Artificial intelligence (AI) and machine learning (ML) are hot topics today. Along with big data and analytics in general these present a daunting challenge to management. Just how much of AI do you need? Should it be done with big data? Does it need a billion rows of data? Can you really use the 100+ machine learning algorithms out there? Do you have a skilled staff, including data scientists, to identify which issues are best for you to address? Can your managers use these technologies to help them manage better?
Organizations require an answer to these types of questions when considering moving into these technologies. However, much of this, but not all, is overkill for many organizations. Managers, analysts and consultants need a basic and core set of analytics that they can apply to a variety of problems and opportunities in an organization. Along with a core set of analytics, a rational and realistic plan should be in place to realize the value and benefits from change efforts.
Some key insights about AI
A little background here helps. There are 6 basic type of artificial intelligence technologies that all the methods can be grouped under:
- Statistical Machine Learning
- Simple Predictive Neural Nets
- Deep Learning
- Expert Systems
- Decision predictive techniques
- Knowledge leverage
Some of these - decision prediction and analysis, expert systems, machine learning - have been around for a long time. Others are very new like the working predictive neural net technologies while some, like deep learning, have very useful applications but are still emerging. Understanding which to use and how and when to use it is the real challenge.
Some basic suggestions
When embarking on a path involving these technologies it is useful to follow some simple guidelines:
- Avoid launching enterprise wide project. Past experience with very new technologies shows this is dangerous and fraught with failures. Indications are that complex problems are still difficult to solve with deep learning neural nets and have led to a number of failures. Small projects on the other hand have a high degree of success.
- Have a clear objective for each project
- Develop an AI risk profile for your AI portfolio
- Grow your skills as you increase your AI use
- Get good help and make sure there is knowledge transfer
KCI can help
KCI provides several services that can help you achieve success in your AI journey. Here are some of the ways we can get your AI projects moving on a successful track:
- Professional team training in AI and ML
- On the job training and mentoring for project teams
- Tool selection and matching to need
- Focus on AI requirements and plan
- AI Risk analysis
- Knowledge transfer
