Artificial Intelligence and machine learning are squishy, inexact terms for most people. Particularly those who work for midsize businesses. Management in these companies hear great, wonderful and futuristic work being done in AI for other (larger) companies but have a hard time translating that into actions they can make in their own business.
For one thing, there are no widely agreed definitions of artificial intelligence, machine learning or predictive analytics. These are terms which have nuances and specificity to some people and may vary from practitioner to practitioner. Having a common language is fundamental to understanding but is lacking in the big hype called AI.
For the sake of brevity of this writing, allow me to lump everything into artificial intelligence or AI.
Part of the hype is the sense AI is going to change the world of business. Probably true in many senses, but how can we adapt it to a midsize company today? Big AI ideas using big data sources get big headlines and results. More mundane applications of statistics to transactional data to derive usable predictions in existing business processes just don’t get the glamour, but still delivers results.
What a lot of companies need is not a data scientist or another technology industry generated title. What executives at midsize companies need is a person to guide and advise – translate the technology into real business processes delivering results.
The ideal person needs to understand the business process like an executive, comprehend the data meaning like an analyst, navigate databases like IT, and apply statistics like an academic researcher. This person has to be able to extract meaningful data and create AI models where the output is directly usable in business processes. Also, this person needs to be able to work with outside application vendors to tailor AI for the upcoming releases of enterprise software (think ERP, CRM and other vendors who are building AI type functionality into their general products – it needs to be fitted to an individual company’s business processes).
This person may or may not already exist on the payroll? Perhaps this task is better done by an outsider who does this on a frequent basis. This may be ideal for midsize companies may not need this person to be a full-time employee especially when they are investigating and learning.