- Any Questions?
- 080-4710-6006
- [email protected]
By Dhruv Rastogi – VP & Head of Data Science at IKS Health
AI product management involves solving customer problems using data enabled by artificial intelligence and machine learning. In the past, data was mostly used for analytics, where once the product is launched, you would run some data reports, and analytics and then find some insights based on that information. Based on these insights, you would go back and program something and create products.
Artificial intelligence has been a backbone in shaping product management in recent years. It has been responsible for shaping products from standalone black box solutions to integral, end-to-end systems.
Firstly, product managers need to understand the role of AI and ML in product management. We look at it from two different perspectives- from a feature to a core dimension, and from a standalone black box to an end-to-end solution dimension.
If we think of certain products, for example, speech-to-text products, or products that help in image recognition, object detection, or language enhancement, these are standalone black box products. They typically don’t need anything else, and they are the core as well. But if we think of something else like a self-driving car or a surveillance drone, these are solutions in which AI is at the core, and without that the product doesn’t have any meaning. There is no meaning in a self-driving car if it cannot detect objects around and make the right decisions.
Another dimension that you see is areas in which AI acts as a feature. This includes examples like Amazon Prime, Netflix, or Hotstar. All have a recommendation system. But the recommendation is a feature of these products. So these products can work stand-alone as well, essentially a repository for content but how recommendations help us is they help improve the user experience and help them in being on the platform for a longer time. But the core system even works even if there isn’t a recommendation system.
Many times humans are required to make AI usable. So AI is not a magic box. Most of the AI systems that we use today, whether it is speech to text, anything in language, or images, all have been built with a lot of input, intelligence, and support by humans themselves. So how do people help in building AI systems?
1. First, they help in handling the edge cases. Many times a machine can’t figure out what to do in such a scenario. This is where the human’s judgment supersedes. The reason why machines don’t work in such cases is because such cases did not have enough past training data. If they didn’t have enough past training data to handle a case they won’t be well equipped to handle that case.
2. The second is, building the training data sets themselves. Hence, if machines have to learn somebody has to build a training data set for the same. That is where humans help in building these data sets.
3. Generating heuristics is another important aspect. If there are no specific solutions to a product problem or the time required to find one is too high, a heuristic function may be used to solve the problem. The aim here is to find a quicker or more approximate answer, even if it is not ideal.
4. Quantifying the sample accuracy is also an important role. For example, even when machines are trained, someone has to say that the machine is currently operating at a particular accuracy level.
We can say that while AI and ML help hugely in product innovation, human collaboration remains indispensable in refining their capabilities. The relationship between human intelligence and AI continues to drive success and advancements across diverse product development processes.
AI in product management utilizes artificial intelligence, deep learning, or machine learning to solve product problems.
It is highly unlikely that AI will fully replace product management. AI is used to assist product managers and enhance aspects of product management, like data analytics and customer support. But it cannot completely replace a product manager.
In traditional product management, product behavior is usually binary and predetermined. AI product management deals with probabilistic outcomes.
Challenges associated with AI product management include working with more stakeholders, ambiguity of outcomes, difficulty in explaining the rationale behind the outcomes, addressing fairness and bias concerns, adapting to new infrastructure and tools, and selecting the right problems to solve with AI.
About the Author:
Dhruv Rastogi – VP & Head of Data Science at IKS Health
