Welcome to the world of Infosys Engineering! It is a half a billion plus organization that takes pride in shaping our engineering aspirations and dreams and bringing them to fruition. We provide engineering services and solutions across the lifecycle of our clients’ offerings, ranging from product ideation to realization and sustenance, that caters to a cross-section of industries - aerospace, automotive, medical devices, retail, telecommunications, hi tech, financial services, energy and utilities just to name a few major ones.

« Imagining a digital twin for your connected car | Main | Delivering (renewable) Energy as a service »

Amplify connected vehicle services through continuous learning

Humans naturally are learning species - we start doing things better once we get a "hang of it". There's always been belief that we can create intelligent systems that learn and mimic human behavior. While we are still away from building a replica of the human brain, there's technology advancements to accelerate this journey. Using a simple framework of Learn, Anticipate, and Validate, I will illustrate how the connected services can be enhanced and expanded to deliver superlative seamless experience.

Before we begin, commit to a journey that is evolutionary and iterative it is important to create a product strategy. The first step towards creating learning systems is to discover what aspects of behavior do you want to emulate, Secondly user acknowledgement and realization is crucial for adoption. Thirdly, having the roadmap that provides direction to continuously build and improve the intelligence quotient of these systems.

  1. Discover the appropriate learning by asking probing questions about what we want to learn. I'd go a step further and prioritize these questions as a roadmap starting with foundational. So for instance in the connective vehicle scenario - the various categories are in-vehicle experience, driving behavior, vehicle health, dealership experience. In each of these categories, we could ask questions on a spectrum such as 
    • What are the most commonly used features of the IVI platform?
    • What are the user preferences during morning vs. evening commute?
    • What do the vehicle health reports suggest?
    • Is the dealership prepared to handle the customer when they walk in for service? 
  2. Prioritize these questions and Outline the product roadmap. The roadmap lays out the product features, and how they address the particular questions. A well-defined roadmap is an outcome of asking the right questions - learning systems by definition have to learn and hence trying to solve all possible scenarios upfront is not feasible

  3. Measure and Refine the features based on the real time feedback from the field. This is crucial since learning systems have to continuously measure if the output as expected by the user and then refine the features and roadmap. The product manager has to make these decisions based on multiple inputs and develop product in such a way that models are updated 

The above framework helps to define the product and roadmap, the core learning strategy is based on continuous learning, developing algorithms to anticipate and finally validate with the user. 

  1. Learn customer behavior - in the vehicle, outside the vehicle, based on their digital footprint. Learning is a continuous process and it's important to create models that are dynamically updated based on behavior patterns. For instance - commute to work is a pattern that consists of context, location, behavioral attributes, vehicle health - which can be modeled with entities and their attributes. 

  2. Anticipate behavior - once these base models are created, the system starts predicting new scenarios and behavior and perform a fitment analysis as to which is the closest model that it relates and then anticipate what will be the action. These predicted behaviors have to be shown to the user for confirmation and validation

  3. Validate system generated output - it is important to present the system output to the user and get a confirmation or validation if it matches what the user was thinking. It then becomes a closed looped learning system where the system continuously learns and improves the confidence in recommending the next set of actions
The Learn, Anticipate, Validate framework must become integral to product roadmap. Product managers must be critical of ensuring that features have the framework built in and continues to be enriched. Simplicity must be at the core, as these learning systems must not become bloated and lose the essence. The widespread availability, adoption (albeit increasing) of connected services has increased customer expectations - the key is to balance it with well thought of products. When you look around, online/mobile advertising, web applications, smart phones have all been building these learning systems for a while - however few are successful because they provide the users what they want or thinking about when they want it.  Build learning connected services systems that are not annoying! 

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

Please key in the two words you see in the box to validate your identity as an authentic user and reduce spam.

Subscribe to this blog's feed

Follow us on

Blogger Profiles

Infosys on Twitter