How Natural Language Processing Will Reshape The Software Landscape

Artificial intelligence (AI) and machine learning have revolutionized the way consumers do business. Thirty years ago, people bought a newspaper, looked in the yellow pages or went to the library if they needed information. Today, everyone goes to the web, where they can type in a request and the information appears before them in seconds. This same technology, known as natural language processing (NLP), is now popping up in the business application world, and it’s about to reshape the software vendor landscape. Here’s why.

Since the early 1990s, the world of corporate software has been dominated by large ERP (enterprise resource planning) software vendors. The reason these vendors are so entrenched is that they rely on complicated workflows that users must master and step through to perform their daily tasks. Learning how to use these systems often requires months or years of training and is to the point where acquiring these certified skill sets has come to be the equivalent of getting an advanced educational degree. This has allowed the likes of SAP, Oracle, IBM and others to cultivate massive user groups that bring demand for their products wherever they go. SAP had over 30,000 people attend its 2018 Sapphire conference.

However, the capabilities provided by NLP enable users to shortcut these laborious menu-driven processes to jump right to the information they are looking for, thus removing the need for much of the extensive training and custom configuration mandated by the ERP vendors. These service decreases will continue to erode a large source of revenue from providers and their associated service partners, which have already been in decline due to the movement of applications from on-premise to the cloud over the past decade-and-a-half.

But of larger importance is the impact NLP will have on the dogged loyalty of these large user groups. A hospital administrator who uses Google to find the nearest pharmacy will also want to use an NLP search in their HR system to find a scrub nurse in Philadelphia who speaks Spanish. They won’t want to step through multiple menus and drilldowns to get that result. Thus, as less training is needed due to more intuitive NLP shortcuts, the less important ERP certifications will become.

Further accelerating this shift is the fact that leaders of a new generation are taking center stage. The oldest millennial is 36 years of age in 2018. This generation has grown up with smartphones and Google. They want things now and have been accustomed to getting them. If they have to go through extensive training and can’t get timely answers, they will go elsewhere. Add in the economic factors of quicker access to information, which means less employee cost and more ready insights upon which to take action, and the C-Suite will be on board.

However, although NLP is here today, it relies on machine learning to get smarter. When an AI engine is first implemented, it’s loaded with extensive glossaries and grammatical rules, but it’s through continued use that the system picks up different ways of stating things. When someone does a Google search, the return is a ranking of the most likely results. But the AI in Google then uses the selections made after that to continually learn what people are looking for and then re-ranks the results. It also incorporates other words and nuances that might often be repeated in the resulting content to continually learn and produce more optimal future search results.

That said, it won’t be difficult for AI engines to ramp up in the business-to-business space. The reason for that is the glossaries of terms and structure of queries for such corporate disciplines as HR and finance are much narrower than the entire world of language that Google and other internet search providers have to deal with. Take the human resource software space, for example: Most glossaries in HR consist of fewer than 1,000 terms (see SHRM and HCI). Granted, when you search with more filtered queries, the universe of results expands exponentially, but it’s still a very finite universe.

Although this shift may not happen overnight, it is coming. The combination of technical innovation, ease of use, new generational leadership and positive bottom-line impact will create a wave that corporations won’t be able to resist. The large vendors will adapt in some way, but the result will leave a large opening for new vendors to get a foothold and alter directions going forward.

This article first appeared in Forbes.

Connect with Tom

Tom McKeown was recently the CEO and Co-founder of TrenData, which was acquired by isolved HCM in 2021. He currently manages the product team and business unit for their People Analytics offering.

Explore Topics

Tags