Blockchain Intelligence (AI) in Value Chains

There is a lot of speculation about the impact of Artificial Intelligence on Blockchain. Part of the problem is the folklore surrounding Artificial Intelligence and Blockchain. Here we elaborate on someof the core synergies between Blockchain and AI – focusing primarily on the advantages of AI for all the layers of a robust Blockchain reference architecture. The business value of the AI is realized in the context of end-to-end value chains.

End-to-End value chains orchestrating people, enterprise applications, trading partners – while leveraging Blockchains and IoT/IIoT – is where the Internet of Value achieves its full potential. Value chains are business processes that are digitized and automated through a Digital Process Automation platform. AI is the nervous system of the value chains.

AI

AI has been around since the mid-1950s! The well-known and respected Association of Artificial Intelligence (AAAI) was established in 1979. But AI had lots of ups and downs with the associated hype. There was a time when rule-based Expert Systems were predicted to take over the world. They didn’t. However, more recently, we are seeing an incredible rekindling of interest and deployment of AI. A perfect storm of technological advances in computing, databases, and networking have caused the emergence of AI as one of the most significant digital transformation trends. However, the hype is still there! It is amazing that even after so many years, there is still some confusion as to exactly what should be considered “AI.”

Artificial Intelligence, at its core, is trying to emulate humans: make intelligent decisions, sense, see, move, carry out tasks, talk, and interact like humans. One of the major areas impacted by AI is work. AI can be the core engine for automation – either through physical robots or software robots, also known as Robotic Process Automation; or through intelligent virtual assistants for AI-assisted work; or through involving cognitive workers who think for a living.  There is no question that many repetitive and well-defined tasks will eventually be replaced through AI-enabled automation: these span smart manufacturing shop floor functions to front or back-office repetitive tasks. AI also attempts to learn continuously. The ultimate objective is to make it difficult in specific highly technical or specialized domains (work) to distinguish between AI and humans. Furthermore, robust AI systems can intelligently correlate and respond to real-time events. The combination of analytical – especially predictive – models with real-time event processing is a powerful enabler for responsive systems. In many applications, AI is providing augmentation and assistance to humans.

Of course, AI can be applied in all areas within consumer and enterprise applications – not just work automation. However, as Blockchain evolves into value chain the emphasis on intelligent automation becomes imperative. This spectrum of automation enabled and empowered through AI, is critical to Blockchain solutions – discussed next.

Blockchain

Blockchain also has its own share of speculation, connotations, and confusion. Blockchain is a type of distributed and decentralized ledger. Each Blockchain involves many copies of a decentralized ledger database that exist on nodes all around the globe. Compared to AI, Blockchain is in its infancy. Blockchain traces its inception to Cryptocurrencies – and especially Bitcoin. It was launched only about a decade ago, and the first-ever block of the Bitcoin Blockchain was unleashed by Satoshi Nakamoto on January 3, 2009. We still do not know exactly who is (or who are) the famous inventor(s) of Bitcoin – which is amazing in the digital era of leaks and disclosures! Whoever this person(s) is (are), they are good communicators.  

There are many salient characteristics of Blockchains, and part of the confusion is that “Blockchain” connotes many disciplines and technologies. Artificial Intelligence is relevant to each one of these but impacts them differently.  In addition to the peer-to-peer decentralized and disintermediated recording of transactions, some Blockchains also execute Smart Contracts. Smart contracts are the policies, rules, and mutual agreements between parties that get executed in the Blockchain.

Sources of “Intelligence”

Before delving into this important synergy between Blockchain and AI, we need to frame the sources of intelligence. There are four main sources, and each of these has a tremendous impact on Blockchain applications.

• Regulations, Knowledge, Policies, & Procedures:  There is an enormous amount of knowledge in written documents. In any application domain or industry, there are policy and procedure manuals with inherent and embedded knowledge. This spans operations manuals, organizational procedures, as well as regulatory compliance documents. Rules-based systems, as well as language processing, can be leveraged to extract the knowledge and then operationalize it in Blockchain applications. Blockchain Smart Contracts can be leveraged to digitize the business rules in the Policies & Procedures (e.g., by-laws or operational regulations) and execute them on the Blockchain. For more complex policies, declarative business rules can digitize the policies and work in conjunction with Smart Contracts.

• Human Intelligence: Cognitive workers – people who use data or information to do their jobs – have a lot of deep knowledge in their heads that needs to be harvested. As noted above, AI-assisted human work often digitizes the knowledge of cognitive workers. Their expertise also needs to be understood, digitized, and operationalized, sometimes in intelligent virtual assistants or bots.  Blockchain Decentralized Applications can leverage the cognitive workers’ knowledge either through the digitization of business rules or through fully or semi-automated intelligent assistants. Cognitive workers participate in end-to-end digitized value streams – operationalized through Digital Processes Automation (DPA).

• Legacy Code: Another important source of business intelligence is embedded in legacy code that contains business logic. The embedded policies can become ossified, with little or no business visibility. They are difficult to change or extend. The challenge is to leverage the intelligence in legacy systems while allowing the organization to modernize and be agile. Blockchain Architectures involve orchestration of Smart Contracts with API invocations of AI decisions through secure exchanges. Here again, the DPA layer can be an intelligent modernization bridge between legacy applications and Blockchain messaging – much the same way it is now with business-to-business protocols.

• Data: Blockchain stores all the business transactions: the addresses of the senders/receivers of exchanges, if applicable, the amounts of exchanges, the meta-data of transactions, and the execution of code in the Blockchain, etc. The chain in Blockchain only grows and is never deleted. Information cannot be modified. Only validated blocks can be added to the Blockchain. So, each Blockchain grows. Blockchain Data Mining and Machine Learning techniques with Blockchain transaction data sets are a tremendous source of intelligence. There are patterns of knowledge and intelligence hidden in the Blockchain data. AI can learn from the Blockchain data and potentially predict outcomes (positive or adverse) before they happen. Decisions based on AI predictions can be fed back to the data source to improve the precision of the predictions continuously.

Blockchain is an important and robust source of ever-growing data to mine or gain intelligence from a plethora of AI techniques. However, there are other sources of data. Some of these sources include social networks, connected devices (IoT/IIoT), business application transaction data – and more. Combining and aggregating these data sources with Blockchain transactions constitute the “raw” Big Data for analysis. No easy task but a tremendous source of continuous optimization of Blockchain solutions. Data mining, as well as machine learning, can then generate and optimize the intelligent decision models.

There is a caveat, though. Even after the combination and aggregation of the data involving Blockchains, the value of the discovered or analyzed models can be achieved only through execution. The AI decision models need to be made the “nervous system” of end-to-end value streams that are digitized and operationalized through DPA.

From all of the aforementioned sources of intelligence, the discovery of intelligence needs to be followed with action in Value Chains

This action is the execution of decisions in the context of end-to-end value streams orchestrating people, enterprise applications, devices, robots, and of course, the Blockchain! Each is carrying specific tasks assigned to them through the underlying AI-enabled DPA. 

Example: Warranty Value Chain

We wrap up with a pragmatic example: the application of Blockchain AI in Warranty Value Chains. You can read the details of Warranty Value Chains here.  This innovative solution was created through the strategic partnership between Pegasystems Inc., Samsung SDS, and Tech Mahindra.

Warranty management has been one of the most significant expenses for manufacturing companies. Despite a reduction in the number of claims issued, there has been a noted increase in costs across the warranty value chain. Existing legacy systems are disparate or standalone, requiring manual interventions and hand-offs, with zero immutability of records.  The onus is transferred to the customer to prove ownership and warranty coverage to OEMs via physical invoices, contributing to decreased customer satisfaction, duplication of work, and inability to handle counterfeiting and fraudulent claims. The dual combination of DPA + blockchain aims to target these inefficiencies by optimizing the warranty value chain. Organizations gain the power to track the part with a digital twin for the entire lifecycle of the warranty value chain, as illustrated here.

How about AI? There are multiple opportunities and use cases for all categories of AI in Warranty Value chains:

• Warranty Business Rules and Policies: these are decision tables, decision trees, constraints, calculations, and expressions, etc. that are authored by Warranty experts. Business rules can be leveraged in all the milestones of the Warranty Value Chain.
• Predictive Analytics for Repairs: The connected device IoT Data can be mined for predictive analytic models that can be digitized and automated in especially the repair milestone of the Warranty Value Chain.
• Machine Learning and Adaptive Analytics: In addition, the combination of business rules, predictive models, and continuous learning feedback loops can be leveraged to prevent repairs ahead of time and avoid costly fixes for devices or vehicles under Warranty.
• Customer Experience Next-Best-Actions: The customer context and customer interactions are another source of data that could be analyzed for both predictive and machine learning prioritization for actions that optimize the customer experience.
• Blockchain Ledger Analytics: The transactions that are recorded on the Blockchain for the entire value chain are another rich source of potential predictive and machine learning models that could be periodically analyzed to optimize the entire value chain.

AI is an integral part of Digital Process Automation, and the potential of AI optimizations for on-chain (Blockchain transactions) and off-chain data (IoT, customer, etc.) are tremendous. AI is the nervous system that automates and drives end-to-end digitized value chains to successful completion.