Natural language processing chatbots bring conversation to AI
A natural language processing chatbot that focuses on intent can boost the effectiveness of bot technology. By evolving conversationally, bots can become digital coworkers.
Companies using natural language processing chatbots should look to multitask if they want to evolve processes like IT service management capabilities with machine learning. When developers consider design, personality and interaction, bots can join the workforce as employees, not just technologies.
AI-enabled conversational agents that are user-designed and understand flexible human languages and questions generally outperform stagnant chatbots when it comes to long-term user adoption of AI technology.
Who is CINDE?
Los Altos-based IT operations management company Symphony SummitAI added a new chatbot in the latest version of its SummitAI IT service management (ITSM) suite. CINDE, the suite's digital agent, can converse across different platforms to communicate with users wherever they are.
Akhil Sahai, chief product officer at Symphony SummitAI, said the tool seeks to use AI and machine learning to make companies' service desks functioning members of the workplace -- not simply to automate or augment an individual process.
Using natural language processing and by focusing on integrating tools with employees, AI bots can understand user intent better -- something Sahai said most chatbots are missing.
Pranay AgrawalCEO, Fractal Analytics
AI digital assistants leap ahead of traditional chatbots and single-use machine learning programs for one key reason: They can understand intent. Natural language processing chatbots seek to understand the intent behind a help desk ticket instead of just solving the problem, which boosts their overall effectiveness. They can solve a single problem -- reset passwords, set up virtual private network access -- but can also answer follow-up questions, learn from requests and dig through previously closed tickets.
What sets natural language processing chatbots apart?
As chatbots become a staple in AI-enabled enterprises, some versions are proving to be limited in their functionality and ease of use. Most chatbots require specific question formatting and deliver bland, formulaic answers to questions -- they can't hold a conversation.
Natural language processing chatbots can develop something of a casual personality -- responding to multifaceted questions in a conversational manner and seeking to understand what users are looking for, not just responding to keywords. You can phrase your question multiple ways and still receive an applicable answer.
"You want to have a conversation with an employee and not give them a straightjacketed Q&A," Sahai said. "Then, [the agent] gets to learn from [the chatbots]."
Natural language processing chatbots with friendly, flexible interfaces and that can take over low-level processing jobs can override fears of AI in the workplace -- thus integrating the technology more seamlessly, said Pranay Agrawal, CEO of Fractal Analytics.
"In the work setting, one of the biggest challenges in implementing AI is skepticism," he said. "A hurdle [to implementing AI] is getting too caught up in the technical fanciness of technology without giving adequate attention to the users and how they're going to use it."
The future of chat
Traditional chatbots automate processes for entry-level workers, but natural language processing chatbots can ideally evolve and start tackling processes with higher skill levels.
Analysts say that the rise of AI in the workplace will lead to a cohabitated work environment where bots can take on short term and low-level work, leaving humans to take on longer term projects or skilled assignments. Whatever the trajectory, it's clear that technology has to be designed with existing employees and their processes in mind.
"Just AI is not enough. You need to bring in a combination of AI, engineering and design," Agrawal said.
Agrawal notes that the technology itself is not beneficial in the long term without focusing on conceptual user intent, the user interface and the overall likability of a product.
"AI is the ability," Agrawal said. "Engineering is the ability to integrate large amounts of data into the algorithm and then output from these algorithms into operational systems, where decisions are made or actions done. Design is the concept of 'how do we keep the end user in mind' so that we can create applications that users can actually adopt."