Creating intelligent experiences: The role of an AI Product Manager

With the expanding field of data analytics, the integration of Artificial Intelligence (AI) has become a powerful tool for data-driven decision-making in product management. In this article, we share a thought process to enable a product manager to leverage AI for product thinking to build AI driven products that solve customer problems.

8 min read
Share on

Exploring how AI empowers various industries and its current applications is a fascinating journey. Until now, AI product management primarily resided with platform Product Managers who utilized AI to build core features. However, this paradigm is rapidly changing as Product Managers across different spheres and verticals embrace Artificial Intelligence.

Although we are making advancements in Artificial Intelligence, it’s important to acknowledge that machines still face challenges such as accuracy, truth, and bias. We’ll delve into these aspects further later on.

Is AI relevant for my work?

As a practice of building products, we always begin with ‘why’? Why do we need to solve a problem? Why is it a problem to be solved now etc? And we then decide the solution and possible use of additional technologies. But with the colossal upcoming of generative AI technologies, Product managers need to add an additional arsenal to their toolbox. Product managers should now think of how AI can build and break their use cases. If you have concerns about being potentially obsolee as a product manager, we propose a valuable framework to determine the necessity of embracing AI:

  1. Assess disruption potential: Evaluate whether AI could disrupt your organization’s core bottom line. Analyze if competitors or market trends indicate a need to adopt AI to remain competitive. Learn from examples such as Chegg who got disrupted and other such examples to understand the impact of AI adoption on similar industries.
  2. Evaluate applicability: Determine if AI or AI tools can effectively enhance or automate your product area. Identify specific tasks or processes that could benefit from AI integration. This evaluation helps decide whether to build AI capabilities in-house or seek external solutions.
  3. Transition or expand: Consider whether it’s time to transition to other product areas where AI can add value or expand your current product offering by incorporating AI features. Explore opportunities to leverage AI assistants to streamline operational work, allowing you to focus on strategic product thinking.
  4. Embrace technology: If you have been solely focused on building user experiences without embracing technology, it’s time to shift your mindset. Embrace AI as a tool to enhance user experiences and create innovative solutions.
  5. Enhance skills: Upskill yourself in customer empathy, creative thinking, and customer discovery to gain influence over your ideas and make informed decisions about AI implementation. Understanding the potential pitfalls and challenges of introducing AI in your product is crucial for successful adoption.


The different problems that AI could solve for us (Source: Google Images)

Whatever the outcome of the above framework, it is time for PMs to up their game in AI. And one possible path is to learn and cultivate a culture of AI in their professional growth and as well as of the organization.

Junior Product Manager:

a. Enroll in online AI courses or certifications to build a solid foundation in AI concepts

b. Collaborate with AI experts to seek opportunities to work with data scientists

c. Stay updated with AI trends by following relevant communities

Mid-Level Product Manager:

a. Dive deeper into AI algorithms and frameworks to expand your understanding of AI algorithms and frameworks

b. Incorporate AI-driven user research methods into your product development process

c. Take initiatives to lead projects that involve AI integration

Senior Product Manager:

a. Mentor and Share your AI knowledge and experience with junior product managers.

b. Evaluate AI vendors and technologies for their capabilities, reliability, and alignment with your product strategy

c. Collaborate with AI research teams to get access to cutting-edge AI research, emerging trends, and potential opportunities for innovation.

Product Lead/Manager:

a. Foster a data-driven culture across your product team and organization

b. Develop AI roadmaps and strategies aligned with your product vision and goals.

c. Build cross-functional AI teams to strengthen your product team’s AI capabilities and enable seamless AI integration.

Director/VP of Product:

a. Shape AI product vision for long-term vision of AI-driven products within your organization.

b. Influence AI research and development by establishing partnerships with AI research organizations, startups, or academia.

c. Develop AI thought leadership.

Let us present some real-life use cases if this new technology still seems a mumbo jumbo. The versatility of AI can be seen from the smallest use cases to most sophisticated machines impacting life care and National Security.

Until recent months, that was the most used and novel idea in AI. But the recent advancements in Natural Language processing have pushed a bar notch higher. Generative AI, or generative artificial intelligence, is a form of machine learning that is able to produce text, video, images, and other types of content. This has made the most impactful web search,

Real-life applications of AI to boost productivity at business (Source: Google Images)

Case studies

Case I:

Here is a real-life application in the content streaming industry where AI could help achieve hyper-personalized content leading to increased viewership and revenue.

A content streaming company aims to provide the best possible viewer experience by offering personalized content recommendations that captivate and engage their subscribers. The company plans to build a solution to leverage AI and enhance their content recommendation capabilities to deliver a more tailored and satisfying viewer experience.

The typical content recommendation system primarily relied on manual curation and simple user preferences. Recommendations were based on limited factors such as genres, previous viewing history, and popularity. This approach had several limitations. It struggled to understand nuanced viewer preferences, failed to capture evolving tastes, and often resulted in generic recommendations. As a result, viewers faced difficulties in discovering new content, and their overall experience was suboptimal.

By implementing an AI-powered content recommendation tool, we revolutionized our approach to personalization and viewer experience. Leveraging machine learning algorithms, we could now analyze vast amounts of viewer data, including past viewing behavior, contextual information, viewer demographics, and real-time interactions. Our AI-powered recommendation tool utilizes collaborative filtering, content-based filtering, and deep learning techniques to identify patterns, similarities, and correlations between viewers and content. This allows us to recommend content that aligns with each viewer’s unique interests, preferences, and viewing habits. Additionally, the tool continuously learns and adapts based on viewer feedback and behaviors, ensuring the recommendations become increasingly accurate and relevant over time.

Implementing the AI-powered content recommendation tool brought significant improvements to the viewer experience at the company, which was measured by different metrics as follows:

  • A X% increase in Click Through Rate, indicating improved engagement with recommended content.
  • A Y% increase in average viewing time, demonstrating higher viewer satisfaction and interest in recommended titles.
  • Positive user feedback, with a Z% increase in user satisfaction scores.
  • A K% improvement in viewer retention rates, highlighting the impact of personalized recommendations on subscriber loyalty.

Case II:

Let us present an additional case study that highlights our strategic approach and utilization of AI. While engaged in a project that required large-scale implementation of ML models, we encountered a challenge related to the loss of significant data signals, which directly impacted the company’s profitability. Considering the immense volume of data we processed on a daily basis, measured in terabytes, the loss of these valuable signals posed a critical issue in decision-making.

Our initial step was to assess the consequences of the missing signals. We trained our models using the available signals and conducted an offline A/B test to evaluate their performance. Through continuous monitoring and analysis of the model’s output, we made iterative improvements based on our findings. Recognizing the need to compensate for the lost signals, I undertook the responsibility, wearing multiple hats as a product manager, to explore potential solutions such as establishing new data sources or integrating with existing ones.

Although securing new data initially proved to be a challenge, we eventually obtained a stream of signals that significantly enhanced our model training process. We maintained a continuous cycle of training and testing until we achieved satisfactory levels of precision and recall. These metrics, analogous to the combination of peanut butter and jam, needed to be finely tuned to comprehensively evaluate the model’s performance.

Upon reaching a point of confidence in our A/B test evaluation and acceptable performance thresholds, we decided to deploy the models in a production environment, initially targeting a select group of test customers. Another crucial consideration throughout this process was collaborating with an individual well-versed in data privacy, ensuring compliance and addressing any related concerns. Although the beta rollout demonstrated the benefits, our customers expressed reservations about the “black box” nature of AI usage. Consequently, collaborating with the product marketing team to craft a compelling narrative that resonated with non-technical customers and aligned with the sales strategy became paramount in delivering a successful product. This underscores an essential product management principle for future AI endeavors.

Summary

As we wind up the ideas of being an AI product manager, just realize that we are living in the most interesting times of our era of AI revolution. If we do not want to miss the boat and advance our careers and make meaningful impacts, now is the time to adapt to new learnings.