With the ongoing talk about all things AI, it is important for companies to make sure they have the right AI skills within the organisation. But what should businesses consider when they are attracting and retaining critical AI talent? Sid Bhatia, Regional VP & General Manager, Middle East, Turkey & Africa at Dataiku, outlines the dos and don’t when hiring people with the right AI skills.
Just when businesses thought they understood all the angles of Artificial Intelligence (AI) – what it could do, how to implement it, its business impacts and its ethical and compliance implications – along came Generative AI (GenAI). Even an everyday AI business, where the organisation has matured to the point where data and AI have become part of its culture, would have been taken aback by the capabilities on display. One projection for the Arab Gulf region puts the overall economic impact of GenAI at some US$23.5 billion per year by 2030, with Saudi Arabia and the UAE taking the majority at US$12.2 billion and US$5.3 billion, respectively.
However, when it comes to attracting and retaining critical AI talent, GenAI does not force a return to the drawing board. As long as an organisation remembers to align hiring targets with business needs and takes steps to cater to the expectations of different AI talent groups (especially expectations of upskilling and professional development), it should be able to maintain its everyday AI culture.
The AI professional is not a single category with a single approach to attraction and retention. Different roles call for different considerations throughout the employee lifecycle. When hiring an analyst, for example, their skillset would be so broad as to necessitate a thorough assessment of their understanding of the business and its specific needs. When hiring a data scientist, their predilection towards examination, curiosity and problem solving presents a challenge in presenting enough challenges to attract and retain them. Leaders, architects and engineers have their own unique recruitment and engagement hurdles too. The road to everyday AI is a twisty one. Many bends are talent-acquisition issues, so let’s examine the trail as a series of dos and don’ts.
Do pursue balance
Data scientists are gems of talent but without enough data architects to ensure the right database architecture, the organisation will not be able to efficiently deploy, enhance and scale Machine Learning models. A lack of business impact can lead to low morale and resultant attrition in talent. Likewise, if data leaders are in short supply, things like communication, strategy and prioritisation can fall by the wayside. This can lead to silos and missed opportunities for model reuse. Get the balance right, however, and you can build a slick innovation factory staffed by productive and satisfied employees.
Do hire a diverse workforce
Discussions about AI in the region often swerve towards responsible AI. It has long been argued that the best guarantor of ethical systems is a diverse team. The everyday AI enterprise understands that a diverse workforce, split among teams that do not collaborate, is a waste. Diversity can only add value when as many viewpoints as possible are allowed to coexist in each project. This extends to collaboration between skill levels. Restricting data and AI operations to only the most specialised roles will prevent scalability, sustainability and democratisation. So be sure to mix different abilities and backgrounds. And appoint strong leaders who can get the most out of these disparate profiles to infuse the entire organisation with an AI culture that can deliver unbiased, responsible, transparent systems. This is the kind of organisation to which young AI talent will flock.
Don’t fly blindfolded
Before recruiting, you should know what your expectations of the new hire will be. What projects will await them? From where will the data come? What business goals do those projects address – entire operational chunks like the supply chain or a production line, or smaller tasks like new reporting or self-service features? The answers to these questions will help you with the job description and the questions you ask in interviews.
Don’t chase unicorns
The truth is it takes a village to form an AI team. Trying to locate a one-person band with every required skill is unrealistic (unless you are Google, Facebook or Microsoft) and unnecessary. It is also counterproductive because if you were able to find such a person, they would quickly become irreplaceable, which would represent a threat to the AI function. Instead, concentrate on what skills you need in order to fulfil your specific business goals. If you can define a business problem that you wish to solve, present it during interviews and ask candidates how they would design the project. Never forget the skills spectrum. While you do not want a unicorn, you do want each recruited AI professional to exhibit multiple abilities. Data scientists, for example, must have good communication skills and a sound knowledge of the business and statistical methods.
Next steps
Success in AI recruitment requires strict mapping of job descriptions (which include technical and soft skills) to business needs. Data teams are eclectic groups of different abilities and backgrounds. While this is true anywhere in the world, it is especially true of the GCC. All enterprises want to build a scalable and sustainable AI strategy. But they must balance their ambitions of future growth with the realities of acquiring and upskilling a talented team.
As GenAI increases its hold on the regional business community, new strategies will emerge, but these initiatives call for a rethink of responsible AI, compliance and sustainability. It has never been more important to look at how AI fits into your business and how this relates to recruitment needs. Data science, Machine Learning and AI platforms must be implemented in a way that makes sense to the business. But without the right people – those who understand the technical side of solutions but also understand the business and can communicate and collaborate effectively – it will be next to impossible to integrate these powerful capabilities in a way that will add long-lasting value. And it will be similarly problematic to build an inclusive and sustainable AI and data-democratisation culture. Collaborative, everyday AI enterprises emerge through everyday consideration of all things AI, and that includes the dos and don’ts of recruitment laid out above.