We will start to hear more and more about 'Decentralized autonomous enterprise' in the coming 5-10 years as shown in the Gather emerging technologies Hype cycle below. I see that most of these emerging technologies are around AI and Graph in the form of Augmented Intelligence, Knowledge graphs, AI PAAS, Transfer Learning, Adaptive ML, Emotion AI, and Explainable AI that will can be used to enable the 'Decentralized Autonomous Organization'.
So, what is this 'Decentralized Autonomous Organization'?
Currently, most big enterprises may have already automized and optimized their customer facing and supply chain business processes during their digital transformation efforts using the latest technologies. Several AI-driven solutions are deployed to help manage the customer journey and marketing campaigns in the form of personalized offers and campaigns. AI is also already used in several operational processes such as Fraud Detection, Document processing, etc. This is where we are today...
However, even if the insights delivered so far help in the decision making, they do not learn from their actions and optimize the decisions made by themselves. For example, a fraud detection system might flag a transaction as a fraud but it might turn out to be a valid transaction or vice versa. In this case, the only way to fix it would be to re-train the model based on newly available data and labels. Human assistance is required to help fix this issue. Adaptive ML and other approaches might be required to address the capability to learn from its mistakes. Thus, it seems to be the next level of maturity in the journey to a 'Decentralized Autonomous Organization'. That is not the only roadblock to the 'Decentralized Autonomous Organization' though, there are also ethical issues as rightly set by the CEO of Microsoft and other industry leaders in the field as defined below...
The ethical rules of AI adoption:
- “A.I. must be designed to assist humanity” meaning human autonomy needs to be respected.
- “A.I. must be transparent” meaning that humans should know and be able to understand how they work.
- “A.I. must maximize efficiencies without destroying the dignity of people”.
- “A.I. must be designed for intelligent privacy” meaning that it earns trust through guarding their information.
- “A.I. must have algorithmic accountability so that humans can undo unintended harm”.
- “A.I. must guard against bias” so that they must not discriminate against people.
If we limit AI/ML technologies based on the above AI adoption ethical guidelines, we should never go beyond the predictive analytics where the machine only makes recommendations... Also, the reality is that when it comes to decision making, people will not trust machines and would like a responsible body for mishaps. Overall, it will take a long time for the current generation to leave the decision making to just machines...
I believe that AI based decision making in most cases is the subject of the next decade, but experiments needs to be there to come to this level of maturity. Ethical restrictions will change over time and AI will eventually cost the jobs of millions (think of self-driving cars) with many economical and social implications. However, we humans are not always rational beings and will go with the hype eventually. Thus, AI taking over decision making in a decade seems the most probable. The organizations which are taking the lead will have the competitive edge where machine based decision making could turn out to be more efficient than humans. If we are not able to prevent its coming, I believe, we need to conquer it!
Currently, I can see two approaches to enable a "Decentralized Autonomous Organization" in the long term. The first one is Intelligent Business Process Automation where all business processes can be automized and self-optimized. They can be configured to manage certain Operational and Strategic KPI's to deliver the organization strategy. However, I see several complications with this approach as the learning processes will need to be guided over long periods of time and the dependency on running all business processes on a BPM engine with rule engines is quite challenging...
The second option is using the Augmented Intelligence. Augmented Intelligence targets to support humans to do their jobs better which is in line with the current ethical rules of AI. As an example, an employee may ask questions in his/her language to a system to analyze terabytes of data and a system would come up with the suggestions after running complex simulations and prescriptive models. No one would object to using such a technology that is quite helpful. One such technology that delivers the promise of Augmented Intelligence is ThoughtSpot, but it is a generic tool and suitable for all use-cases.
After allowing the Augmented Intelligence solution to work with the humans side-by-side, the next step could be possibly asking the system to learn from the recommendations made by the employees and actions taken by the employees. Imagine you are using a chat bot based solution to ask questions and get answers. The system may also learn from your decisions and then be able to autonomously make decisions. As an example, read the following conversation between the system and the employee to realise where this can go...
- You: Is sales forecast inline with the targets for region X and business unit Y
- System: Yes, with additional expected revenue of $100 m for this year.
- You: How can I reduce my working capital?
- System: Transferring inventory from region X to Y will reduce working capital by $5 m
- You: Any predictions for my business unit?
- System (Prediction): Product X may run out in 1 month at London store.
- You: Any excess inventory we need to move to London store?
- System (Recommendation): I have found excess inventory from region X warehouse to fulfill %60 of the stock-out. Do you want me to initiate the stock transfer?
- You (Action): Yes, let’s do it.
- System (Learns from actions): In autonomous mode, it can perform actions on behalf of the role based on daily operations tasks.
Using the approach above, the daily operations of the business can be made autonomously in the absence of the employee as the system can ask the questions to itself and take action based on earlier successful transactions. Such a system can first be deployed as a 'Business assistant' that learns from the user interactions. When the current ethical rules are not a concern anymore, it can then be deployed as 'Digital Twin' to make the decisions itself based on past learnings. Likewise, instead of putting a lot of effort into business process automations, Augmented Intelligence approach can do the job without any additional effort by building its knowledge/expertise over time.
There are several challenges on delivering the hype of 'Decentralized Autonomous Organization' such as technology limitations, complexity, ethical issues, and strategic approaches. One of the options for this futuristic technology might be to capitalize on Augmented Intelligence. It requires less effort to other alternative approaches such as automation of business processes. In 5-10 years time most organizational decisions are expected to be made through machines as this concept and hype will be fuelled by the leaders in industry...
There are several challenges on delivering the hype of 'Decentralized Autonomous Organization' such as technology limitations, complexity, ethical issues, and strategic approaches. One of the options for this futuristic technology might be to capitalize on Augmented Intelligence. It requires less effort to other alternative approaches such as automation of business processes. Unfortunately, In 5-10 years time most organizational decisions are expected to be made through machines as this concept and hype will be fueled by the leaders in the industry...