitive enterprise usually has a hybrid infrastructure where it has deployment models in hybrid, secure, scalable, and standardized. And sometimes multi-cloud environments. The processes undergo huge transformations when going into a cognitive enterprise.
And this is enabled through analytics and smart decision making through support and communication data is the fuel for cognitive enterprises. And this can be open data, public data, subscribed data, enterprise proprietary data, pre-processed data structured and unstructured data and heterogeneous data
When discussing cognitive enterprises technologies usually end up in the spotlight. So remember that the technologies cannot alone transform a company into a cognitive enterprise. It really relies on these five other components to make the picture complete.
So these technologies we'll get into a little bit later into in the presentation, but some examples are of AI machine learning, blockchain, virtual reality, augmented reality, IOT 5g, or 3d printing. So one or more of these may be used in cognitive enterprises. And lastly applications APIs are a really foundational connection point in cognitive enterprises, and they allow the interaction between platforms, cloud, native applications, legacy, or adaptive applications.
So let's take a little deeper look at one of these six components, the cognitive technologies, the robotic process automation.
This is something that we'll continue to talk about later in the presentation. So only briefly introduce it here, physical robots. This is the automation of repetitive manual tasks with intelligent robots on the factory floor or at the point of sale. And this can improve the quality of the work performed. It can reduce the risk of errors and of course, increase productivity. Natural language processing is often called NLP, and this helps to improve the overall usability of applications. It enables data insight by creating intelligent communication models.
For example, chat bots, policy based systems are not AI per se, but it is the implementation of intelligent business trans business platforms that can be reduced that can help reduce the need for manual processing and complex scenarios. So again, not AI, but using intelligent systems, machine learning is using techniques to automate analytical models building. And this is using algorithms. So this helps companies do mostly predictive work like predicting customer behavior or product recommendations that all has many other uses. For example, in fraud detection.
And lastly here will address deep learning, which in a very general way, you can think of deep learning as machine learning onto steroids. And this is solving a class of problems such as image classification and object recognition. This is already used today in online gaming product categorization, adaptive speech recognition and state of the art authentication mechanisms. So I mentioned that we would come back to robotic process automation and here we are.
So let's take a closer look at what this means and how this fits in with both artificial intelligence and inside the cognitive enterprise. So a simple definition of our PA is that is emerging as a form of business process automation technology. It is a more advanced form of what we have seen already. And so this is based on the notion of a metaphorical artificial intelligence where physical workers or soft software workers enable enable automation of multiple step processes. So RPA systems develop their action list by watching the user perform that task.
And then they perform the automation by repeating those tasks directly in the graphic user interface.
And this automates the interactions with the graphic user interface, allowing data to be handled in and between multiple applications like email or ticketing, bookkeeping scheduling, and many more. So RPA is impacted by artificial intelligence, namely that AI can boost RPA specifically in complex use cases that can be used in conjunction together.
And so a few of these use cases we can see here, for example, anomaly detection, where in our PA system can help identify anomalies and out lower outliers in the information that is processed by the robot in text to speech to text again, simply as it sounds where an RPA system can convert text to and back again, and interact in various languages seamlessly. And in text understanding an RPA system bolstered by AI can automatically analyze and process large amounts of structured and unstructured data in bots.
RPA systems can interact with customers in a pseudo human way course in an automated way as well for self improvement, RPA systems supported by AI improve the quality of work continuously by assisted or autonomous learning. And in decision making RPA systems can support decisions and improve the quality of those decisions based on their learnings.
So if we want to leverage the power and the efficiency of RPA systems, we have to make sure that our use case really can support artificial intelligence as well. So there are a few prerequisites to finding an appropriate use case for AI.
These per requisites are first to have access to historical data. This is used for the training reference and improvement of AI systems. There must be a constant flow of real time data. You must be able to achieve a critical, massive input for operating and to have potential self-improvement in your AI tools.
There must be a repetitive task to be completed. This is the only way that you can really leverage automation and this unlocks the cost savings. If there is not a repetitive task to be completed, then it doesn't make sense to train your algorithm, to do this task.
And related to this, there must be a limited scope of control. The organization can operate in a contained system without significant interface with other influential actors. And so you can think of this as limiting the, the variables or the potential unforeseen situations that your AI tool would get into. If you have a very limited scope of control, you can limit and reduce the number of unforeseen circumstances that it would have to deal with. There's a very wide scope of control. You can't always control, and you can't always train your tool to deal with those unknown variables.
Well, so now that we have our prerequisite, these four items that we know our use case must have access to in order to be a good fit for AI, we can take a look at some different use cases. These five examples in different industries, we can use to take a look at in comparison with our four per requisites.
So in terms of historical data, these five fit the bill pretty well in predictive maintenance and manufacturing in network coordination and any energy production, fraud detection, and finance employee support, and AR and predictive demand in supply management, all of these should have access to historical data from within the organization.
We're shared within an ecosystem, their realtime data, given that you have the appropriate sensors in place, on your manufacturing machinery, or on your edge devices, in a, in an energy production capacity, you should have access to realtime data in order to apply your tool. These are all repetitive tasks and your employees would benefit from having support, having automated support to free up their creative capacities. And the scope of control is a bit more complicated. We can't describe it in simply a checked or unchecked box.
So taking a look at the first column, predictive maintenance and manufacturing, this seems to be a good fit.
In most cases, this would be limited to the own operations of a single enterprise, having access to their own machinery information about their own processes and not being too concerned with the processes and the wear and tear on the machinery of their neighbor company, neighboring company. So when focused on their own enterprises, the scope of control is limited, which makes it possible in terms of network coordination and energy production. This is very, very different.
This often includes monitoring edge devices, which may or may not be private property. This opens up the system to a lot of different variables, security variables, the validity of the data.
Yes, all of this makes the, the system or puts the system in an interaction which cannot always be controlled and makes it much more complicated to implement in fraud detection. It interfaces with members of a transaction, which is typically contained in an ecosystem or in a, in one company's branch of relationships, but not always. So here it depends. And employee support support in HR.
This is often limited to the employees of a single organization and in predictive demand for supply chain management, this is limited to your own operations, but in a very general way, because each relationship is usually interfacing with another organization and link in the supply chain, which can become very, very complicated. And so here, it's a bit harder to say what the scope of control is, and it's assumed that there are more and more unforeseen variables that an AI tool would have to deal with here.
So taking a look at the next steps and cognitive enterprise does very well because it can leverage an ecosystem. And so these are the ways that an enterprise can access or interact with AI products and solutions within an ecosystem. And we can start to see how this is built out. So first in the most narrow sense, we have AI data, and this is the foundation of AR artificial intelligence machine learning. Curtis. We need to be able to share data, have access, to open and public data. And all of this enables AI to improve because there is more to learn.
So next we have AI platforms.
So AI can be delivering baseline capabilities like text to speech, text recognition, visual recognition. And these are all increasingly provided as a service. And in order to access the service, you move through a platform. And so these platforms are used sometimes with, or without sharing data. So as we just mentioned, these AI services that apply based on capabilities, they can be targeted for much more specific use cases, such as automated text, understanding and analysis for legal departments.
And these are accessed through the AI platforms and they're combined with specific often shared sets of data.
Then we have AI solutions which deliver to the user, for example, by providing natural language processing or by translating text into other languages by augmenting the view of the factory workers.
So many, many different solutions, and they share services and data. So next on the largest scale, we have AI networks, more complex situations will not be built on a single AI solution. So for example, smart cities will combine vehicle data, which is comprised of several different AI solutions along with traffic and smart energy management data, and much, much more. So there were some improvements that we know for sure that we can access with cognitive enterprises, and these can be summarized as accessing efficiency and gaining added value.
And so these are the, the achievable benefits which we can have from using modified processes through the transformation towards a cognitive enterprise. So the first is trust, just balancing security and the user experience.
So trust can be gained in many different ways, but very important and foundationally through KYC processes, fraud, detection, and cybersecurity. And it's quite important for an enterprise to balance the security and trust with a seamless customer and employee experience.
So next is user experience, and this is moving from a personalized or customized experience to an individual experience. So the goal here is to move from today's first personalizing systems to real time systems that understand customers as individuals and enable personal and contextualized customer experiences. So you can think of the example of moving from a generic chat box saying, do we want to chat to, hello, Annie, what do you need today? Are you satisfied with the product you got last week?
So third is efficiency and we do this through automation and the, the automation enables an organization to cut costs. So these routine activities that are continuously repeated in the same or similar form are ideal candidates for the use of cognitive technologies, because remember that automation enables the burden from employees, from humans to access that creative work. And lastly, we have ecosystems, which is the really unique part here. A cognitive enterprise is able to rewire and exchange data and really leverage the power of a reciprocal relationship.
So for example, specialization and an intelligent supply chain could offer new concepts for highly integrated adaptive collaboration. And remember that in ecosystems, partial services are also possible. And if everybody delivers, what they're really good at applicable ecosystem is created, that enables new services. So for example, we can have one party which contributes speech to text in English and another party, which contributes English to German translation. And so these in combination deliver a tool with a much higher functionality.
Now, remember that APIs here are essentially the glue that allow the integration between these different tools and different platforms.
So how do we get there? This is the next logical question. The process two becoming a cognitive enterprise is of course, very dependent on your industry and the business case, which is made for it. And this is in fact, the first piece of, of advice that you should take with you is that there needs to be a strategy and a blueprint.
First, there must be an alignment of the business case made, and the technology, this alignment is crucially important, and if it is not aligned, then you won't get very far in this process. There must be a long term vision with actionable goals and built in adaptability.
So next really work on understanding how to leverage and contribute to a network and an ecosystem, cognitive enterprises, Excel within ecosystems, because there is this reciprocal relationship, particularly of data where you can provide and use data and technologies within partnerships, be able to scaffold one upon the other and deliver higher value. So remember that this is an iterative process and there must be continuous implementation. The automation of routine processes is key here and intelligent processes are the basis of new business models up new opportunities for you.
And remember that you can then dedicate your employees to higher value activities when you've unlocked that opportunity through automation.
So that concludes the main part of our webinar. Today. I'd encourage you. If you have a question, please submit it to the, go to webinar question panel and I'll wait a moment in case there are a few last minute questions. Okay. If there are no further questions, then like to close out the webinar by reminding you to familiarize yourself with the keeping your Kohl services and everything we have to offer.
And I hope to interface with you either through our reports, through an event or an upcoming webinar, you can be familiar with some related research on this topic. And I look forward to the next time we hear from you either with a report, a webinar or an event. So thank you very much.
