Natural Language Processing (NLP) learnings from the entrepreneurial front lines

NLP – what’s in it for you?‍

I love solving a good challenge. Gathering bits and pieces of info, making sense of the patterns and figuring out a new solution is what drew me to engineering to begin with. It’s also a big part of why I’m an entrepreneur focused on the tech space.  

This fall marks two years since my latest initiative – an AI-backed solution changing the game for corporate recruiters and organizational security teams – took shape at the R&D stage. I dove into this project because I fell in love with the problem: recruiters were commonly seeking online info to support decisions about candidates, partners and other stakeholders. But when done manually, that process was a whole lot of hit or miss. Recruiters couldn’t be sure they were searching the right “John Smith”. Context was critical, but open to subjectivity from one recruiter to the next. Source credibility was debatable. And above all: the process was massively time-consuming.

We knew natural language processing (NLP) could be an excellent match to help humans accomplish these tasks better, faster and – importantly – at scale, as candidate numbers grew exponentially. What we didn’t know was just how much things would change as we launched Valital, the output of our NLP work. Yet, even as the world continues to shift, I’m amazed by how much the entrepreneurial fundamentals that spurred our progress continue to ring true.

Whether you’re building an NLP solution or another tech-based platform, listening is key. One of the core benefits of AI like this is its ability to reduce judgement, prejudice and bias. That’s what makes solutions like ours so powerful. NLP helps machines read, understand and make sense out of what humans say or write (even, in Valital’s case, classifying text and articles). Our team has done a lot of listening in the months since we built our initial solution for recruiters. We’re all ears, and what we’re hearing has fueled our success, enabling us to subsequently launch a bespoke solution for security teams. Enterprise risk groups are now leveraging our NLP platform to assess people’s conduct faster, and more comprehensively, than ever. We got here by developing the same way our technology works: assume nothing, ask everything, and adapt models to what the data and feedback dictates. That’s huge.

True, too, for developing AI solutions with real people in mind. Employing NLP – or any form of AI – isn’t just about reverse-engineering solutions to what helps people work and live better. It must also come down to doing so in a way that’s ethical, and focused on leveraging the power of tech in the right ways. For us, that meant focusing relentlessly on privacy by design. Without that, we knew our solution would miss the mark. 

Because our platform scans public online records in minutes and then collects, validates, analyzes and classifies information about an individual’s conduct, the way we manage it is absolutely vital. Without privacy at its core, it doesn’t matter if our tools help people make objective decisions, standardize processes and move their businesses forward with confidence. For our team, making the most of AI continues to mean doing so in ways that have meaningfully net positive impacts for stakeholders and their stakeholders. That’s foundational.

Above all, this year has reminded me that the answer to the most complex problem is often the simplest solution. NLP and other forms of AI represent endless potential. No doubt as the pandemic continues, tech entrepreneurs will play a vital role in solving the challenges this crisis poses, and the broader economic difficulties born of it. But that starts by getting down to the heart of what your users are struggling with, and building directly from there. Simply, efficiently and with immense attention to the user’s reality. That’s the bottom line.

Watch this space for our next R&D blog on building an effective NLP model, and some of the best open source tools available today.