Artificial Intelligence Process Automation for Enterprise Business Communication

Ciarán Daly

September 4, 2018

5 Min Read

By Dynamic AI56

According to Forbes, 85% of business communication is repetitive, time consuming, and costly  - but imperative for enterprise success. In the near future, artificial intelligence systems will automate simple tasks, while many areas are already covered today - just think of your bank SMS assistant. Companies are deploying automation solutions to facilitate internal and external business communication across diverse channels. Technological leadership is crucial for companies in the age of automation.

The clearest benefit of automated communication is cutting response times for customers and employees down to seconds. Companies that manage customer expectations have not only made AI part of their technology strategy, but invested in their own workers who can benefit from more free time. When most simple tasks are handled by process automation, human operators have more time to deal with complex cases and create a unique experience for customers.

Artificial intelligence is a crucial technology for enabling automatic learning and the flexible interpretation of the automated business communications. Unlike chatbots, FAQ-bots, and automated responders, which are usually trained for a predefined set of cases, an AI system can learn new cases of process automation from real-time operations. Typical applications of process automation include the labelling & categorizing of requests, customer routing, and responding to simple queries. Despite diversity of process workflows across companies, AI-based systems can also learn specific actions.

More complex cases, like legal inquiries or long emails, can only be served by AI systems that use features of computational linguistics to get a sense of the message not to be based on a simple list of keywords or phrases. The ability of AI-based systems to learn enables the discovery of similar requests in operations history and proposing the relevant action. Another advantage is the ability to process complex workflows involving multiple messages, e.g. a long e-mail thread with pieces of information spread across many messages.

One of the problems of automated business communication, however, is unpredictable accuracy rates. Incorrect responses to customer requests can have a negative impact on the customer experience. This is why many companies never deploy an automated system. Furthermore, the ability to control correct action rate for an automated system is crucial to keep overall interaction impressions high. Companies continuously seek a balance between the predicted rate of correct actions and system efficiency that a volume of traffic system can handle automatically. 

Dynamic AI Inc. has developed a system for AI process automation for enterprise business communication that automates communication and databases for enterprise customers and employees to save time and cost, and boost customer satisfaction. The system is able to predict the probability of a correct response and routes ambiguous requests to human operators. Their feedback, as well as that of automated responses, is used to train the system to increase accuracy and cover more types of request.

We have developed a patented novel Natural Language Understanding (NLU) architecture built upon an in-depth analysis of text structure using recent breakthroughs in computer linguistics and a set of self-organized machine learning approaches. Modular architecture enables integration with various communication channels e.g. Slack, Skype, and e-mail, and can be integrated into company’s CRM or HR database. Distributed processing pipelines allow the system to scale in terms of traffic volume.

Implementing such a system is impossible without creating in-house AI technologies to get the best out of both Natural Language Processing (NLP) and machine learning. While the most well-known AI technologies are machine learning and neural networks, other technologies and approaches are emerging. Our patented technology is based on rethinking a well-known concept of genetic algorithms with unique applications. We use genetic algorithms to conduct a search of other algorithms, which in turn solve distinct problems in fields of computational linguistics and machine learning. Such an application can hence be thought of as a form of artificial intelligence which produces algorithms to analyze accumulated knowledge in various ways to make decisions and hence achieve automation.

Automation of customer care, communications for internal IT support and HR departments are key applications for our system. Imagine a new employee writing an e-mail to IT support requesting a performance workstation with proper reasoning. The system consults its knowledge base to find if he/she is eligible for such a request and locate the appropriate approval chain for it. The necessary information e.g. hardware type, department, due date, etc. gets extracted from the message. Then, a formalized ticket is created in a service management database with proper labelling. Managers responsible for approval are notified by Slack to vote on the ticket. Then, the system responds to employee’s emails, stating the successful outcome of automated processing and resulting ticket number.

The HR department is usually responsible for onboarding procedures and answering frequent questions about the company culture. There is a Slack chat for this purpose. When a new question arrives, the system searches for similar the request in the historical knowledge base. If the answer template is found, then it is filled with actual information, e.g. name of responsible employee or working hours, and then sent back to chat. Employee asking a question can vote whether he/she is satisfied with such answer and this information is used to train the prediction of correct response probability. In case, the worker is dissatisfied or if system could not provide an answer, a HR employee is invited to a chat. The response is captured and is used to extend the knowledge base of the system. So next time the similar question can be answered automatically.

The impact of AI technologies on business communication is constantly rising. Most simple tasks will be automated in the coming years and new technologies for dealing with complex interactions will become live. Companies with technological leadership on the automated business communication landscape will boost customer and employee satisfaction. From customer self-service to complex business processes, every form of business communication will definitely be deeply impacted by trend of automation and for enterprise executives it is important to know what’s on the horizon for AI process automation.

To learn more, catch up with Dynamic AI at this month's AI Summit San Francisco

Keep up with the ever-evolving AI landscape
Unlock exclusive AI content by subscribing to our newsletter!!

You May Also Like