In a significant real-world learning experience for the MTech students of Product Design and Management, IIITH recently hosted a conclave of tech leaders, heads of product, domain product companies, AI researchers, and product managers of various corporates, and startups to deliberate upon the impact of AI on product management and product innovation.
Cognisant of the fact that AI, particularly generative AI holds immense potential to disrupt the way product development has traditionally been approached, a round table conference was organised by IIITH to deliberate on AI and product management. The panellists discussed not just the obvious impact of AI on product development but also deliberated on the very practice of product management as we know it.
AI To Complement Not Compete With PMs
Kickstarting the session was a discussion that focused on how AI is reshaping product management. The experts spoke at length on how domain-specific AI tools and purpose-specific AI tools are the need of the hour especially when precision, expertise and speedy results are required. As a “partner” to the product manager, AI helps in a number of ways such as processing large amounts of data, allowing PMs to draw valuable insights and assist in decision-making, being a part of the product lifecycle through A/B testing and usability studies, analysing data to understand which features ought to be prioritised and so on. While AI’s potential is undeniable in elevating user-centric product innovation by enhancing personalised experiences, or providing first-level support via conversational bots or HR management systems, it is imperative to keep in mind that it’s ultimately the PM and not AI who is the user’s advocate. For instance, ChatGPT can only supplement and support the PMs and never actually replace them because there will always be human elements in the process of product development such as an innate understanding of the market and users’ that are irreplaceable.
The Buck Stops With PMs
The speakers deliberated upon how product management without the AI element seems unthinkable of today. They were of the opinion that while PMs need not be AI developers, understanding the possibilities of AI and staying abreast with emerging tech is crucial for their relevance. For example, It’s not enough just knowing about the advantages of AI but also about its limitations. PMs can add value by incorporating user-based inputs to make the systems smarter. When AI is used to arrive at a recommendation, again the PMs are expected to evaluate the output and understand how it was arrived at, thus putting the onus of accountability upon themselves. This not only ensures AI reliability but also builds trust among stakeholders.
AI in various domains
AI is being used in industry-specific data flow analysis where organisations can leverage advanced algorithms and machine learning models to extract valuable insights. For instance in supply chain management, Amazon is harnessing AI to achieve real-time visibility into the locations of its equipment, revolutionising operations, enhancing efficiency and auto-correcting errors seamlessly. Similarly in healthcare, AI is now routinely used in identifying diabetic retinopathy, and melanoma. However, the diagnostic process in healthcare is much more complex involving intricate decision-making that goes beyond simple symptom-matching. A blind reliance on GPT is therefore not recommended in this scenario. Much like Unified Modeling Language tools provide a conceptual framework for software design, AI too can serve as a mental framework for PMs. But AI tools can’t replace a software developer in its entirety. In the case of customer support too, there are risks involved in integrating GenAI for processes such as refunds. Using AI as an assistive tool alongside human support is more advisable than as a sole decision-maker.
Limitations of AI in Product Management
AI models particularly LLMs are prone to biases emphasising the role of humans in training AI models. PMs may fall prey to FOMO (the fear of missing out) and rush to adopt AI solutions at every step of the product management process which is not advisable. Plus there are many privacy and data confidentiality concerns in the usage of AI tools. For example, some companies have banned the use of ChatGPT due to the potential IP issues and risk of proprietary data leakage. There are other ethical concerns too surrounding critical decisions made through AI systems like in healthcare where sometimes recommendations for treatment are based on the race of the person. In addition to this, ‘force-feeding’ or supplying of extensive data or specific inputs for training of AI models in order to facilitate a specific outcome poses another ethical risk. Just like humans are faced with ethical choices they have to make, AI systems too have to deal with complex ethical dilemmas emphasising the human element in decision making.
Conclusions
With the increasing role of AI in various domains, AI in product management too holds great promise, from aiding data analysis, to enhancing productivity, there’s a lot it can do to simplify the process. However, there is a need to strike the right balance while using it as a supportive tool and retaining the essential human elements required in the process. It was unanimously agreed upon that AI education for PMs is critical and in fact usage of AI tools such as Canva and Amber (for making roadmaps and user personas), Taskedo (for customer tickets and feedback), Webex (for summarising meetings), Teams (for transcription and feedback) many more are extremely crucial for PMs in the adoption and navigation of the tech landscape.