One of the trends that I find fascinating in the field of artificial intelligence (AI) is the possible emergence of what I call “XAI-embedded service level agreements (SLAs).” The X, in this case, stands for explainable (more on that later). These are contracts that quantify the performance of a service by using qualitative and subjective indicators such as customer engagement, client satisfaction, brand loyalty, or ease of use.
In this blog post, I will explore some of the recent digital innovations that are paving the way for XAI-embedded SLAs and how these can benefit service providers and customers.
The First Digital Innovation: Algorithmic Contract Types
If you want to create and sell bonds, you need to specify how and when you will pay back the money to the buyers. If these financial assets are converted into digital tokens that represent their value and ownership, you can use smart contracts to code these rules into the tokens. Smart contracts are self-executing agreements that run on a blockchain. They can automatically check if the conditions of the contract are met and trigger the payments accordingly. This way, you can make sure that the bondholders receive their money on time and that you follow the regulations of the market.
By tokenizing digital assets, financial operators can create more efficient and transparent services for their clients. Here are some of the benefits of tokenizing digital assets:
- Real-time balance sheet: Tokenized digital assets can be tracked and verified on a distributed ledger that records transactions and ensures their security and immutability. With these, financial operators can provide their clients with accurate, up-to-date views of their assets and liabilities without relying on intermediaries or manual processes.
- Deterministic cash flows and price discovery: Tokenized digital assets can be programmed with smart contracts, enabling financial operators to automate the execution of payments, dividends, interest, and fees based on predefined triggers and events. Moreover, tokenized digital assets can be traded on decentralized exchanges, where buyers and sellers can discover the fair market price of the assets without intermediaries or manipulation.
- Transparent liquidity terms with financial contracts: Tokenized digital assets can be used to create more flexible and customized financial contracts, such as loans, bonds, and derivatives. By tokenizing the underlying assets and the contractual terms, financial operators can offer their clients more transparent and tailored liquidity options, such as early repayment, collateralization, and risk management.
The Algorithmic Contract Type Unified Standard (ACTUS) Financial Research Foundation1 developed and curated the ACTUS Data Standard, which defines a universal set of legal terms used as parameters throughout the different financial agreements. It is implemented in the form of a data dictionary with attribute applicability by contract type. A second standard, the ACTUS Algorithmic Standard, defines the logic embedded in legal agreements that eventually turn the contract terms into actual cash flows or, more generally, business events.
With the core data of digital assets tokenized and embedded within the ACTUS standards, financial operators can offer clients real-time balance sheets through deterministic cash flows and price discovery. This enables machine-auditable and automated execution of transactions and more transparent liquidity terms with financial contracts.
Nucleus Finance, a fintech company, is developing Tokenization-as-a-Service solutions that translate financial instrument data into ACTUS smart financial contracts on the Casper blockchain2. As a native financial digital asset, the token is machine-readable and executable for faster trading and settlement.
The Second and Third Digital Innovations: Fuzzy Logic and XAI
Fuzzy logic and XAI are two distinct concepts that can be used together to create more transparent and interpretable AI systems.
Fuzzy logic is a mathematical framework that allows for the representation of uncertainty and imprecision in data. It is based on the idea that, in many cases, the concept of ‘true’ or ‘false’ is too restrictive—that there are many shades of gray in between. The fundamental concept of fuzzy logic is the membership function that defines the degree of membership of an input value to a certain set or category. The characteristics of an entity subject to personal evaluation (e.g., easy to use, improves quality of work) can be assigned quantifiable metrics based on the collective (i.e., membership) assessment constantly reviewed and calculated by the underlying algorithms.
XAI, on the other hand, refers to the development of AI systems that can provide clear and understandable explanations for their decisions and actions. It contrasts with the concept of the “black box” in machine learning, where even their designers cannot explain why the AI arrived at a specific decision. Fuzzy sets can offer an effective paradigm for supporting an accurate understanding of natural language and building efficient linkages to human intelligence through concepts and computing with membership functions for XAI3.
Temenos, a banking software company, launched a Software-as-a-Service XAI model based on ‘type 2’ fuzzy logic that leverages static and transactional data patterns, coupled with human judgment and expertise, to provide easy-to-understand explanations in human language to users, helping them identify why exceptions occurred and relay that information back to regulators and customers4.
So, What Are XAI-Embedded SLAs?
One of the challenges of digital innovation is to establish SLAs that reflect the value of the solutions provided. Traditional IT services can be evaluated based on quantitative indicators such as availability, performance, and reliability. However, digital solutions often affect the experience of users in more subtle and subjective ways. Therefore, SLAs for digital innovation should include qualitative indicators that capture the aspects of user experience (UX), such as customer engagement, client satisfaction, brand loyalty, or ease of use.
UX is a broad term that encompasses the emotions, perceptions, and behaviors of users when they interact with a product or service. UX is often measured with subjective or qualitative indicators that are hard to quantify, but they are essential for creating value and differentiation in the competitive digital market. So, how can providers incorporate the UX value of one provider into their SLAs in a way that can be measured according to a standard reference and comparable with other providers’ UX?
One possible way I envision to implement this approach is to use XAI-based algorithmic contract types that ingest “fuzzy” indicators and apply them as digital tokens to a financial transaction. While “traditional” indicators for UX are measured through surveys, feedback, reviews, ratings, or other methods that collect user opinions and preferences “offline” of the transaction, the data to measure XAI-embedded SLAs is collected “online” by the token throughout the transaction life cycle. The token feeds the collected data into the AI engine that calculates and attributes the relative quantitative values, compares them with the SLA metrics, and produces the results that the “explainable” component of the engine allows the human to interpret. A score for the “UX value” can be determined and compared with the SLA threshold. All this while the transaction is running live.
Why Are XAI-Embedded SLAs Important?
XAI-embedded SLAs are important for several reasons. First, they can help establish trust and transparency between service providers and customers, especially in domains where transactional systems have high stakes or social impacts, such as health care, education, or in our case, finance. By setting clear expectations and accountability mechanisms, XAI-embedded SLAs can reduce uncertainty and risk for both parties and foster confidence in the quality and value of the service.
XAI-embedded SLAs can also help drive innovation and improvement in the field of AI. By creating incentives and feedback loops for service providers to optimize their transactional systems and processes, XAI-embedded SLAs can encourage continuous learning and adaptation, as well as ethical and responsible design and deployment of XAI solutions.
Finally, XAI-embedded SLAs can help create new opportunities and markets for financial services. By enabling more standardized and comparable evaluation and benchmarking of transactional systems and services, XAI-embedded SLAs can facilitate interoperability and integration across different platforms and domains. Moreover, by allowing customers to choose and customize their desired level of service quality and performance, XAI-embedded SLAs can create more diverse and personalized offerings and experiences for different needs and preferences.
Algorithmic contract types, fuzzy logic, and XAI are powerful tools that can help providers incorporate qualitative and subjective indicators, such as UX value, into their SLAs and deliver solutions that impact the experience of their users.
Fintech providers should start considering XAI-embedded SLAs as innovative propositions to quantify and improve their performance in a transparent and accountable way. UX is just an example. Many other qualitative and subjective metrics can be processed in the same way. The algorithms behind an XAI-embedded SLA can also use the feedback loop of the data captured by the token throughout the life cycle of a transaction to adjust the scores and help users optimize the performance of the offered products and services according to the needs and preferences of their customers.
Banks should also consider these tools to create value and differentiation in the digital market, providing a new way to quantify (and consequently monetize) increases in customer satisfaction and loyalty.
For more information on this topic, contact me here.