Post by account_disabled on Mar 5, 2024 1:17:32 GMT -5
That must be taken into account when evaluating the usefulness of solutions involving llms. One dimension, of course, is the degree of linguistic fluency of which the system is capable. Conversational user interfaces, exposed by chatbots or voice bots and digital assistants, smart speakers , etc. They have existed for a long time. These systems are capable of interpreting the written or spoken word and responding accordingly. This response can be written/spoken or initiating the requested action. One of the main limitations of these more traditional conversational ai systems is that they are better at understanding than - for lack of a better word - expressing. By relying on well-trained machine learning models, they are also often able to find a correct solution to problems in the area for which they have been trained, as they often work based on pre-trained intentions. Based on the training data, they usually give quite precise answers to questions in their domain. The problem: they are usually limited to a fairly small number of domains. Machine learning programs (llm), for their part, are usually trained to “understand the relationships between words, phrases and sentences in a language. The goal is for the llm to generate results that are semantically meaningful and reflect the context of the input. This is part of chatgpt 's answer to the question: what is the purpose of an llm?. The training set for an llm typically consists of a large amount of “real world” knowledge that typically comes from public sources.
The internet. The output itself may be in written, graphical, or other format. What llms excel at is generating answers to questions in a human way. And they can respond to a wide variety of topics. When focused on text, they are meant to generate coherent and meaningful responses. The problem: sometimes they lack precision and confidently give wrong answers. And what is worse, erroneous or inaccurate results are not easily identifiable by a user without the necessary Buy Bulk SMS Service knowledge. See again the example of google bard, which made $100 billion disappear (temporarily, at least) from google's valuation. And it's not just about google; there are many examples that point to chatgpt or you.Com, among other tools. How accurate are llms like chatgpt? The question is whether both dimensions always matter equally or not. In a business sense, it can be argued that precision always matters. Receiving factual errors in a business conversation is not only an example of poor customer experience, but in extreme cases it can even lead to legal problems. What is also important to understand is that the more precision required, the more the need to integrate additional systems to increase llm increases. An llm on its own is not much more than a form of entertainment. Even in search engines, llms only augment search by enabling natural language queries and delivering results in human language instead of a mere list of links. At least, this is what they should do. That being said, what are the business use cases for an llm? As already said, reasonable precision is needed. Obviously, they also require fluency as a precondition, as fluency is the core differentiator of an llm.
Let's look at some use cases in no particular order of priority. Chatgpt business use case examples tell stories I would start with something I would call “storytelling.” it is basically the creation of market-relevant documents that describe the capabilities and differentiating factors of a product, solution or service. Being somewhat related to marketing (no offense intended) and being a first point of contact for customers, it has to be easy to understand without requiring great technical precision. At the same time, it should not be incorrect. A simplified version of this could be the (improved) generation of content for social networks, for example tweets. The benefits of this are faster creation of high-level content for general websites, but also, more specifically, for abm scenarios and landing pages. In order to create this text, an llm needs to be connected to internal systems that contain requirements, specifications, as well as communications between the people involved. This is also a use case that should be implementable in the short term. Emails one of the main tasks of people is writing and, above all, responding to emails. Especially in sales scenarios, where customer queries can be formulated and suggested based on previous emails and the context provided by the crm system, for example on proposals made.