AI in enterprise: How will machine learning and data science transform the marketplace?

AI will be essential for companies who wish to remain competitive in future

Most CIOs agree that AI will be essential in order to keep pace with their competitors in coming years, but the implementation of AI in production is proving more difficult than some might have imagined.

AI in enterprise generally refers to the use of various tools including natural language processing, big data analysis, machine learning and image recognition.  A Deloitte study conducted late in 2018 showed that machine learning and natural language processing/generation were among the most-used skills in the AI toolbox, with a 63% and 62% adoption rate, respectively.

AI still has its limitations and we’re not at the stage yet where it can complete creative tasks by itself.  But “augmented intelligence” that can be used in decision-making or other similar tasks has the potential to give a huge boost to productivity.  According to Chris Howard, Gartner’s research VP, “If you are a CIO and your organization doesn’t use AI, chances are high that your competitors do and this should be a concern.”

AI adoption faces barriers in lack of skilled workers, outdated software

Despite the eagerness to adopt AI, enterprises are finding it difficult to implement production with AI.  Svetlana Sicular, vice-president of research at Gartner, explains that the process is proving difficult for many companies because they currently run on outdated software.  As a result, whereas 21% of CIOs had said they planned to be in production with AI this year, only 10% of companies in this study had managed it.

A 2019 Gartner study based on data gathered from CIOs of major companies across 89 countries revealed that 37% of companies have currently implemented AI.  This number represents a growth of 270% compared to 2015, when only 10% of companies had adopted AI.  But the same study also found that one of the biggest challenges when it comes to implementing AI in enterprise is the AI skill shortage, a common theme which also surfaced in a recent O’Reilly study.

Open-source software will pave the way forward

A popular plan of action seems to be using open-source and cloud platforms as an aid to implementing AI, instead of developing proprietary AI capabilities.  It’s not necessarily that all companies will start using open-source software, says Sicular.  But they are likely to turn to open-source for ideas on how to implement AI in their companies, later relying on commercial vendors to actually carry out the transformation to production with AI.

To succeed, AI must be implemented across all levels

The same Deloitte study found that the three most salient advantages of AI in enterprise were in helping to enhance current products, optimize internal operations and make better decisions.  Every enterprise will be able to find a different use for AI, but the key is to think strategically and apply AI to the areas where your company will need it most.

Conduct experiments, carry out research, keep an eye on what competitors in your market are doing and don’t spread yourself too thin when it comes to implementing AI.  Sicular advises aligning ambitions across different levels.  Since a key setback to implementation of AI seems to be the lack of skilled staff, training staff will be a crucial step in the successful incorporation of AI in production.

Slowly but surely, we're seeing AI successfully implemented in more and more companies. There's no doubt it will be an essential tool for any company who wishes to remain competitive from now on.