Artificial intelligence and large language model tools in business  – a sandbox for creators or a key to growth?

Language technologies
News title with a portrait of Kaspars

The potential of generative artificial intelligence (AI) and its tools has long been a hot topic of public conversation, yet their application and the true business potential remain unclear. It is important to understand the possibilities unlocked by large language model tools, in order to make maximum use of their advantages. 

The number of users of generative AI tools has been growing rapidly: since 2020, it has doubled, reaching 250 million in 2023, and experts believe that by 2030 AI will already be used by 700 million people. However, despite the fast development of AI and the increasingly larger number of users, we at TILDE have also noticed that many companies are still unsure about whether they should use large language model (LLM) tools in their business processes. The main reasons for this uncertainty are concerns over the ethics of using such technologies, data privacy and security. Nevertheless, those tools are actively used by company employees, primarily to save time. 

There are numerous AI tools and they need to be explored to gain significant efficiency improvements in organisations. 

LLM tools unlock many opportunities, starting from information retrieval and processing, translation, speech recognition and text-to-speech technologies to image and video production. These types of applications can significantly improve efficiency and shorten the time spent on specific tasks. In 2023, Microsoft conducted a survey involving 31 thousand respondents. 70% of respondents admitted that they are quite eager to use AI tools to reduce at least some of their daily workload. To address the concerns of company management and find out which business processes and resources of the organisation could be optimised in a safe and acceptable manner, companies should conduct in-depth analysis and testing of AI tools.  

Unproductive meetings are the biggest time thief.  

A large part of employees spend a lot of time in various meetings, and inefficient meetings are the number one disruptor of productivity. Microsoft’s survey suggests that 55% of people feel that the next steps at the end of a meeting are often unclear and that it’s hard to summarise what was discussed at the meeting (56%). 

Inefficient communication is a waste of time and staff resources, as the discussion has to be subsequently repeated to understand what it was about and what the next steps should be. One of the approaches to make virtual meetings more meaningful and effective is the use of LLM tools. More and more AI tools are being developed that automatically transcribe meetings, summarise what was said, as well as create to-do lists. This reduces the risk of some important information possibly being missed.  

There is no need for employees to waste time on an information search if it can be done by the company’s internal assistant.  

Another way of using LLMs is information search, knowledge extraction and response generation based on information sources selected by the company: documents, website content and specific information databases. This is particularly handy in companies with a large number of employees who handle a significant amount of information and data on a daily basis; for example, in insurance, finance, medicine, education and other sectors. 

Solutions that work as the company’s internal ChatGPT are already available on the market, including in TILDE’s product portfolio. They combine complex technologies such as Retrieval Augmented Generation or RAG, semantic search and LLM, in order for the system to be able to process a variety of internal information of the company: procedures, documents and other resources. Such solutions can find the required information and answer the employee’s queries in a matter of seconds by extracting information and knowledge as well as indicating the sources and formulating recommendations as to where to look for additional information. 

Man sitting by computer

Chatbots improve the accessibility and efficiency of customer support teams significantly, while at the same time also creating a positive customer experience.  

Chatbots or virtual assistants are no longer a novelty: they provide significant improvements in both the efficient use of staff resources and customer experience. Public demands as to the speed of customer support are constantly growing, and people want to receive support immediately and at any time of the day. However, not all companies are able to provide customer support round-the-clock; therefore, chatbots are indispensable partners to the existing team. Organisations using TILDE’s virtual assistants report that at least 30% of their customers use chatbots to obtain information in the evenings or during weekends and holidays when customer service centres are closed.  

Chatbots are often surrounded by various misconceptions and concerns; for example, a lack of confidence that they will be able to understand the customer’s problem and provide meaningful support. Most of the time, such concerns are unjustified because smart chatbots are able to understand a customer’s query even if it contains grammatical errors, as well as provide personalised support to authenticated customers. Moreover, the latest LLM technologies enable the even faster training of virtual assistants, significantly expanding their knowledge base and possibilities to provide support. Unlike previous-generation chatbots, the latest virtual assistants use LLMs that are able to generate responses automatically based on vast amounts of unstructured data; for example, web page information, instructions, product and service descriptions and other documents. This innovative solution enables more effective training of the chatbot, a broader knowledge base, as well as diverse functionality and integration options, thereby considerably improving the quality of customer service and user experience. 

The opportunities provided by AI need to be explored and put to use. 

AI tools do not replace but rather complement the team. These tools are definitely worth exploring in order to minimise the time spent on tasks that can be streamlined. Saving time on repetitive and time-consuming tasks also increases productivity, as employees are able to dedicate more time to more complex and creative assignments. The automation of less important duties by means of AI tools can help companies reduce their costs, improve various business processes and be increasingly more successful.