How artificial intelligence is moving into EIM and what we need to consider
Do chatbots dream of electric sheep?
André Vogt, Director EIM at CENIT AG.
Published in: DiALOG - THE MAGAZINE FOR ENTERPRISE INFORMATION MANAGEMENT | MARCH 2018
Published in: DiALOG - THE MAGAZINE FOR ENTERPRISE INFORMATION MANAGEMENT | MARCH 2018
ChatBots for customer communication, AI for selecting, recognizing and structuring large amounts of data, and autonomous aircraft, vehicles and robots. IoT, networking and cloud, these are the topics of information management in the here and now.
With the modification of Philip K. Dick's book title "Do Androids dream of Electric Sheeps" from 1968, I try to transport the thoughts of that time on the question of man, machine and artificial intelligence into the context of EIM.
In 1968, the ideas about androids or replicants and the questions about the boundary between humans and technology were pure fiction. In the first film adaptation by Ridley Scott in 1982, there was at least already the network and the vision of neural networks and thus artificial intelligence, and in 2017 with the last film adaptation, we must ask ourselves, where do we stand? Professionally? Technologically? In the context of enterprise information management?
Responsive Image
My Enterprise Information Management understanding comes classically from the management and provision of corporate data and information. The automation of internal processes and thus the standardization and safeguarding of procedures has become a quasi industrialized service. Thus, the current focus is on the extension of processes to the actual customer in order to map a complete "360 degree customer communication" in an automated way. CENIT sees the extension of additional channels, but also of additional media as the task of the near future. Messenger and its chat protocols, audio information from phone calls and conversations, and personalized and interactive videos are our focus.
Natural language or "natural language processing" is a major trend that is being driven ever faster and further by artificial intelligence and self-learning systems, but is really nothing more than an additional channel as part of the multi-channel strategy of "360-degree customer communication". In an ideal world, a call, a fax, an email and the use of Siri, Alexa and Cortana would have to result in the "same" IT-supported process flows.
ChatBots promise the solution here - they are increasingly populating the Internet and becoming universal service providers. They assist as news suppliers, learning aids for foreign languages or service mediators - and you can "chat" with them almost as you would with humans in customer service. But this actually "old" technology did not experience a real breakthrough at the beginning. Online retailers were the initiators. Various of them have already switched off their rule-based chatbots again, because a chatbot is also a kind of signboard for the quality claim associated with the brand, which does not suit every product.
In 1966, Joseph Weizenbaum used "Eliza" to demonstrate the basic principles of linguistic communication between humans and computers. This was actually the start of what is currently being promoted in the media as the chat bot trend. The actual trigger, however, is not so long ago, but is related to the mobile and cloud trend of the last decade.
Eight years ago, Siri took away the fears and worries about natural language communication with machines in an almost playful way. Rule-based approaches were already plentiful, but by "humanizing" it as a personal assistant, new desires or requirements for communication settled into the minds of users - and thus software vendors.
The other major platform providers such as Google (2012) and Microsoft (2014) quickly followed suit. Naturally voice assistants served as conversation partners for private use. The acceptance of these solutions in the private sector positively influenced the openness to imagine, test and introduce solution scenarios in the business sector as well.
With Alexa, AI has made its way into the living room. The questions in our consulting projects are currently much more focused on the deployment scenarios, target groups, and dialog management than on the technology. It has been recognized that the technology or platform does not really matter, since the added value of computer-supported communication cannot be evaluated in terms of IT costs and IT functionality.
What the solutions have in common is that a new customer communication channel - chat - is considered to be quasi-established and is provided via various end devices (browsers, tablets and SmartPhones, etc.). However, the technological implementation of the dialog differs considerably among the solutions and platforms on the market. In addition to very simple rule-based approaches, the spectrum on the other hand ranges all the way to artificial intelligence-based approaches such as those at Microsoft or IBM, where the integration of additional cognitive capabilities is virtually standard, so that image recognition, translations, and predictions of offers for the next action in a chat bot are possible in the simplest way.
So what does the ideal ChatBot look like? It must be capable of learning. To do this, it interacts with the customer, observes him, notes what he likes and what he doesn't like so much, registers his feedback, and thus builds up an aggregated knowledge base that grows over time. A central prerequisite for this is that the ChatBot is provided with the appropriate data, on the basis of which it can learn who the customer is and what he likes. This cannot be mapped without Artificial Intelligence, so you have to deal with this topic in a positive way.
CENIT offers a maturity model for the classification and grouping of chatbots in order to narrow down the right approaches for the start of the digitalization strategy in the customer communication environment in a focused manner. This is the only way to reduce the risk that the idea of introducing a quick solution is not destroyed by the discussion about an "eierlegende Wollmilchsau".
Let's not kid ourselves. The number of people who support digitization with AI and chatbots is statistically above 68% - but only up to the moment when they are personally involved in the conception and description of requirements. Suddenly, concerns, problems and fears are addressed - almost a bit like the questions of 1968, motivated by Philip K. Dick. Data security, the cloud discussion, the concern of job substitution, and again and again the quality of communication - in terms of content and technology. Thus, AI and ChatBot projects must always be accompanied by appropriate change and associated motivational communication. We recommend pragmatism when using AI-based dialog systems.
"Just do it" is the motto. At CENIT, we prefer to show the solution directly so that we can work out the special features in the dialog and process together with the customer. And here we don't just look at the frontend or channel, but also at the integration of the existing systems. This creates significantly more change than the discussion about the legal issues of a cloud solution or questions about data sovereignty and IT technical requirements.
We were able to gain good experience with our SmartChat solution for vehicle damage reporting: In addition to a special workshop method for the joint development of the dialog flow, we were also able to gain technological experience in the implementation on various platforms, which was incorporated into our maturity model. The actual challenges, however, are more likely to be solved at the interdisciplinary and procedural/methodological level. Accompanying the change is key.
So is AI now taking over? Are people being replaced? The human being is still at the center. Not only in the project for the introduction of a ChatBot as part of a digitization strategy, but also as a user. There will continue to be communication that will only take place in personal form. Trust, genuine emotions and the experience represent a value in themselves. But do I have to pay for that if I want to report a defect in my Internet connection at night or if I'm on the road and have already spent many hours talking to other people. These are my positive examples where I am happy to use a ChatBot and grateful for this offer.
For every innovation there is a positive and a negative field of application. The choice is in our hands and our attitude toward the possibilities of technology. What will not stop is the human quest for innovation. The often quoted sentence " What can be digitized will be digitized and automated" and the concerns expressed against it we all know from the introduction and/ or use of EIM solutions in our private and professional everyday life. We should see the opportunities and manage the risks. What the ChatBots are dreaming of, I am not able to say in conclusion. But I can promise that chatbots deployed at night will at least let human colleagues in customer service sleep well and dream. You just have to want to see the positive.
My basic attitude is generally one of "Why not...?". But another, recently published forecast from Gartner now makes me personally a little uneasy: As early as 2020, the average person will communicate more often with a chatbot than with his or her life partner.
ChatBots promise the solution here - they are increasingly populating the Internet and becoming universal service providers. They assist as news suppliers, learning aids for foreign languages or service mediators - and you can "chat" with them almost as you would with humans in customer service. But this actually "old" technology did not experience a real breakthrough at the beginning. Online retailers were the initiators. Various of them have already switched off their rule-based chatbots again, because a chatbot is also a kind of signboard for the quality claim associated with the brand, which does not suit every product.
In 1966, Joseph Weizenbaum used "Eliza" to demonstrate the basic principles of linguistic communication between humans and computers. This was actually the start of what is currently being promoted in the media as the chat bot trend. The actual trigger, however, is not so long ago, but is related to the mobile and cloud trend of the last decade.
Eight years ago, Siri took away the fears and worries about natural language communication with machines in an almost playful way. Rule-based approaches were already plentiful, but by "humanizing" it as a personal assistant, new desires or requirements for communication settled into the minds of users - and thus software vendors.
The other major platform providers such as Google (2012) and Microsoft (2014) quickly followed suit. Naturally voice assistants served as conversation partners for private use. The acceptance of these solutions in the private sector positively influenced the openness to imagine, test and introduce solution scenarios in the business sector as well.
With Alexa, AI has made its way into the living room. The questions in our consulting projects are currently much more focused on the deployment scenarios, target groups, and dialog management than on the technology. It has been recognized that the technology or platform does not really matter, since the added value of computer-supported communication cannot be evaluated in terms of IT costs and IT functionality.
What the solutions have in common is that a new customer communication channel - chat - is considered to be quasi-established and is provided via various end devices (browsers, tablets and SmartPhones, etc.). However, the technological implementation of the dialog differs considerably among the solutions and platforms on the market. In addition to very simple rule-based approaches, the spectrum on the other hand ranges all the way to artificial intelligence-based approaches such as those at Microsoft or IBM, where the integration of additional cognitive capabilities is virtually standard, so that image recognition, translations, and predictions of offers for the next action in a chat bot are possible in the simplest way.
So what does the ideal ChatBot look like? It must be capable of learning. To do this, it interacts with the customer, observes him, notes what he likes and what he doesn't like so much, registers his feedback, and thus builds up an aggregated knowledge base that grows over time. A central prerequisite for this is that the ChatBot is provided with the appropriate data, on the basis of which it can learn who the customer is and what he likes. This cannot be mapped without Artificial Intelligence, so you have to deal with this topic in a positive way.
CENIT offers a maturity model for the classification and grouping of chatbots in order to narrow down the right approaches for the start of the digitalization strategy in the customer communication environment in a focused manner. This is the only way to reduce the risk that the idea of introducing a quick solution is not destroyed by the discussion about an "eierlegende Wollmilchsau".
Let's not kid ourselves. The number of people who support digitization with AI and chatbots is statistically above 68% - but only up to the moment when they are personally involved in the conception and description of requirements. Suddenly, concerns, problems and fears are addressed - almost a bit like the questions of 1968, motivated by Philip K. Dick. Data security, the cloud discussion, the concern of job substitution, and again and again the quality of communication - in terms of content and technology. Thus, AI and ChatBot projects must always be accompanied by appropriate change and associated motivational communication. We recommend pragmatism when using AI-based dialog systems.
"Just do it" is the motto. At CENIT, we prefer to show the solution directly so that we can work out the special features in the dialog and process together with the customer. And here we don't just look at the frontend or channel, but also at the integration of the existing systems. This creates significantly more change than the discussion about the legal issues of a cloud solution or questions about data sovereignty and IT technical requirements.
We were able to gain good experience with our SmartChat solution for vehicle damage reporting: In addition to a special workshop method for the joint development of the dialog flow, we were also able to gain technological experience in the implementation on various platforms, which was incorporated into our maturity model. The actual challenges, however, are more likely to be solved at the interdisciplinary and procedural/methodological level. Accompanying the change is key.
So is AI now taking over? Are people being replaced? The human being is still at the center. Not only in the project for the introduction of a ChatBot as part of a digitization strategy, but also as a user. There will continue to be communication that will only take place in personal form. Trust, genuine emotions and the experience represent a value in themselves. But do I have to pay for that if I want to report a defect in my Internet connection at night or if I'm on the road and have already spent many hours talking to other people. These are my positive examples where I am happy to use a ChatBot and grateful for this offer.
For every innovation there is a positive and a negative field of application. The choice is in our hands and our attitude toward the possibilities of technology. What will not stop is the human quest for innovation. The often quoted sentence " What can be digitized will be digitized and automated" and the concerns expressed against it we all know from the introduction and/ or use of EIM solutions in our private and professional everyday life. We should see the opportunities and manage the risks. What the ChatBots are dreaming of, I am not able to say in conclusion. But I can promise that chatbots deployed at night will at least let human colleagues in customer service sleep well and dream. You just have to want to see the positive.
My basic attitude is generally one of "Why not...?". But another, recently published forecast from Gartner now makes me personally a little uneasy: As early as 2020, the average person will communicate more often with a chatbot than with his or her life partner.
CENIT is the partner for successful digital transformation. With CENIT at their side, customers have far-reaching options for optimizing their horizontal and vertical business processes. Innovative technologies from the areas of Product Lifecycle Management, Digital Factory and Enterprise Information Management create the basis for this. The competence of CENIT consultants results from the combination of interdisciplinary process understanding and deep technical expertise. The end-to-end consulting approach gives CENIT customers the certainty that their solutions are created with an understanding of their entire value chain. As a holistically positioned partner of its customers, CENIT assumes responsibility for everything from consulting to the introduction of innovative IT solutions to economic operation. The CENIT team adapts to the specific situation of the company and thus ensures the practical relevance that makes measurable operational optimizations possible in the first place. For more than 25 years, CENIT has been realizing competitive advantages for renowned customers in key industries of the economy. CENIT employs around 800 people who serve customers worldwide in the automotive, aerospace, mechanical engineering, tool and die, financial services, retail and consumer goods industries.
www.cenit.com
