Artificial Intelligence (AI) is a comparatively new field that is redefining business processes across a wide range of industries. There is no official guidebook for investing in these technologies, but there are lessons to be learned from others who have traveled down this road before.
What you need to know about investing in an enterprise AI system
Artificial Intelligence (AI) systems are redefining business practices across verticals and around the world. While simple plug-and-play chatbots and automations might fall under some official definitions of AI, true enterprise-scale disruptions arise from complex autonomic systems that bridge the back- and front-end offices and tap into machine learning (ML) to improve over time. These multi-faceted cognitive systems demand a level of investment that’s probably beyond the reach for your local Mom-and-Pop store, but isn’t solely the domain of the Fortune Global 500 either. As your company moves forward with an AI solution, here are some things to consider:
There are different costs for different solutions
There is a wide spectrum of AI functionality, with more sophisticated systems demanding additional investments in time, human resources, and money. The first step to implementing AI is determining which solutions will offer the most impact for your business.
For example, a digital colleague like Amelia can be trained to automate many customer-facing interactions including everything from answering basic FAQs (“Amelia, what is the closest bank branch to my house?”) all the way to independently resolving complex customer needs from end-to-end (“Amelia, I’d like to open a new joint checking account”). These interactions require some training and resources, but the latter task obviously requires additional preparations.
The good news is that as Amelia gains more real-world experience, we are able to imbue her with “out-of-the-box,” industry-specific functionally, which means our customers need to invest fewer resources to get up and running.
AI does not necessarily mean instant ROI gratification
AI opens the door to unprecedented efficiencies in an organization, but companies shouldn’t expect these returns to materialize immediately. Some specific features and functionality can be installed beforehand, but AI’s true potential lies in its ability to recognize its limits and independently expand them by observing human-human interactions. To that end, these systems’ ability to self-optimize necessarily takes time.
While there is a ramp-up period until AI systems are able to function with extremely limited human oversight, the returns become very tangible once they do. As with any new hire, there is a learning period before their full potential is realized. However, unlike a single human hire, the training period doesn’t result in one capable employee – it provides an entire workforce that scales with your company’s need.
How to move forward
Companies shouldn’t think of cognitive solutions as a one-and-done investment, but rather as a prolonged relationship that needs to be sustained over time. In a recent blog post, Allan Andersen, IPsoft’s director of enterprise solutions, explored what companies need to consider when designing an in-house cognitive Center of Excellence (CoE).
Specifically, Andersen said the key is to view the process in stages where companies steadily gain more independence to develop their solution over time. As he put it: “The first stage or level, is a ‘seed’ level. This means it doesn’t require a huge investment with lots of staff, but just enough to have a presence within your business. We recommend establishing this shortly after the first Amelia production use cases.” Additional steps and levels lead to a point where the CoE takes operational control over the AI systems.
While there is a ramp-up period until AI systems are able to function with extremely limited human oversight, the returns become very tangible once they do.
Many companies choose to have AI platform master one part of their business and then expand its roles in stages. For example, an international financial services company recently hired Amelia for a pilot program where she quickly reached a high 87% accuracy during conversations. After this initial success, the company deployed her for an expanded role in which she will take nearly 40% of all IT requests, from routine L1 issues to more complex L3 incidents, and resolve them from end to end. (You can read more Amelia case studies here.)
There are no set rules about what an AI system implementation should or will look like. Ultimately, customers should use AI as part of a broader strategy for achieving maximum business value and sustained growth – taking the path that will serve their operations best.