James Taylor, CEO of Decision Management Solutions and a leading expert in helping companies adopt digital decisioning.
The last few months have seen a dramatic increase in excitement around Large Language Models (LLMs) and generative AI. While not new, recent performance improvements and immense increases in the size and complexity of the models mean that they can manage more complex problems. The ability of these models to summarize complex documents, pull together multiple threads and generate readable text—of any length and in any style—is quite remarkable. Every company is justifiably investigating how best to take advantage of them.
Many companies are especially keen to allow customers to interact with an LLM that “represents” the company. Such superpowered chatbots would make decisions on behalf of the company, giving customers answers 24/7. They would relieve over-burdened call centers of calls, allowing human staff to focus on higher value-added activities. In this way, they could reduce costs while improving customer service and satisfaction.
However, this powerful and compelling use case is threatened because of the way LLMs make decisions. Trained to predict what the next word, sentence or paragraph might be, LLMs struggle to provide transparency about how they decide. They may give answers that are not compliant with published regulations or company policies, potentially creating legal exposure for the company. They have the potential to “hallucinate,” supporting incorrect responses with seemingly valid references and explanations, which could lead to significant issues for the company promoting their use.
How, then, can you take advantage of an LLM without these risks? One option is combining it with digital decisioning systems. Based on my years of experience as a consultant for companies building digital decisioning systems, here is what I have learned regarding how these two technologies can complement each other.
Understanding Digital Decisioning
Leading organizations around the world have faced the need to automate business decision-making for years. When transactions or customer interactions require quick response times or the ability to manage very large volumes, companies automate. When a need for automation is combined with decision-making complexity, especially when it’s not clear what the best response should be, typical methods of system development may fall short. That is why I recommend Decision Management, or Digital Decisioning, models when decision-making complexity such as ensuring regulatory compliance, applying multiple company policies, or deep knowledge of best practices is required.
Digital Decisioning systems utilize expert knowledge in software through the use of decision models, business rules and decision tables. These business-oriented representations have the advantage of allowing domain and subject matter experts within a company to directly specify, update and manage decision-making. This typically eliminates the need to create separate IT requirements documents for implementation by a different team. Decisioning platforms and business rules management systems can help ensure this logic can be executed at scale. The combination delivers transparency because the design is typically clear, even to non-technical people, and because the execution is logged.
Advanced analytics complements these expert-based approaches. Predictive analytic and machine-learning approaches extract business insights from historical data. The resulting analytic models can be deployed to replace rules of thumb. They can be combined with expert guidelines and business understanding to improve the accuracy of decisions.
However, these systems are generally inaccessible to customers. They are most often deployed as Application Programming Interfaces (APIs) where enterprise applications and back-office systems can access them. High-performance processing of loan applications, claims handling, fraud identification and much more is possible, but these systems are rarely accessible to customers or the company representatives they talk to.
Combining Digital Decisioning With LLMs
With the arrival of LLMs and generative AI, customers are no longer willing to settle for intermediaries. They want to be able to self-serve on even the most complex of transactions. Yet, as mentioned above, LLMs can’t yet be fully trusted to make repeatable, transparent, compliant decisions. When LLMs are effectively combined with expert-based decisioning, however, it can facilitate AI-driven interactive experience for customers combined with the transparency, repeatability and compliance of digital decisioning.
To see how this might work in practice, consider claims handling. Say an insurer needs to apply federal and state mandates, company policies and individual plan designs to see if a claim is covered by a policy and to calculate how much will be paid. It needs to use historical data to see if there are patterns of fraud or waste that should be flagged. It needs to do this at scale for millions of claims. The decision-making can be designed using a decision model so it matches the way the company operates. This model can then orchestrate the business rules and analytics needed to make the decision. The system can log how this decision was made. It is consistent, transparent, accurate and reliable.
However, it is only accessible by submitting a claim to the claims system. A patient gets an arms-length explanation of benefits, eventually, but only after a claim is submitted. Providers, too, must submit claims and make plans for payment based on their understanding of the contracts and terms they have signed.
Add an LLM, though, and everything changes. An LLM can power a provider or patient chatbot. It can find the relevant documents and summarize them, helping people understand the basis for the decision. It can interact with members and provider staff to identify the patient and gather the information required by the digital decisioning system. Once it has the information it needs or gets it from an internal system, it can call the digital decisioning system to get the official answer: What WOULD the insurer say if this claim was submitted, at this time, by this patient, with this medical history? The result will be the same as it would be if you went ahead and submitted it.
Furthermore, because the LLM manages the conversation, it can take the technical explanation or log provided by the decisioning system and turn it into something the patient or member understands. The member gets an easy-to-use, chatty and human-centric interaction and an accurate, transparent, compliant and reliable decision.
LLMs are a powerful and game-changing technology, but they can complement, rather than replace, expert-based decisioning solutions.
Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?
Read the full article here
