Artificial Intelligence (AI) as fuel for business value has been saved
Artificial Intelligence (AI) as fuel for business value
The convergence of AI into a value generating engine for business is accelerating and has become a serious option to consider.
While this may sound great, it cannot be realized by simply tapping into an ocean of data and “magically” turning this into money. However, there are clear steps that can be taken to improve the chances of success and value of AI initiatives. That is why this series of blogs elaborates upon Deloitte’s experience in the field of AI and brings you a pragmatic and multidisciplinary approach to turn data into value through the so-called Deloitte Artificial Intelligence Loop (DAIL).
By Naser Bakhshi, Sjors Broersen & Daniël Rennings
Go directly to
- Deloitte Artificial Intelligence Loop (DAIL)
- So what can you do with DAIL?
- What to expect next?
It’s hard to imagine a world without data. It even sounds surrealistic to have such thoughts nowadays. Almost everything we do in our daily life generates huge amounts of information. Never before did companies have access to the vast amounts of data they store today, ranging from data on customers and financials to operations and ecosystems. Exciting as this may sound, swamps of data are useless unless they’re put into action. This is where another “hard to imagine” concept comes into play: Artificial Intelligence. A buzzword, a hype or reality? Well, it depends on how you look at it. We have written a whitepaper about the fundamentals of AI, so for the sake of simplicity we assume you have been through it for a reference on practical applications of AI (which we also call applied AI).
In this set of blogs we will zoom into the concepts behind AI and elaborate upon the process from data acquisition to value creation and back.
After reading this series of articles, you will have a better understanding of how applied AI works, and what the key components of sense (data collection), reason (data processing), augment (data utilization) and improve (learning) encompass. Let us describe the so called Deloitte AI Loop (DAIL) framework that comprises of these steps, based on our experience from many projects in the field of AI.
Deloitte Artificial Intelligence Loop (DAIL)
Human cognition has been a fundamental inspiration to both the early works on AI and the solutions we use today, as both typically mimic the human process of cognition. This process of cognition has also been the blueprint for DAIL, which encompasses sensing, reasoning, augmenting and improving. In these four steps it goes from data and algorithms up to value creation and continuous improvement. Let us illustrate these terms this with an example.
Imagine, your department is tasked with developing new products to put on the market. There is a tremendous amount of internal and external resources at your expense, ranging from documents with concepts, prototypes and products that are already on the market, to regular web search engines and licensed books and surveys. Typically, gathering and interpreting data from these sources takes up 30-40% of the time in developing a new product. As a result, using some of these materials, you can scramble together a proposal for a new product every few weeks.. Although it is great that we, humans can combine various sources of information to come to a concrete proposal, we must realize that the limitations in this example are not hard constraints in today’s digital era. What if AI could be used to improve this process of gathering and interpreting data?
This is where DAIL comes in, as we will describe next. Instead of scanning a small subset of the available materials ourselves, what if we would set up an infrastructure so that we can exploit (SENSE) all available material at once? Such an infrastructure would then allow us to utilize algorithms to REASON about the available resources, deciding on what is relevant and what not. As a result, we can AUGMENT the process of defining a new product, as we only go over materials that are already deemed relevant by our cognitive solution, accelerating both our efficiency and effectivity. And let’s not stop there: if we can directly utilize user input to IMPROVE the system over time, we can even increase our efficiency beyond the initial gain from the cognitive system without human intervention. From our experience in implementing cognitive solutions we have seen that in order to be successful in implementing AI, these four steps need to be taken into account. As we also did for the particular example we laid out above, to develop a cognitive solution for one of our clients.
AIDA Deloitte AI Assistant
The Deloitte AI loop is enabled by three critical enablers: Privacy, Security and Ethics, Change Management and Technology Stack. If you are curious to hear more about those, or want to dive deeper into the SENSE, REASON, AUGMENT and IMPROVE, please read the deep dive we provide at the top of this page.
So what can you do with DAIL?
We see DAIL both as a conversation starter and a reference framework: it gives you an impression of what value AI solutions can add to the organization and provides you with a blueprint of factors to account for in rolling out an AI project. This enables the development of cognitive capabilities that become part of the organization that matures in the cognitive space, solutions that are there to stay. Not because they are force-fed, but because they bring true value through careful design and supporting enablers.
What to expect next?
In a series of blogs will have deep dive in the different components of DAIL and describe these in a more in-depth fashion. For more information, please do not hesitate to contact Naser or Sjors via the contact details below!