In our previous blog post, we talked about the need of reskilling to succeed in the age of AI (1). Now our focus will be on who should be re-skilled and hired if an organization wants to become an AI Pioneer? Before we go further, however, it will help readers of this blog to classify their organization into which stage of AI development they are in. For this purpose, we stick to a classification provided by MIT (2):
- AI passives – Hardly any AI adoption and little to no understanding of AI technology landscape.
- Experimenters – Organizations piloting and adopting AI without deep understanding. Basically, learning by doing.
- Investigators – Organizations with knowledge in AI technologies but stuck in pilot stage with hardly any company wide deployment.
- Pioneers – Organizations that both understand, have adopted AI and are incorporating AI in their offerings and internal processes.
Hopefully, you are now able to classify your company into one of the above groups so let’s focus on key training and hires an organization could opt for when moving through different stages of their AI journey.
The AI passives
Training: For companies in the AI passive stage, before hiring any data-related people, the C-level executives and/or management team should be trained on topics such as AI awareness, what AI can and can’t do, what it takes to become AI-driven and modus-operandi of AI in an organization (hub vs spoke model (3)) . The next level of training should involve managers and team leads in topics similar to the ones for C-level executives plus data-driven business models, how to identify AI use-cases, how to draw AI project canvas, what are the MVP requirements and how to assemble the right team (4).
Hire: Data Scientists (someone with enough experience in Statistics, Machine Learning, Deployment, and some Data Engineering) and Data Engineers (ideally 2-5 Data Engineers per Data Scientist).
Training: The first level of training for these companies should also involve C-level executives, managers and team leads in space very similar to AI passives. The next level should be for people working with data (including IT professionals) in concepts in Statistics, Machine Learning, Data Engineering and Deployment. Most such companies usually struggle having a general knowledge about what AI can (and can’t) do and waste time in trial and error. Providing key training to people working with data can help reduce time-waste going into trying things.
Hire: The first and foremost person such companies should hire is a Chief Data Officer who can help the management team define data strategy and the corresponding AI budget. The next hires depend upon the issues you are facing. If you haven’t hired Data Scientists, then these should be your first priorities (profiles similar to AI passives). If you have hired Data Scientists that are struggling to bring use-cases into the pilot stage, you might want to join them with the Business personnel via training in AI Translator profiles (5). Data Scientists in such companies should also have knowledge in the AI deployment space. Finding such profiles, however, is very challenging (1) and the right way to proceed would be to train Data Scientists and DevOps people in AI deployment. Along with Data Scientists, these companies should also hire Data Engineers (again, ideally 2-5 Data Engineers per Data Scientist). Finding Data Engineers might again be very challenging (trust me on that) and if this is the case with you, it would be better to train your IT professionals in Data Engineering, Cloud and an understanding of Machine learning.
Training: Investigators display good knowledge in AI technologies, however, they are stuck in the pilot stage. This shows that these companies have hired Data Scientists but they lack a company-wide AI vision and systemic knowledge in AI deployment. Investigators should train their management team in understanding what it takes to become AI driven and what might be the right AI strategy for them. The Data Science managers and team leads should be trained in an Analytics Translators role along with an overview of deployment strategies e.g. how to assemble a right team (4). Along with Managers, Business Analysts represent ideal candidates to be trained in Analytics Translator roles. Ideally a pre-defined deployment-related strategy and infrastructure should be available to all Data Scientists and they should be trained in AI deployment-related technologies. To increase communication between IT professionals and data people, IT professionals including DevOps should be trained in concepts in Machine Learning and Deployment.
Hire: Similar to Experimenters, Investigators should hire a Chief Data Officer who can drive AI beyond the pilot stage and define AI strategy and key hiring space. A heavy focus should also be on hiring Machine Learning and Data Engineers (6) to move pilots into the deployment stage.
These companies are front-runners in AI and do not encounter issues similar to other three groups. They have well defined use-case prioritization and deployment strategies.
Training: Continuous training at different levels for hired Data Scientists in Deployment, IT professions in Machine Learning, Business Analysts and Managers in Analytics Translators’ role.
Hire: Product Managers and Product Owners in the AI space. If finding such people is a challenge, train Business professionals into these roles. Pioneer should also hire AI researchers for continuing their ongoing effort in AI development.
My goal from this blog is to bring an awareness around how companies can become AI driven while keeping humans in the center of their journey. There is no single recipe for AI success and relying solely on hiring new employees is certainly not a right strategy (talent supply is working against you as well). A mix of recruiting + training has led companies to move from being passive into AI pioneers and if your goal is really to become an AI Pioneer organization, then focus on right hires and right training.