I was once tasked with creating a roadmap for AI for a business support organization. The popular belief is that it is a game-changer, potentially disruptive, and is supposed to make life and business better for everybody in general. But how do I demystify the power of AI? Where is the start?
Setting a vision or desired future state is the first step in the process of defining the desired pathway for a technology’s deployment. You will need to identify what you hope to achieve. I started with use cases that were easy but also the pressing problems in the operational world. I started scouting the pathways prescribed by the gurus in the art of problem-solving and soon found out that AI would not readily fit. Even though the epics were broken down into several stories and use cases, navigating in the traditional process posed several hurdles.
So, I decided to start on the other end — using Porter’s three generic strategies as a reference, identified areas of how AI technologies will impact cost leadership, differentiation, or focus to give you a competitive edge. I chartered the course into three tracks: RPA, Machine learning, and NLP. Each of the technologies addresses different types of issues encountered in the gamut of activities in operational processes. The conglomeration of these tracks would then converge to a solution that would completely address the business problem.
RPA automates high volume, low complexity, routine administrative tasks — of the types that one would drop into an outsourcing bucket. However, setting up an outsourcing contractor that would ‘pass’ all security clearances is a daunting and expensive affair. By deploying RPA solutions, the administrative staff is freed up to handle more complex issues.
Machine learning as a discipline tries to design, understand, and use computer programs that learn from experience (e.g. photo analysis), without being explicitly programmed for specific modeling, prediction, or control tasks. Human analysis of a photo is not just time-consuming; it is also requiring domain knowledge. Using this ammunition in AI, a high volume of pictures from the field is analyzed and used for quality assurance.
The capabilities of machines referencing types of NLP used in generating natural languages such as text summarization, machine translation, spam detection, sentiment analysis, and information extraction. Existing NLP tools can directly be applied to customer feedback, order notes, and troubleshooting notes to scout for process gaps.
The use of artificial intelligence to resolve simpler problems provide an opportunity to foster user adoption. And the collective advantage of the collaborative efforts from people, processes, and technology was more visible than ever before. This approach has a cascading effect and paved the way for rapid developmental activities.
In conclusion, if you are unable to launch AI in one fashion, there is an alternate strategy that works. Disparate solutions in RPA, ML, and NLP all come together to solve a bigger problem and help organizations tread forward in adopting AI technologies for the greater good. Often, getting started is half the problem solved.