Some of the tech industry's heavy hitters promise to give businesses access to high-powered machine learning algorithms that can parse text or speech. The releases are another sign of fierce competition in a market that seems to heat up every week.
The upshot for enterprises could be decreased time to development of advanced machine learning applications with speech and text analytics capabilities. That maps with growing interest in natural language processing and analysis. Twenty-three percent of 614 survey respondents said their organizations are using text analytics tools, with an additional 30% saying they're either piloting projects or planning them in the near future, according to TechTarget's 2016 BI and Big Data Analytics Market Landscape Study.
First, IBM released Watson Conversation, a cloud-based tool that enables enterprises to train chatbots using the Watson cognitive computing engine. The service is delivered through the IBM Bluemix cloud platform.
Developers don't need to be experts in developing machine learning algorithms. The service works by having the developer specify examples of end-user queries and appropriate responses. Watson then uses these examples to train its machine learning models and develop a bot capable of responding using natural language to a variety of queries. Bots can be deployed on web, social media and mobile platforms for a variety of customer service and engagement tasks.
Next up, Google delivered its Cloud Natural Language API. This service lets users analyze the content of text files for sentiment, content and syntax. Users can specify the text to be analyzed through an API call to Google's cloud-based service. Machine learning algorithms can analyze and parse something as short as a sentence or as long as a news feed.
Together, the two technologies continue the trend of cloud-based, managed machine learning services. Experts often tout the potential of advanced machine learning, and some leading companies have scored big wins using it. For example, Amazon has driven large sales volumes thanks in part to its recommendation engines built around machine learning algorithms. However, these models can be technically complicated and difficult to deploy. They demand a high level of expertise.
This has created an opportunity for the big tech companies to push managed versions of text- and speech-based machine learning applications, and competition in the space is heating up rapidly.
Scale has advantages for the software vendors. Machine learning models get better and better the more data they have to train on. As more customers bring their data to machine learning services, it represents new opportunities to train and improve the underlying models. This is one reason why competition is so fierce right now among vendors, as they vie to amass the largest user base.
In addition to IBM and Google, Facebook launched a service this year enabling businesses to build customer service chatbots in its Messenger app backed by its natural language processing algorithms. And Microsoft has made the development of chatbots a centerpiece of its strategy going forward.
Whichever company is able to obtain the most users is likely to develop the best services and win the machine learning platforms war. IBM has jumped out to an early lead with Watson, and the release of Watson Conversations solidifies its position. Google's new API will help it stay at the top of mind of potential customers. But look for the other players to remain active. This is one area of analytics that isn't likely to cool off anytime soon.
Machine learning algorithms are set to be a big part of many industries
Everything enterprises need to know about machine learning and AI
Emerging data preparation tools increasingly rely on machine learning