Artificial Intelligence (AI) seems to be a buzzword lately, promising techniques and tools that could influence our lives, our work and the way we do business. For many designers and enterprise architects, the question becomes: What would be the role of designers on all levels (strategic, EA, BPM, data, technical) in incorporating AI in a company? There are even ethical questions that may arise when instituting AI in a company, which you must take into account before making the change.
In this blog, I offer some starting points for thinking about and supporting AI transformations in your enterprise.
If we believe what analysts, media and other opinion makers are telling us, this is the year of Artificial Intelligence. Last year it was blockchain (or more specifically, the rise of the Bitcoin and other cryptocurrency) – this year it will be artificial intelligence. As we all know, this “hype” is not new. Most of the technology that is hyped has already existed for quite some time. Although we might characterize this technology as hype, some of it is still very relevant and useful for a lot of enterprises.
You are likely familiar with the well-known examples of AI in our everyday lives. The majority of people know AI from examples like Tesla (using AI to predict upcoming road situations), Netflix (recommendations based on the behavior of other viewers), Google’s Nest (self-learning thermostat) and Apple’s Siri (digital personal assistant).
But next to these well-known examples, AI is being used in other contexts as well, e.g. insurance companies using AI to calculate premiums (on a personal level) or the police predicting where crime will occur based on computer analysis.
Although not every enterprise will explicitly take steps to start using AI in the short-term, many companies will use other products that have AI capabilities. Applying new technology without thinking about the consequences might lead to problems, but new technology also brings new opportunities if used correctly.
The discussion about using AI, how it can impact your business and how it can improve your bottom line should start with strategy makers. With them, the long-term strategic goals are set, then the Chief Information Officers and Chief Technology Officers will think about how AI can help realize those goals. Then, the team will determine which data is available and how they can fully exploit the knowledge hidden in it to support achieving these goals.
Together with technology experts, data and enterprise architects, product developers and other experts, you can explore what AI can do for your enterprise. As agile development is the norm these days, you can start experimenting with AI, learn what benefits it offers to your organization and, if promising, choose to scale up to incorporate more processes, bigger customer segments, etc.
In that evolution, models can typically be used to keep everything aligned and coherent so you can evolve these models at the same time. Enterprise Studio can help you plan the implementation of new AI technology as well as provide you with models for all iterations to help you implement AI in your company.
The quick rise of AI technologies also introduces new challenges. One of these challenges is what happens when you process data or make decisions in your business process based on AI. Next to possible ethical questions, you might also need to deal with privacy regulations like the European General Data Protection Regulation (GDPR). A stringent EU regulation on privacy protection, which just got into effect in May 2018, affects any company worldwide who processes data from EU citizens.
GDPR demands that you can explain automated decisions in a meaningful manner to the individuals affected. This puts limitations on the use of AI, e.g. if a neural network decides you don’t get a loan, “The computer says no” will not be an acceptable explanation. You will need to be able to explain why the computer says no. Enterprise Studio is a perfect way to identify and maintain your registers on where and how personal data is processed in your enterprise. Although it will not help you understand the complex self-learning algorithms and neural networks of AI, it can help you to explain where and which automated components make a decision and which data it takes into account.
Although we cannot predict what AI will mean for us, we can start taking some precautions and measures. This might reduce the possibilities of “accidents” like e.g. police algorithms that are engaging in racial bias or road accidents with self-driving vehicles.
Precautions could include measures like using open source implementations of AI technology or making your AI algorithms transparent, which will build trust and fulfill the needs of the customers and regulators. Of course this comes with problems. Explaining how a neural network has learned to answer a certain question might be impossible. It is not just about the algorithms, but also about the data used to train all these self-learning algorithms. On top of that, how do we know the data used to feed a neural network is correct and obtained legally? Businesses will need to discuss this, and might even want to discuss if you’re prepared (and have the resources) to handle AI technologies ethically.
As the use of models can help you design and implement AI in your enterprise, it can also help you to keep track of the places within your enterprise where AI might be applied to decision-making. Next to practical reasons (e.g. help with compliance regulations, impact of change analysis, informing people), it can also be a reference point to discuss the use of AI in your enterprise on a more ethical level.
AI is very promising and can be of great value to many enterprises and their customers. If you are already using AI, you need to consider if you really know where AI is involved in your business processes, where the technological AI components are, and which data is being used. If you are unsure when answering any of these questions, you might want to consider having a chat with a BiZZdesign representative so we can help you with this discovery.
If you are considering using AI and need to define its use and want to discover how it will impact your business, we recommend that you start using architecture models with a platform like Enterprise Studio to develop and design your AI capability. By modelling your enterprise and its ecosystem in Enterprise Studio, you can start using Enterprise Studios analytics and AI capabilities (e.g. natural language generation, predictive view suggestions, etc.). Please get in touch if you’d like to see it in action!