pio3 - Fotolia
Natural language processing is an emerging technology that has the potential to revolutionize the way businesses engage with customers.
NLP technology uses text and speech analytics to identify patterns in speech, a capability that can help businesses gain insight into how they can improve their products. IT services firms now have an opportunity to build NLP-based customer- facing platforms for a variety of industries.
The NLP market, which research firm MarketsandMarkets estimates will grow from $7.6 billion in 2016 to $16 billion by 2021, is fast becoming one of the biggest artificial intelligence (AI) market segments. NLP applications are popping up in vertical markets where customer interaction is essential for business growth. These verticals include the banking, financial services and insurance industries, retail, media and entertainment, and healthcare, among others.
"NLP promises a lot of transformational benefits for businesses," said Kalyan Kumar, CTO at HCL Technologies Ltd. "The ecosystem of providers is constantly expanding in this space and we are seeing demand for this technology across verticals and regions."
HCL Technologies, which is a multinational IT services company, headquartered in Noida, Uttar Pradesh, India, has developed DRYiCE, HCL's autonomics and orchestration platform. The platform, which has more than 40 modules, incorporates AI technologies such as machine learning, NLP, deep learning, cognitive computing, predictive analytics, robotics and neural networks.
Assessing the demand for NLP technology, Kumar said clients across verticals are looking for intelligent and comprehensive solutions for automation and orchestration of tasks. Changing customer requirements, behaviors and client experiences are the primary drivers for enabling such solutions, which are fueled by the surge in smart devices and technologies like cloud and automation.
NLP technology: Chatbots find varied use
He also said there's a growing effort to continuously enhance the interaction between humans and machines using chatbots, which are becoming prevalent across the spectrum of services, especially where customer-service scenarios exist.
"Chatbots are finding varied use -- from personal to professional uses of human interaction and service needs. We expect a good demand on various design and implementation aspects in this area going forward," Kumar said.
For IT service providers that want to implement NLP offerings, David Schatsky, managing director at Deloitte LLP, said they'll need to discuss with their customers three key areas of a project's development:
- what text the customer needs to work with, and what the potential training set looks like (the training set refers to the sample text initially fed to the system to improve its ability to recognize and respond to queries);
- what key performance indicators are being targeted with an NLP solution; and
- what process will be deployed for customizing and training the solution and maintaining and enhancing it over time.
Schatsky said systems integrators, consultants and other IT service providers have to understand the concepts underlying NLP, the various architectures and approaches being used, and the limitations of the technology.
Kalyan KumarCTO, HCL Technologies
"Vendor training programs may be necessary, but they are unlikely to be sufficient. Providers should invest in developing a deep and broad knowledge base in order to provide sound advice and good solutions to clients," Schatsky said.
He added that a big challenge for the NLP market is that there is still no such thing as a domain-independent natural language understanding.
"A solution that works for, say, federal taxes, will not work for procurement or customer service. NLP systems need to be trained and customized for each application," Schatsky said. "Another challenge is predicting how long training will take, how accurate the results will ultimately be, and how accurate they have to be to deliver a strong business benefit."
Lalit Dhingra, president of the U.S. division of NIIT Technologies Ltd., a global IT service provider, said his company has embarked on a few pilot NLP implementations and what he has learned is that user friendliness is one of the key features that customers demand.
Finding NLP partners
To help with its NLP projects, NIIT Technologies recently partnered with Artificial Solutions to utilize their NLP technology platform to build artificially intelligent applications. Artificial Solution's software aims to help companies that are looking to improve customer experience in areas such as mobile personal assistants, bots, wearables and internet of things interfaces.
HCL Technologies' Kumar said finding the right partners can help IT solution providers organize and plan their NLP development strategy early. He cited the example of DRYiCE Lucy, HCL's NLP-based virtual assistant, which integrates with a number of technology providers. The DRYiCE Lucy ecosystem includes Amazon Alexa (for intelligent voice recognition capability) and IBM Watson Natural Language Classifier services in Bluemix (for returning the top matching pre-defined classes for short text inputs using machine language algorithms), according to Kumar.
Additionally, building an NLP solution that addresses linkages and relationships with various entities, while keeping the process and services aspects of the application in mind, is a good thing.
"Partnerships play an important role in the entire solution setup. The right systems integrator who can orchestrate through an ecosystem of providers and develop a pragmatic and business-aligned solution to address client business goals will be successful," Kumar said.
Read about a task assistant for Slack that uses NLP
Learn more about the Viv virtual assistant, which uses AI
Gain insight into the relationship between NLP tech and population health
How component pieces of chatbots can be used on their own for business value