How to Use AI in Product by Intel Senior Product Manager

In the era of AI and IoT, it is incumbent for companies to keep pace and innovate constantly. 

When building an IoT product, it is essential to think of an end-to-end solution. When doing so for an AI product, having the right data set and Machine Learning models are crucial. In this article, Charunethran Govindarajan, Product Manager at Intel, describes the importance of being multidisciplinary and explains ways in which any company can successfully integrate AI/IoT into their products.

Multidisciplinary PM

  • Product Management is multidisciplinary.
  • Being Multidisciplinary doesn’t essentially mean Multitasking.
  • Nobody is expecting you to single-handedly build the product and make it a success.
  • Haters are going to hate no matter the amount of effort you put.
  • Essential tips:
    • Don’t act or try to be an expert in all disciplines.
    • Programming, Biz Strategy, Roadmap Building, Dogfooding, UX Research, UI Design, Market Research, Tech Writing, Customer Support, and Architecture are few of the disciplines. You don’t have to be an expert in any of these.
    • Knowing sufficient information about these domains is enough since you’re still responsible for the entire product development lifecycle.

You should be constantly aware of the changes happening in all sub-domains and thereby, the butterfly effects on the product backlog. The 5 goals of PM are technical knowledge, customer/user focus, design, leadership skills, and S&M.

Customer vs Industry

  • Neither the customer nor you or the company are right. Understanding this plays a very important role in determining the success of a product.
  • Empathize with the customer and read in-between the lines.
  • There are two different scenarios when the customer and industry are involved: 
    • Case 1: Customer and Industry are aligned
      • Evaluate what the value add is to the customer.
      • Be aware of the competitors in the market.
    • Case 2: Customer and Industry are not aligned
      • This provides a scope for innovation, so think innovation.
      • Reassess your Unique Selling Proposition (USP).

Product Management for the Internet of Things (IoT) and AI

  • The Expectation vs Reality for IoT is pretty hyped. For example, the IBM investor briefing in 2012 estimated a trillion interconnected devices by 2015.
  • When dealing with an IoT product, think of an end-to-end solution which covers right from the device/gateway to data management and messaging.
  • Most of the customers AI/IoT requirements are quite generic and usually include keywords such as Smart, Machine Learning, Robot, Deep Learning, and Pattern Recognition to name a few. Spend more time with the customers and understand their actual requirements.
  • The Secret Recipe for a successful IoT/AI product is data accumulation → product launch → customer acquisition and repeating this in a cycle.

ai-product-management-iot

  • Having a strategic data set and knowing the right Machine Learning models to use, go a long way in building a successful AI product.
  • If there is a change in the Product Requirement Document (PRD), the below aspects need to be considered:
    • Strategic data acquisition.
    • Unified data warehouse.
    • Spotting pervasive automation.
    • Keeping pace with AI.

Meet Charunethran G.

Charunethran Govindarajan

 

Charunethran Govindarajan has a Masters in Electrical Engineering from Stanford and is currently working as a Senior Product Manager at Intel. He joined Intel in 2013 and was one of the first 3 engineers to work on Interconnect Consortium. Currently, he focuses on AI and IoT and believes in being a jack of all trades than a master of one.

Enjoyed the article? You may like this too: