On Nov 30, 2022, OpenAI created a “What Hath God Wrought!” moment with its ChatGPT announcement, setting in motion a deluge of frenetic “We are AI-enabled” and “We are AI-powered” pronouncements by companies big and small. The truth is that not many companies know what to do with Artificial Intelligence (AI), let alone how to do it effectively. However, this is changing rapidly, as companies are facing investor and competitive pressures to quickly become more efficient by utilizing the benefits of AI technology. While exhilarating for some, AI has its unique challenges for successful development and implementation of solutions developed using this technology. A thoughtful organizational structure with key skills will ensure success in identifying valuable initiatives that create value for the enterprise and mitigate the challenges. 

Gloom, Doom, Zoom, or Bloom? 

Reid Hoffman, the founder of LinkedIn, describes in his recent book Superagency four mindsets or ways in which people relate to AI and categorizes them as: 

• Doomers, those seeing AI as an existential threat 
• Gloomers, those thinking AI will lead to job losses 
• Zoomers, those enthusiastic about AI wanting to go full throttle 
• Bloomers, those who are cautiously optimistic and want a balanced approach. 

Most organizations and their leaders would be better served by recognizing the importance of having a Bloomer approach to harness the transformative potential of AI technology. If you are a mid-size or a small organization, sooner or later your technology team needs to be able to build and implement AI solutions.  

Skills and Roles for an AI Team 

Let us start with the foremost concern in the minds of leaders implementing AI in their organization: Is what we are doing and how we are doing it ethical? Authors Beena Ammanath and Reid Blackman in a Harvard Business Review article indicate that to vet the ethical risk of a project, the most vulnerable constituents of an organization are procurement officers, senior leaders, and data scientists and engineers. They argue that not just them, everyone in an organization needs to understand AI ethics. This points to the unique challenges in building an organization at the heart of development, implementation, and use of AI technology in various organizations.  

To build and implement AI solutions, a range of skills are needed that help with various dimensions of building and implementing AI solutions: Technical, Functional, Process Integration, Change Management, Ethical, Legal, and Environmental. These skills require different roles to be in place: Business Analysts, Data Engineers and Scientists, Machine Learning Engineers, Ethical AI Developers, and Change Management Experts: 

• Business Analysts: Help identify problems, gather requirements, and seek out process impacts. This team also helps management to make informed decisions about what a suitable AI initiative for the organization could be. A simple framework that we recommend you use is to plot projects or various use cases of AI implementation on two axes: Tolerance for error in results and Intensity of use. The highest usage volume use cases in the organization with the most tolerance for error are the ones to pick first.  
• Data Engineers: Prepare data pipelines that are suitable for training AI algorithms on. Let us face it, the organizational data, especially for a small or medium organization, is scattered and not in an AI-consumable form. By having the skills to prepare organized data pipelines, the team will take advantage of converting generic and topical AI solutions into meaningful sources of insights for the organization.   
• Machine Learning Engineers: Build algorithms to continuously learn from data and fine-tune the underlying models. They optimize the performance and scalability of AI solutions and prepare them for wider deployment. 
• Ethical AI Developers: Ensure data privacy, mitigate biases in AI models, and promote fairness and transparency in AI applications. Use of AI has wide implications given that the base models have been inevitably trained with biased training data and the organization must ensure that such bias is eliminated. In addition, there are strong environmental impacts to using AI technology in general, with each Graphical Processing Unit (GPU) – a core element of AI infrastructure – using up to 4 times more electricity compared to a regular computer’s Central Processing Unit (CPU) and then needing many times more cooling. Responsible usage means reducing negative network externalities.  
• Change Management Experts: Well before an AI solution is conjured up, it is critical to identify what Dr. Heath calls “What is the goal of the goal?” Only when we understand the fundamental rationale and utility of doing an AI project, not simply for its buzz-word value, that we will be able to manage the change the organization must go through. The users need to see the value and must adapt the new tools for effective returns on investment. 

How to Hire for AI? 

There is a shortage of talent in this space and many organizations are trying to figure out how to fill the talent gaps rather quickly. In the rush to fill open positions, it is critical that the organizations are aware of key considerations, given the roles and the unique challenges of the technology: 

• Specific technical or functional skills for various roles are table stakes and must be strong especially for AI/ML and Data Engineers. Technical skills needed span a strong understanding of the science of data, a strong foundation in algorithms, and an expert knowledge of programming 
• While looking for talent, we need to shift our mindset from evaluating talent on “what you remember” to “how you think.” There are three most important soft skills needed in the assistive-AI era: 

1. Critical Thinking 
2. Problem solving 
3. Collaboration 

• After hard and soft skills, the next fit to evaluate is the harmony of values between the team member and the organization. Any conflict in values is very risky, and these gaps widen when faced with dilemmas that can only be resolved by tapping into the values of the organization.  

There are techniques that an organization can use, and much has been written about them, to evaluate technical/functional skills, soft skills, and to identify values in potential AI hires. The Interviewing should also include responding to moral dilemma type questions or case studies to get insights into the thinking of the person you are trying to hire for your AI team. In addition to building an in-house team, there is the option of partnering with firms specializing in AI integration and AI talent augmentation, the former you can outsource to, while as with the latter you get the benefit of not having to handle the administrative burden of full-time hiring while scaling your team easily. 

Developing and implementing AI solutions comes with its own set of unique challenges, given that the technology is nascent and has multi-dimensional impact. However, by establishing a thoughtful organizational structure and hiring team members with relevant skills, enterprises can successfully identify impactful initiatives and execute them to create significant value for the organization. 

Schedule A Free Consultation