According to the Gartner 2019 CIO Survey, 37% of organizations have implemented some form of AI — that’s a 270% increase over the last four years. Yet despite this soaring investment, research shows that 54% of companies are still unable to build on AI and natural language processing (NLP) technologies themselves due to talent shortages. This means many businesses are left looking at third-party tools that are inadequate for their needs and can’t be modified to fit their specific use cases.
But your business shouldn’t settle. It’s important to build an in-house AI team that can build the products you need. Here’s your blueprint for creating an AI team in this challenging hiring environment.

Find your dreamers

Every great invention begins as a vision. It’s your job to find a group of visionaries. For AI and NLP, this often means turning to PhDs and other academics steeped in the nuances of the field. For example, my company’s NLP team includes PhDs in computational neuroscience. They understand the technicalities, they understand what’s possible, and they aren’t afraid of the unknown — in fact, they’re excited by it. These are your researchers, and every great tech team needs them. That said, hiring academics comes with a very important caveat: They build for perfection. This is a positive, but it also means slower turnarounds. They’re happy to spend months — or even years — on one question, but that doesn’t jibe with today’s continuous delivery culture.

Follow the code

To counterbalance the dreamers, you need implementers. Hackers — the engineers, the people building the actual product — are the backbone of your AI team. These are the women and men who thrive on iteration, who love to test and make mistakes and learn from them. They don’t necessarily need to understand the theoretical; they just need to know how to bring it to life with code.
You can’t build the product without them, and that’s why, in today’s competitive hiring crunch, you need to get as specific as possible in your search. Don’t waste time with a booth at a college job fair. Instead, set up a symposium to present directly to the computer science department. Don’t expect giant tech events like SaaStr to turn up recruits; instead, send your technical leads to smaller AI and NLP events focused specifically on the vertical you’re building for.

Strengthen the foundation

This new crew needs a leader. Your executives need transparency into the projects. So you need someone who can play both sides and bridge the gap between your executive team and your AI team. On the surface, this can seem like an insurmountable challenge. AI lingo is intimidating and doesn’t exactly attract the average project manager to give up their current role for something unknown. As a result, businesses often make the mistake of either 1) putting someone too technical in this role, who is then unable to effectively communicate back to the executives or 2) giving the job to someone who is too theoretical and removed from the actual problem the team is trying to solve and thus cannot direct effectively.
Your leader should excel at taking inputs and creating outputs and be someone who closely understands the customer’s problems. This does not need to be a coder or an executive. This person doesn’t even need to be an expert in AI. The AI leads on my team, for example, are good at people- and project-managing. They’re also smart. Smart people can do anything — they can sell, they can design, they can build a product. Businesses should be open to different perspectives here and think outside the traditional team. Find someone who understands the problem, can learn the basics of the tech, and can distill and communicate it in plain English.

Build with diversity

Diversity in hiring has become one of the most important conversations in the last few years, and one of the greatest benefactors will continue to be AI technology. There cannot be great AI without it, because AI needs lots of data and lots of training to work well. The best data and training come from variety. The more perspectives, experiences, ideas, and worldviews you can bring to the table, the better.
You may think diversity in hiring means more legwork. It doesn’t. It just means different legwork. For example, those of us in Silicon Valley naturally think of CalTech or Stanford for hires, and there are great recruits there. But there are also excellent recruits at the University of Edinburgh, where they have a specific program for speech recognition. We naturally think about hiring coders, but what about bringing in linguists to support NLP development or diversity and inclusion experts to help build AI for HR functions? The more you broaden your scope, the more diversity you’ll naturally bring in. You should always be thinking about remote employees, students at overseas universities, and non-traditional recruits. In short, don’t throw out the non-traditional resume.
Overall, think long and hard about the challenges you’re specifically trying to solve for. Your team — no matter the makeup — should understand and believe in the solution.