Shipping Fast and Small with AI: Breaking Down Effective Product Management Techniques at SKILup Days 2024

Hi, everyone. Uh, thanks for joining me today. Uh, my name is Depo Joti, and I'm excited to bring you a presentation on product management, uh, via which we can drive innovation of customer satisfaction, uh, by injecting ai.

Um, so in this presentation we'll explore how an AI enabled customer centric centric approach to, uh, product management can lead to an innovation and, and high heightened customer satisfaction, uh, all the while, uh, creating, uh, sustainable long-term, um, business value. Um, so let's start with outlining the key topics that we'll cover today. Uh, so first we'll define, uh, what a product management is in contact of, of project management as well.

Uh, I define how it's super critical to the business. Uh, next we'll kind of delve into how AI is sort of empowering, uh, customer centric, uh, product development approach, and how we can tie a lot of things and to end in the entire journey, uh, by making it more efficient, making it more, uh, sustainable by injecting, uh, AI powered tools, um, into the mix. Um, and then finally, we will explore, uh, some strategies and, and mental model, uh, that has helped me as a pm uh, kind of navigate my career across Amazon, Maersk, you know, to kind of deliver sustainable, uh, value, uh, to our customers from a, from a quick and nimble, um, customer centric, uh, approach.

Cool. Uh, so I think, uh, in terms of, uh, what product management is project management, uh, so I would be using both the terms interchangeably, and I think because in the context both will kind of derive the same value out of the presentation as, as my hope would be. Uh, so essentially they sit at the cusp of business and technology, um, empowerment, uh, where, uh, where, uh, we drive, uh, using technology, using, using platform, using project, we di we try to drive significant value, um, of, uh, the program or the product, uh, that we want to kind of launch, uh, for, for our customers or for achieving, uh, the business value.

So, uh, how, and, and how is, how has AI kind of helped you? Right. So AI is super critical here because it has kind of transformed, uh, uh, the entire pm uh, journey by automating a lot of mundane tasks, uh, providing real time data driven insights, uh, that, uh, has helped us, uh, make informed decisions faster and, and real time.

Uh, so that brings us to the role of, uh, so the role is, uh, very evolving, um, uh, due to, uh, you know, ai, uh, integration into market research, into, into customer feedback, uh, into analysis, into sentiment analysis, into product road mapping. So the entire journey, uh, has been, uh, impacted, uh, by the power of ai. And of course, I will kind of detail, uh, how each and every field, um, has been empowered for us, uh, uh, in, in, in our, in our subsequent, uh, uh, section.

So, um, so AI has essentially helped a product manager, um, uh, you know, predict trends, prioritize features, uh, and manage resources efficiently, uh, by kind of an analyzing vast amount of data and gen generating, um, insights out of depth. Uh, so let's then now start breaking the entire, uh, life cycle. Uh, so, uh, the first bit, uh, to the, to the, to the, you know, to the AI powered, uh, customer backward approach to product cycle is to kind of identify what the customers actually need.

Uh, this is where the crux of everything, uh, should start from, uh, correct, because at the end of the day, we have to deliver features, uh, for our customers and ensuring that, that it's the most impactful, um, of, of, of, of all, all the laundry, laundry list of items that, that we could do for our, for our customers. So, um, ai, AI power tools, uh, can create, uh, more personalized, more customer centric, uh, products. For example, um, AI algorithms can analyze large data sets, uh, to, to predict customer preference and behavior leading to a more targeted, uh, product designs.

Uh, so we start with, uh, we start with, um, how do we do that, right? Like, how do we kind of perceive the customer needs? Uh, so there are a lot of aiwas tools, uh, for example, Google and analytics, uh, mix panel, uh, amplitude track, uh, uh, which an amplitude, which kind of tracks the user behavior, um, enabling the PMs to gather a deeper insight into the customer, uh, preferences.

It also helps create something that is very important, uh, which is the user persona, right? Like, for which set of customers and their persona are we kind of delivering the value? So there are a lot of AI driven persona creation tools as HubSpot, uh, you know, where we can make personas, uh, and that's the most accurate, uh, based on, uh, the real world data, based on the target segments, uh, that we want to kind of, uh, focus on just eliminating a lot of guesswork.

So the start of the journey on identifying, uh, a persona is something that is very important, uh, which, uh, which is got a real help from, uh, from the, from the, um, um, aforementioned, um, uh, you know, ai, AI tools. Uh, the, the second aspect is about the entire product development process, right? Like, um, so when we have developed and kind of identified what the personas that we would be targeting with, and then we would come on the features that would help, uh, uh, with our target segments in, in the most impactful, uh, uh, manner, we then need to come into, you know, like, uh, uh, process flows, for example, um, AB testing, uh, code development, uh, or even prototyping on, on, on what we would want to release as a feature.

Uh, tools like GitHub, uh, copilot, uh, can assist the developers write code faster. Um, uh, uh, you know, uh, prototyping tools like Figma, uh, can really help. And of course, with the corresponding AI plugins can really help streamline the entire design process and thus, uh, en enabling and enriching the, the UX or the ui, uh, that, that we will, um, uh, uh, launch, uh, for our, for our feature, uh, for our product.

Uh, so this entire journey from writing a faster code to creating the most impactful prototype of the ux, uh, could all be achieved by injecting, um, uh, you know, various ai, AI technologies, um, into it. Uh, so now we'll move to the timeline where we have, um, uh, where we have to utilize and analyze the customer feedback. So how do we do that?

Uh, so ai, uh, has again, helped us here in a big way on gathering feedback. Uh, there are multitude of tools like, uh, AI enhanced chat bots, uh, surveys that can automatically collect the customer feedback in real time and, and help us kind of draw insights, uh, on, on how the customers has perceived it, um, or what is the features that they were expecting versus what we have developed. Is there a gap?

And essentially, you know, like a lot of surveys on usability score, net promoter score, uh, stuff like that, which we can absolutely use it as, as a, as a part of our continuous improvement cycle. In our, in our net iteration, our, our net, our next, uh, feature launch. Uh, there are, uh, AI powered, uh, surveys, for example, by SurveyMonkey, uh, Zendesk, uh, which, uh, has really impacted on how we kind of gather process and generate insights, uh, from customer feedback, uh, uh, real time.

Uh, there are also an LP tools, right? A natural language processing tools, uh, uh, you know, like, like monkey learn, uh, uh, Google Cloud ai that can also perform sentiment analysis on our customer reviews. Uh, so not only does it help generate the insight, it also tells us how exactly, uh, has the customer perceived our launch or, you know, like, uh, our, our, our product, uh, once they have started to kind of, um, consume it.

Uh, are they happy? Are they sad? Are they angry?

Are they p****d? So we can kind of derive a lot of insight and detail out exactly why, um, are they, uh, and, and why they have exactly felt that. Um, and this could is like, uh, is, is, is is a gold mine of data that we can then use, uh, to kind of make our product or, or, you know, add in into the entire cycle of, uh, continuous, uh, improvement.

Um, the, the last part, this entire feedback tool is where we could leverage AI for a continuous feedback loop. So ai, uh, can absolutely continuously monitor various channels, uh, by automatically flagging any concern. Uh, so it's, so it's not just a static source of data, but it gives us a real time continuous, um, uh, supply of feedback from our customers and on which we can do a lot of predictive analysis, data, data insight analysis, uh, potentially forecast a churn rate.

And, and that kind of suggests improvement on, on, on any feature or any launch, uh, proactively and real time. Uh, so this, uh, the entire gap between when we have launched and the time we have received, the feedback on what we have launched has got reduced, uh, and, you know, like within, within, uh, seconds, now we can, can have the real time feedback of our customers. Uh, so now, uh, we have kind of, uh, uh, delved deep on how, uh, AI tools have really helped us with the product management, with the launch, uh, with, uh, with the customer feedback and customer analysis, and essentially presenting us with a, a very, very good customer centric approach.

Um, on our, on our product product development, uh, we would now kind of, uh, go to the second phase of our, of our presentation where we kind of speak a bit more on, uh, driving innovation and satisfaction, um, of customers, uh, via, via like, let's say an ai, AI driven, um, approach. So here, um, we, uh, where we would kind of understand, uh, the future trends of what, uh, the customer segments are anticipating, uh, the best lever, leveraging the best technologies, um, uh, you know, any, any, uh, any productivity in the customer trend market shifts, looking at the macroeconomic scenarios are all enabled by appropriate AI tools. So, uh, it's not just a human intuition, but also circum, uh, uh, you know, uh, um, um, helped on with a very strong data driven predictive approach by ai, AI tools like, you know, DataRobot, H two ai, where they can analyze and, and kind of simulate various product development scenarios, uh, to drive innovation based on, uh, based on the simulated future trend.

So AI not only helps us to kind of have the most efficient, uh, and most customer centric, uh, approach to product development, but it also helps us, uh, become future proof by, by simulating and generating insights on how the shift in trend or the market segment, um, uh, would look like. Uh, so there are, so there are other ways by which, uh, uh, you know, uh, um, uh, the ai, uh, could help. So in terms of product ideation, uh, predictive, uh, road mapping, where, where there are tools like, um, uh, where product plan or aha, uh, which kind of enable the product manager or the project manager to kind of forecast the future, future demands based on the customer usage, uh, how they are kind of interacting with the system, uh, any historical feedback.

So in our entire roadmap and in, in our entire backlog of what we need to prioritize, we can, again, leverage AI to give us the best backlog or ensure the best trade off that we can do with our limited, uh, resources so that we can, uh, drive and, and deliver, uh, the most impactful, um, output that our customers are, um, expecting. Um, so, uh, AI has also kind of helped us in our road mapping, um, by, uh, by, by, by already suggesting what we should be, uh, focusing on and, and prioritizing. Um, the second bit about this driving innovation is to kind of understand how the customers are satisfied, right?

Because at the end of the day, we can keep on, uh, launching features after features, but if our end customers are not satisfied with it, we will not find adoption, and we can't then del deliver the value that, that, that our feature, our product was expected to kind of, uh, deliver. So, uh, AI has now helped us monitor, uh, real time customer, uh, feedback interactions, um, um, and generate instant feedback on the customer sentiments. Uh, tools like, uh, uh, you know, Cal Bridge, uh, Sprinklr, uh, offers AI powered sentiment, sentiment analysis, um, real time, uh, which kind of help us to take a quick data driven decision to kind of, uh, uh, bridge the gap that our customers are calling out and thus giving us, and thus giving, uh, the customers the best, um, customer satisfaction, uh, that is potentially there to give.

Uh, this is a mark, uh, improvement, um, uh, where previously, uh, the time between our launch to the usage, to the feedback used to easily run, uh, in weeks. So we have now, um, uh, scored an opportunity where we can drive real time, um, uh, enhancements, uh, to like, let's say a, let's say a feature launch. Uh, so with that said, um, we should also be very cognizant that there is a cutting edge innovation that, that AI is bringing.

Uh, but we also need to ensure that it is practical and usable, because at the end of the day, adoption is the key. Uh, and we should be kind of, um, very, uh, pragmatically use, uh, the power of ai, uh, uh, and, and it should always help us take a decision rather than make a decision on our behalf. Um, and at the end of the day, it's, it's our role as PMs to kind of ensure, uh, that we are taking the best decision on behalf of our customer to, to drive, uh, uh, the, the, the most, um, uh, uh, powerful impact, uh, that is there to be delivered.

Um, so essentially by integrating ai, uh, power tunes and techniques into every point of our, of our product management lifecycle, uh, we can then show how the product management is evolving, uh, with, with technology. Um, uh, this really demonstrated how an ai, um, can enhance customer centricity, uh, streamline, uh, feedback loop, uh, drive innovation, uh, and all the while improving on efficiency and satisfaction, um, of our product and feature. Uh, so that brings us to the third phase of our, of our presentation, where I have kind of first taken you through how AI has helped us to create personas to help with the entire road mapping to how AI has then helped us to, to generate better customer insights.

And so help helped us understand with, uh, data enrichment, data insights, sentiment analysis to, to deliver the best feature, uh, with, with, with the real time, uh, efficiency. Uh, so now what, what I wanted to also take you through is some sort of mental model that has helped me as a pm. Uh, I do well in my, in my, in my job to kind of deliver, uh, the most impactful, uh, uh, uh, innovation or to most impactful, uh, feature, uh, for our, for our customers, uh, just generating a long-term sustainable business value.

Uh, so here are some of the product management methodologies that has helped. So one is definitely customer obsession. Uh, I think, to be honest, working at Amazon for a better half of a decade, um, uh, I think the company is really what the company is today, uh, because of its extremely customer obsession, uh, mindset where everything starts from what pain points are we, uh, delivering for our customers, and what is the most impactful now stem.

So at the core of our process, it should not be, uh, you know, like, uh, like, uh, uh, like an insight analysis or, or, or someone's opinion. It should be a hard data driven approach to understanding what our customer really wants and how can we deliver the best on behalf of them. So we have to be very cognizant that we as PMs are the voice of our customers, and we have to live and breathe, um, their life when we are kind of working on, on, on, on the product feature and, and, and the product development, uh, that is the customer obsession and the customer centricity that I wanted to, um, emphasize.

Uh, yet again, uh, the second, uh, decision, uh, you know, or the, or the mental model towards decision is, is one way or two way. Uh, what it really means is that we, we sometimes come into, uh, uh, you know, like, uh, like, uh, uh, analysis paralysis where, hey, uh, we have to take a decision whether we want to go with option A or option B, and we can't take a faster decision probably because we don't have the required data. We don't have the needed confidence to, to move to either e either of the two or, or, or the multitude of the options that we have to kind of proceed.

And that adds up to the slowness of, of how quickly, you know, we, we are, we, we can deliver and impact. It's a velocity, uh, of, of, of a feature, of a feature delivery. So what has helped me here is to kind of go with, uh, one way door or two way door.

So one way door are those decisions where if you go in and if you take it, there is no turning back, uh, probably, you know, like investing on an architecture, uh, where if you have invested on type of an architecture and we have done the entire code base on that, then probably we can't very, very easily go back to like, let's say architecture B uh, then the decisions are one way door, uh, where two kind of drivers our way back is, is a challenge. Uh, we have to give it the needed time. We have to give it the needed thought, ensure that we have got the data, the, the expert opinions, kind of like back it and only then move forward.

Uh, but there are a lot of decisions which are a result of two way door, where if even if we take a decision, uh, between door A and door B and door C, probably if we decide with door CB, we go ahead and then we realize then probably not, you know, like door C was not the best option. We take two steps back, we go to door B. So where the ease of kind of, um, um, uh, going back to the original position by rolling back features, uh, creating enhancements on top, uh, these are some of the decisions where we should not be spending a lot of time on where we can in this the sake of like improving on the velocity of feature delivery.

We can go in and if we, we hit a roadblock, we can always travel us back and then, and choose the second option. So this is something that has really, um, helped, uh, me kind of take faster, better, um, and more accurate decisions, uh, you know, in my, in my during my time as, as, as a pm. Um, the third is kind of like a continuation of the above two, which is deliver small and fast.

Um, what, what, why, why I have always emphasized on delivering, uh, uh, small and fast is because it helps in two very important ways. One is it helps with the adoption of our feature. So if we wait to kind of deliver the entire end-to-end feature directly at one go for our customers, the adoption of us very big feature also kind of works against the cognitive, uh, uh, you know, skills of, of, of, of our customers.

Because at the end of the day, human mind wants to take path of least resistance. So if we kind of help them with smaller chunks of what we want them to experience and adopt, uh, it is best to kind of deliver the big item into small chunks faster and in phases. Um, the second aspect that's also very critical here is to understand the feedback loop, uh, rather than waiting for the entire product or the project delivery.

Um, and then we work on the feedback of the customer. Uh, we would always be, uh, working on a huge time between the launch, uh, between the feedback between getting it coded and then relaunching, versus if the deliveries are small, we can have a faster, um, uh, uh, faster deployment, faster adoption, faster feedback, and then we can already incorporate those feedbacks, those critical feedbacks into, into the second phase or the third phase of our, of our subsequent delivery of our, of our roadmap or our, or our, in terms of our phase wise approach. So, so apart from the entire ai, uh, tools that we have covered, uh, I have also given you a sense on what mental model, uh, we should always have.

That would be, uh, you know, and, and, and, and I think if you can have a, a, a very strong pm mental model along with the aid of the ai, that, that, that, that we will be empowering our, our processes and our platform with, uh, I'm sure we will be able to deliver the best, uh, our value, uh, that is to be delivered, uh, uh, for our, for our customer. And that, uh, brings us to the end of our presentation and, and hope, uh, you guys have been like, thank you so much. Bye-Bye.