Create a Top-Notch Artificial Intelligence Team

If artificial intelligence is going to work, IT must align itself with other disciplines throughout the company and form a cohesive team. This team must be able to handle everything from technical execution to content curation, heuristics and algorithms, along with compliance. How doable is this for most companies, and how can you assemble the best team possible? 

I went to the kitchen to ponder this, and I quickly found an analogous question: How many different chocolate-chip cookie recipes are there in the world? The answer is thousands. Likewise, it’s the same, almost limitless outcome when one considers the broad range of small- to king-size companies in the world that want to use AI. Some companies already have different AI team skills in-house, while others must find ways to work with what they have.  

Identifying the Players on an AI team 

To build and sustain an effective, end-to-end AI team, companies need individuals who understand internal and external IT infrastructure; know how to maintain and monitor parallel processing in the data center or the cloud; and individuals who have the statistical and heuristics background needed to develop questions that can mine troves of data so that pearls of insight can be uncovered.  

Companies also need individuals who can manage the uniquely iterative nature of AI projects, individuals with a comprehensive understanding of the business and its challenges, and individuals with the skills to develop and evolve AI models through iterative machine learning and training. They also need employees with patience, strong communication skills, and excellent training chops who can show users how to use the AI. Finally, there is the governance and compliance side. How do you make sure that there are adequate guardrails, so the AI doesn’t breach subject matter areas that are inappropriate? You need people for this, too. 

Related:Why Every IT Leader Needs a Team of Trusted Advisors

This is a recipe of staff competencies for AI that only a few organizations can afford.  So, what about the smaller companies that want AI but must work out their own recipes with missing ingredients? 

Put Your Best Foot Forward 

The Merriam Webster dictionary defines “putting your best foot forward” as “to try as hard as possible to do something difficult.”  

Building a top-notch AI team can be difficult.  

In fact, Deloitte says, “Assembling teams that can succeed in that [AI] landscape requires companies to re-examine everything: 

1) The teams themselves, including individual job descriptions, titles and career paths; 

Related:Struggling to Build AI Teams? It’s Time to Focus on Skills

2) Team structures, including organizational design, internal alignment and the integration of skills and capabilities, particularly with the increasing reliance on outside talent; and 

3) Team enablement, including culture, communications, collaboration, continuous learning, re-skilling, and up-skilling.”  

The consensus at Deloitte and other technology research companies is that it is difficult, even for large companies, to find all the integrated AI team skillsets in-house because of the STEM shortages in the job market. Plus, there is common acknowledgement that at least some AI skills will need to be obtained through the outside contract market.  

Building an AI Team 

Knowing that you are bound to have AI skills shortages in-house, creating an AI team begins with identifying all the skillsets that you’ll need, and then penciling in those employees whom you believe can handle different roles. Inevitably, some of the boxes won’t get filled in, so you’ll need to acquire the talent for them elsewhere. 

Here is what most companies find: 

A highly skilled project manager and communicator can usually make the adjustment to manage an AI team, and many companies have this person on their application development staff. Most companies also have subject matter experts on the user side who can validate whether the AI outcomes are what one would expect, and also can identify those business areas in need of AI.  

Related:AI Speeds IT Team Hiring

Where the skills gaps are most likely to be found include the following: on the data science staff, where it’s costly and difficult to hire skilled people in a tight job market; on the IT infrastructure side, because IT is familiar with data and systems but lacks familiarity with the parallel computing AI uses; on the regulatory and compliance side, because no one -- not even regulators and auditors -- exactly knows where the legal and ethical guardrails for AI will be placed. In addition, there’s a lack of ability on in-house staff to develop and perfect machine learning models for AI systems, and in the business enablers from both user and IT staff who can effectively both train users on how to use AI in daily work and to revise business processes.   

Filling the Gaps 

While hiring the skills that you need is the ideal solution for every company’s AI team, those skills are in short supply, and many companies just can’t afford them. 

In some cases, similar organizations share costs and resources. A good example would be a cloud-based AI medical diagnosis system that several different healthcare organizations use and collectively pay for. 

In other cases, such as the pharmaceutical industry, it’s important to develop your own proprietary AI system for a competitive edge.  

In still other cases, SMBs that want AI concede that they must use packaged AI from their vendors, although using a generic AI will only give you the run-of-the mill insights that everyone gets. 
The bottom line for most organizations, large or small, is that it’s difficult to form and retain a complete AI team in such a highly competitive job and skills market. You may have to get outside help. This can be a successful strategy if you clearly define where you need the help and how you can transition the knowledge that they bring to the internal knowledge set of your team.   

What we know is that AI teams must be interdisciplinary in nature, and they should be as heavy on business know-how and expertise as they are in IT and data science skills. Uniquely gifted project managers with patience and great communications skills are needed to manage these cross-disciplinary teams, whether the teams are fully staffed internally or a combination of internal and contract hires. 

Finally, like baseball, AI is a team sport. Errors will be made, and insights will be left on base. But as the team matures, execution will get better, and companies will begin to see the results. This is what we’re striving for.