How AI Is Changing The Product Management Job Description by Mayukh Bhaowal "Product people - Product managers, product designers, UX designers, UX researchers, Business analysts, developers, makers & entrepreneurs February 02 2021 True Ai, Artificial Intelligence, Bias, Ethics, mtp engage manchester, Mind the Product Mind the Product Ltd 408 Mayukh Bhaowal speaks at MTP Engage Manchester Product Management 1.632

How AI Is Changing The Product Management Job Description by Mayukh Bhaowal


In this MTP Engage Manchester talk, Mayukh Bhaowal, Director of Product Management at Salesforce Einstein, takes us through how product managers must adjust in the era of artificial intelligence and what they must do to build successful AI products.

Mayukh outlines five new skillsets that product managers must acquire.

  1. Problem Mapping
  2. Data is the New UI
  3. Acceptance Criteria
  4. Explainability, Ethics and Bias
  5. Scaling from Research to Production

Watch the video to see Mayukh’s talk in full. Or read on for an overview of his key points.

Problem Mapping

In the world of artificial intelligence, the number of stakeholders a product manager must be accountable has increased and now includes data scientists and data engineers as a separate entity and not as just part of the technical development team. Product managers must be able to map to a problem and identify the right metrics to asses to reduce case resolution time. Some of the traditional methods are not always application and AI and multi-classification may be required to solve some problems. Product managers must evaluate traditional methods first and then use AI if it proves itself to be the best way to solve the problem.

Data Is the New UI

How clean or messy is your data? Data is an important piece of the puzzle in product management today. Are there prior examples available to teach the machine learning models? This is an example of something new that is absent in traditional methods.

Acceptance Criteria

Product managers need to work with data scientists to determine the right metrics to optimize. Mayukh uses the example of a data hack on a bank to illustrate what metrics need to be optimized. The machine algorithm needs to be trained to pick the lesser of two evils in some cases.

Explainability, Ethics, and Bias

Product managers need to address bias in their AI tools. In different industries, certain factors might matter more. For example, explainability beats accuracy in an industry like healthcare where regulations matter. GDPR, for example, requires you to classify.

Scaling From Research to Production

Understand which products to develop with foresight. AI was in search and ads before but the use of the technology will continue to grow exponentially in the coming years. The key takeaway from this talk is that product managers need to upskill to cross the AI product chasm. There is a barrier against entry where they need to provide specs that add value to engineers.