We will participate on the June 3rd IASA OnPOINT conference panel hosted by SS&C Technologies. Topics covered will include AI, machine learning, deep learning, NLP, RPA and other forms of intelligent automation. There are many prospective use cases for AI and related technologies. However, prior to jumping in with both feet it is important to define the business problem.
Define the Problem, Understand the Data and Don’t Limit Your Imagination
In deploying artificial intelligence (“AI”) or one of its sibling technologies – machine learning and deep learning – the first order of business is defining the business problem. Next, understand your enterprise data and third party data in terms of scope and quality. Once those elements are in place, you are ready to embark on your AI journey upon which your imagination will be the primary limiting factor.
1.) Define the business problem. For example:
- Predicting sales for the second quarter;
- Predicting automobile collisions at a particular intersection;
- Understanding the risk of collision on a policyholder’s daily commute;
- Perhaps you wish to train an autonomous vehicle to drive?
- Perhaps you wish to identify potential underground nuclear weapons facilities across the globe?
- On a lighter note, maybe you wish to deploy an AI to recommend coffee drinks to customers.
These are problems that can be answered by deploying some combination of AI, machine learning, deep learning/neural networks and/or natural language processing (“NLP”).
2.) Once you have defined the business problem it is important to understand the data you may leverage to solve the business problem. Data quality is important – “garbage in, garbage out”. Data consists of both your enterprise data as well as any third-party data you may leverage. Your customer data is an example of the former and licensed demographic data is an example of the latter.
3.) The final element – use your imagination! Computing power continues to increase while cost decreases. AI, machine learning, deep learning and related technologies will continue to advance and new technologies such as augmented and virtual reality (“AR” and “VR”), enable users to add an immersive layer.
AI, Machine Learning, and Deep Learning Examples
Machine learning is effective at identifying patterns in data. Predicting automobile crashes, sales activity, extreme weather events, fraud detection, movie recommendations, and identifying optimal locations for physical retail stores are all examples of problems machine learning is equipped to solve. In the movie recommendation example, Netflix uses machine learning to make personalized recommendations based upon subscribers’ viewing history and other data inputs. Similarly, SS&C Singularity leverages machine learning to observe and optimize trade-related workflows based upon usage and patterns detected by the ML algorithms. Netflix and SS&C leverage these technologies in a manner that is transparent to the user. The user doesn’t need to know (nor should they want to), which technologies power movie recommendations nor how or why investment operations workflows continuously improve over time as “the machine” learns.
- Many cloud-based technology companies (Facebook, Google, Salesforce.com, Zuora and more), use machine learning to better understand customer usage at the individual user level. Companies then use that intelligence to make product enhancements – which subsequently are disseminated across the customer base (in multi-tenant environments) – improving product/service quality.
Deep learning is effective at understanding images, text and voice and has made gains over the past few years in these areas. For example, autonomous vehicles use deep learning models to provide context about the world around them – distinguishing buildings from street signs from people to other vehicles. Deep learning models may be utilized to understand satellite imagery, to power facial recognition technologies and to make products such as Google Assistant, Google Lens and Live Caption possible. Deep learning may be used to infer personality traits from written text. Heck, deep learning is used to autonomously render the background visuals (buildings, mountains, sky, birds) in video games.
Companies typically use a combination of these technologies to deliver a service. For example, machine learning may be used on the back-end of a particular service in combination with a deep learning neural network that powers a conversational AI wrapper (chat bot for example). Or perhaps one wishes to model potential traffic collision rates for a proposed intersection. A 3D simulation that incorporates vehicle traffic patterns, street signs, traffic lights, and video could be quite powerful. Augmented reality and virtual reality add another dimension to 3D simulations.
IASA 2019 June 3rd 1:00pm Session
Should you plan to attend the IASA 2019 conference in Phoenix I would enjoy meeting you Monday June 3rd at the 1:00 pm OnPOINT session – “Working Smarter: How Machine Learning and Intelligent Automation are Transforming Insurance Investment Operations and Accounting.” Feel free to reach me at email@example.com