Bold Moves and AI Trends: Neha Kalani on Navigating AI Product Management Challenges

Step into the fascinating world of AI product management as we sit down with Neha Kalani , an unconventional product manager who's carved her own path to become a Product Manager at Pictory

In this engaging conversation, Neha takes us on an eye-opening journey through her experiences, pulling back the curtain to reveal the dynamic interplay between artificial intelligence and innovative product strategy. With AI-powered products taking center stage in today's tech landscape, Neha's insights offer a truly unique and invaluable perspective straight from the trenches.?

From her winding transition into the world of product management, to the hurdles and high points she's navigated along the way, Neha's story is a testament to the limitless possibilities within this rapidly emerging field. Join us as we dive headfirst into the complexities and nuances of building AI-driven products, exploring the visionary approaches, cutting-edge tools, and key metrics that define success.

Hey Neha, great to be chatting with you! Since we've had our sync call and I want you to steer the direction of this call as that’s my style basically, let's kick things off from the very beginning, shall we? Whether it's your school days, college adventures, or diving straight into work stuff – where do you reckon your journey really started and how's it shaping up for you now?

So you know, after graduating college was when life really started getting interesting for me. I was a mechanical engineering grad with absolutely no background in computer science whatsoever. Product management was the farthest thing from my mind during those college days.?

I got placed right out of college - one of the first people from my batch to land a job at Hero MotoCorp . It was a solid core engineering role. But here's the kicker - for that job, they posted me all the way out to Kerala! A place I knew nothing about, didn't have any friends or connections, and had never even visited before in my life.

My role was territory service manager, which meant I got to travel all around different parts of Kerala - Calicut, Trivandrum, Kochi, you name it - overseeing the service centers in each of those cities. Talk about being thrown into the deep end as a newbie!

Sorry to cut you in between but how did you like Kerala? I am a Malayali by the way!

Oh I loved Kerala! You know, back when I first arrived in Kerala, it was a whole new experience for me.?

One big learning experience from my work there was communicating at the service centers. As the territory manager, I had to interact with technicians, receptionists, customers - people who often didn't speak much English or share my background. It could be challenging to connect and understand each other when we don't know each other's language. But it really pushed me to find ways to communicate clearly across those barriers. After all, a big part of my role was driving up the net promoter scores at each service center I visited.

So I made a habit of spending full days at the centers, from open to close, talking to every customer who came in to drop off or pick up their bikes. With a translator by my side, I'd try to dig into each customer's issues and get to the root of their problems. And I had to be super conscious of whether key nuances were getting lost in translation too. Over time I got really good at navigating those communication gaps.

That whole experience of engaging so deeply with customers and relentlessly listening to their voices - it ended up being hugely formative for me as a product manager today. Because I saw first-hand how vital it is to keep users at the core of everything you do. No matter what role you're in, you're ultimately building something for the users. That's one of the biggest lessons that experience reinforced for me.

After that experience in Kerala, I decided I really wanted to pick up some coding skills. That's when I joined a company called ZS . I built my own tools, did a ton of work in Python and Excel, really diving into understanding how software actually gets made - how the code is written, the logic flows, all of that. By then I had this great mix of coding knowledge plus all my prior business experience engaging customers and solving real operational problems.

It was around that time that one of my friends reached out and said "Neha, I think you'd be perfect for a product management role. I have an opening - want to come check it out?" When I looked at the job description, I had no clue what product management even entailed!

But as I read more about it and talked to people who raved about the role, I realized it was this beautiful convergence of the exact areas I'd been working in - coding and technical understanding on one side, business acumen and customer focus on the other. I thought "This is it, this is the path for me." And when I started in that first PM job, I can 100% vouch for how perfectly all those prior experiences came together to prepare me.

I loved the product so much that I knew I wanted to invest myself fully in making it better. I just don't think I could pour my heart and effort into something I didn't genuinely believe in. So joining a company building a tool that I myself would be excited to use as an end user - that was key for me. And once I came on board here, well, the rest is history as they say!

Neha, your journey from a mechanical background to landing a product manager role seems quite extraordinary, especially considering the unconventional path. You mentioned starting with a struggle in social skills, but now, in product management, it's all about being the communication maestro – talking to teams, customers, and stakeholders. How did you go about leveling up your communication and networking game for the product management gig over the years??

Yeah, I know it was pretty unconventional but it was exciting too. I often share with my juniors at the office that at the core, product management is all about communication. It's about extracting the right information from the right people at the right time. No coding, no designing – just me, my keyboard, a Slack , and maybe a Zoom call or Notion thrown into the mix.

To answer your question, if it weren't for my time at Hero, I wouldn't be as socially active with customers as I am now. My boss threw me into the deep end, setting targets to talk to 20 users a day, which eventually escalated to 50. Speaking to people in a completely new place and even those who didn't speak my language made me realize my knack for understanding things between the lines. It's that skill, spotting the pain points between what users say, that's crucial for a product manager.

Later, when I took on the entrepreneur-in-residence role, they coached us on design thinking. I vividly remember a piece of advice from a Harvard professor – 'Don't believe everything the customer says, believe everything they're not telling you but feeling.' It was an epiphany back then. It taught me that a simple 'yes' or 'no' doesn't cut it. This mindset guided me through my role at ZS, dealing with clients in the healthcare industry in the US.

In ZS, although my role didn't demand a lot of cross-functional collaboration, I actively sought opportunities to understand and assist other teams, honing my cross-functional leadership skills. Engaging with engineers and designers much older than me has been a learning curve, but building trust in my intuition and data is key.

Dealing with diverse teams every day has been valuable in my product management journey. I've learned to communicate and collaborate with people of any age, understanding that users don't come with an age limit.

Absolutely, Neha. I had a different set of things to ask you but your insights have sparked my curiosity. Especially, when you emphasized the importance of asking users the right questions. What’s your approach in defining what constitutes the 'right question'? How do you go about framing questions with the mindset of extracting valuable information?

For me, nailing the art of asking the right question involves starting with a deliberately vague one. Instead of diving into specifics, I throw out a broad one first. For example, rather than asking if someone likes a particular feature, I might kick off with something more open-ended. Let them speak, share their thoughts without steering them in a specific direction. It's like what's happening in our conversation right now – your questions are shaped by my responses.

In user interactions, I often initiate with a question like, 'Why do you even use my product?' This encourages them to delve into a monologue for the first ten minutes, expressing their thoughts freely. During this time, I refrain from interrupting, keeping my camera on and nodding to convey active listening. The goal is to create a space where they feel compelled to fill the silence. In doing so, they naturally highlight aspects they like or dislike. It's in those moments that I pick up cues for my next set of questions.

That’s a great strategy Neha. Diving deep along with them is so essential, I agree! I'm intrigued by how your team dynamics and work style have evolved over time. Moving from companies like ZS or Hero to your current role at PictoryAI - how has the shift impacted your collaboration with AI developers and your overall approach to understanding the tech stack? Essentially, how has the transition to an AI-centric product influenced your team dynamics and work style?

When I transitioned to my current role at PictoryAI, I took on the challenge of being the only product manager, a role I continue to hold. Managing an AI tool meant not just diving into the AI aspects but overseeing the entire product, including regular front-end and back-end issues. This required a dual learning curve – understanding both conventional tech terminology and the intricacies of AI.

My approach was hands-on learning by actively listening to my developers. I never hesitated to ask questions, even if they seemed basic. If a jargon was unfamiliar, I'd Google it on the spot or directly ask my team for clarification. What I found was that most people appreciate questions and explanations. It wasn't just about learning; it uncovered critical issues that might not have surfaced if I hadn't asked the right questions.

Asking questions became a powerful tool, not only for my learning but for uncovering essential problems and ensuring a deeper understanding of the product.

Understood! I can see how important “ asking questions” are in your journey! I'm actually curious to hear your perspective on the common notion that product managers don't necessarily need to be highly technical, just have a basic understanding. However, considering the unique landscape of an AI product, which falls into an emerging and evolving domain, how crucial do you believe it is for a product manager to grasp the intricacies of the tech stack? Especially in these early years where understanding an evolving domain can be particularly challenging?

Without a doubt, Vismaya, I'd emphasize that technical knowledge is not just important; it goes beyond the basics, especially in the realm of AI products. Coming from a non-tech background myself, I've learned that merely understanding broad concepts like technology, tech stack, and AI won't cut it. As a product manager, you're not expected to code, but you are expected to help your development team make informed decisions.

In our evolving AI landscape, even at our company, we're continuously learning. For instance, when consulting with outside experts or deciding on models like GPT-3 versus GPT-4, you need to dive into the nitty-gritty. Knowing the nuances, like response rates, updates in different models, and their relevance to our business, is crucial. It's not about memorizing terms; it's about understanding the intricacies.

This extends beyond AI; even on a general tech product, understanding APIs, handling errors, and estimating the feasibility of a solution for the engineering team are key. You need to be able to ask why one technology is chosen over another and understand the potential impact. I found 'Tech Simplified' by Deepak Singh to be a helpful read when I started here, underlining the importance of delving into technical details to thrive as a product manager.

Tech Simplified is indeed a fantastic book and I highly recommend it to anyone looking to delve into the technical aspects. Coming from a background of development myself, I've always found that having a technical understanding allows me to estimate work better and provide a more comprehensive brief to the team.

Having worked with both technical and non-technical product managers, I've noticed that my ability to comprehend why certain things are or aren't feasible from a technical standpoint gives me an edge. It's not about comparison but rather understanding why a solution might not be technically viable.

I'm aware that in product development, coding is just one piece of the puzzle, and for AI products, the more intricate decisions often revolve around choosing the right model and training the data. As the founding PM of PictoryAI, how did you navigate the delicate trade-off decisions, such as selecting the model, determining the data inputs, and making crucial decisions on what data to rely on?

Very, Vismaya. The decision-making process primarily revolved around our business needs. Considering that we're a startup with limited funds, the choice of models for PictoryAI was deeply tied to pragmatic considerations. We assessed factors such as the number of users, response time requirements, desired accuracy levels for specific tasks like video creation, and the overall skill set of our engineers – both in-house and those we hired externally.

The evaluation process involved extensive testing, spanning weeks if not months, where we rigorously tested the same models with different parameters, tweaking training methodologies. While I may not delve into the intricacies of the technical details, I keenly observed and comprehended the end results. Collaborating with my founders and the team, decisions were then made based on a holistic consideration of cost-effectiveness, accuracy needs, user volume, and the technical expertise available to us. These were the four parameters we initially had.

I believe these metrics provide a solid foundation, especially when starting on model training or deciding on the most suitable model for the task at hand. On a relatable note, Have you integrated any AI tool stack into your product strategy and roadmapping phases? Given your experience with Picture AI, I'm interested to know if AI plays a role in shaping the overall product strategy and roadmaps.

I'm a daily user of GPT, Notion AI, and our own tool, leveraging AI extensively in our product strategy. Specifically, in the case of churn analysis, I've shifted away from traditional data analysis tools. Our cancellation flow involves gathering detailed feedback from users, often in the form of text, and, being non-technical, I make use of Notion AI a lot to promptly feed all this text data into Notion. While the volume of data initially posed a challenge, breaking it down solved the issue. Notion AI did an exceptional job, providing insights into the top reasons for churn. With this information, I can strategically prioritize upcoming sprints. For instance, if a specific feature is causing churn, addressing it becomes my top priority in the next sprint. This is just one of many ways AI plays a crucial role in my day-to-day operations.

If you don’t mind, I have a tool suggestion for you that might ease the process of compiling customer feedback. It's called Zeda.io , and what's unique about it is that it incorporates emotional analysis in feedback. Considering the challenges you might face with tools like Mixpanel, which I'm familiar with, I thought Zeda.io could be a smoother option for a non-tech role like yours. It might save you the hassle of manual data input and analysis. Just a suggestion to explore!?

Neha, in the world of AI-generated content, have you encountered any resistance from the larger audience regarding the credibility of the content? How do you navigate and address concerns about the authenticity of AI-generated outputs from your product?

Speaking of our approach at PictoryAI, we're fortunate to have a great marketing manager who also serves as our community manager. When users express concerns about content promotion on platforms like YouTube, our marketing manager conducts webinars emphasizing the importance of caution when using AI in content creation. Specifically, if users heavily rely on visual stock visuals, it's advisable to refrain from using AI voices and instead opt for their own voice.

From a product perspective, we've equipped Picture AI with a feature we call 'Auto Synchronization of Voice.' Users can record their voice using a quality microphone, upload it, and with a simple click, apply it to the entire project based on the text. This feature eliminates the need for manual adjustments that a traditional video editor might require, offering a seamless and efficient solution.

We believe in a balanced approach, leveraging both AI capabilities and the unique voice of content creators to ensure their content stands out in the cluttered landscape.

Neha, you're spot-on! Content creators should treat AI as a tool, not a crutch. It's about stacking AI on top of their talent, not replacing it. This balance ensures their content stands out in a crowded space, blending technological efficiency with their unique voice. The key is using AI strategically, as an ally to enhance creativity, not overshadow it. Given that platforms like YouTube and other social media apps have introduced features like YouTube Create, providing tools for script filling and video editing, how does PictoryAI perceive its competitors in this evolving landscape?

Yes, while YouTube Create isn't a direct competitor, the landscape is highly competitive, resembling a red ocean with numerous tools emerging constantly. We view this not as a threat but as an indication of the growing relevance of AI-powered content creation. Our focus is on continuous improvement, staying attuned to user needs, and ensuring our features are not just cutting-edge but also user-friendly. As for sustaining our position, our strategy revolves around a thoughtful balance of product modification, user retention, and staying adaptable in this dynamic space. To navigate this challenge, we emphasize two crucial factors. First, we prioritize speed, ensuring quick responses to market dynamics. Second, and equally important, we've defined our niche. We specialize in script-to-video transformation, excelling in a specific segment, such as healthcare advertisements, where talk visuals play a significant role. Rather than chasing the entire content creator market, we focus on identifying our strengths, catering to a specific audience, and moving swiftly to maintain our competitive edge in this dynamic space.?

Given the challenges of sustaining and growing a startup, especially as it matures beyond the initial stages, how do you envision selecting and cultivating your niche user base to ensure product retention and development? What strategies do you plan to employ as PictoryAI transitions into this phase of its journey?

Absolutely, it's a daily challenge for us to navigate the competitive landscape and focus on retaining our user base. Retention, being a lagging metric, demands a comprehensive effort from the entire team to ensure users stay with us. Our strategy is centered around a thorough understanding of our target audience, particularly our Ideal Customer Profile (ICP) users.

To address retention challenges, we meticulously analyze why our ICP users might be leaving, aiming to pinpoint the underlying issues. This process involves collaborative efforts from various teams such as product, design, marketing, sales, and technology. Once these teams unite, we work together to define solutions tailored specifically to our ICP users. As the leader of the product team, my role is pivotal in planning and executing these solutions. Furthermore, aligning the messaging of our marketing team with our product goals is crucial to effectively connect with our intended audience.

Absolutely Neha, it's a valuable insight to consider retention as a form of repetition and continuously communicate the product's capabilities and use cases to users. Over-sharing and over-delivering to users, especially in the initial stages, can indeed set the tone for standing out in a crowded market.

Now, for those looking to delve into AI, particularly product managers scaling or transitioning to AI products, where would you recommend they commence their journey?

Yes! So, diving into AI product management is like starting? a thrilling adventure! To kick things off, get hands-on experience with AI tools. Don't just skim the surface – roll up your sleeves and actively explore what these tools can do. It's all about learning by doing!

In parallel, familiarize yourself with AI-specific terminology. Concepts like tokens, request rates, and other technical jargon are integral to the AI landscape. This involves delving into the intricacies of how AI processes information, handles requests, and generates responses.

To stay updated of the latest developments in the AI space, consider following thought leaders, experts, and practitioners on platforms like Twitter. This provides valuable insights into emerging trends, discussions, and advancements in the field.

Yes, I get it! I didn’t want to get too technical with you keeping my distinct network but one more question here. So, broadening our view a bit, how do you foresee the role of a Product Manager evolving with AI stepping onto the stage? Are we looking at a transformation in how PMs operate, especially with AI playing a tag-team role??

Honestly, Vismaya, I have been thinking about this a lot lately. When it comes to the role of Product Managers evolving with AI, I imagine AI handling a significant chunk of the technical aspects, like prioritization. With custom GPT models being trained on historical data and product backgrounds, AI can suggest the next moves, taking care of a substantial part of the prioritization process. However, PMs will still play a crucial role in gut checks and intuition, ensuring that decisions align with business goals and customer expectations. The human touch in cross-functional team management and program oversight will remain pivotal, complementing the efficiency brought in by AI. So, while technical cycles may see more AI involvement, PMs continue to bring their unique skill set to the table.

Agreed. I believe the decisions that a PM will take would change. The decisions PMs make would more of less be like choosing the right model or deciding on training processes, and may evolve into more strategic and nuanced considerations.?

Absolutely! Looking into the future, it's fascinating to see how AI is progressing. Take the recent example of the rabbit product unveiled at CES 2024. The concept revolves around creating an AI companion that deeply understands users on a personal level. It comprehends your preferences, interactions with apps, and can perform tasks on your behalf without the need for you to have a phone.

Sam Altman’s vision is to have AI, like GPT, eventually work so seamlessly with users that it can write applications, mimicking the user's writing style and preferences. It's like having a highly sophisticated assistant that not only understands your needs but can also perform tasks more effectively than you might do yourself.

The beauty of models like GPT-3 is their ability to continuously learn and refine themselves based on the data they receive. So, as users prompt and interact with these models, they contribute to the ongoing training and evolution of AI.?

I feel the continuous influx of data into AI models like GPT ensures that they keep improving over time. As more data is fed into these models, their accuracy rates inevitably increase. It's a proven fact that the more data these large language models receive, the better their accuracy becomes.?

However, the real challenge lies in ensuring authenticity and standing out amidst the clutter. With so many people using similar AI tools, the key is to differentiate oneself and maintain authenticity.?

This is my last and a question I ask everyone :?

If you had to choose one word that best describes you, both personally and professionally, what would it be? Also, is there a word you'd prefer not to be associated with under any circumstances?

Interesting question, Vismaya! I'd describe myself as either a learner or a multitasker. I thrive in constantly picking up new skills and adapting to various tasks simultaneously. You can throw anything at me, and I'll dive in to learn it. Well, almost anything – I did struggle a bit with those dense IIT-JEE books.

As for what I don't want to be associated with, that would be a settler. I'm not one to settle into a comfort zone, even as I age. I always want to venture into new territories and explore roles that might not even exist yet. My journey has taught me not to make decisions based on FOMO, and I never want to settle for what everyone else is doing. I crave the excitement of trying out new things and embracing whatever challenges life throws my way.

As I wrap up my conversation with Neha, I'm left feeling inspired by her journey through the dynamic world of AI product management. Neha's vibrant spirit, coupled with her knack for learning and ability to juggle multiple disciplines, have truly set her apart in this rapidly evolving field. Her insights into the intricate interplay between human ingenuity and AI capabilities offer a refreshingly nuanced perspective for those of us navigating the areas of AI-driven product development.??

What resonates most profoundly from my chat with Neha is the paramount importance of staying laser-focused on the end user, and the vital need to harmoniously blend uniquely human skills with cutting-edge AI tools and technologies. A heartfelt thank you to Neha for being such an open, friendly, and generous guest - freely sharing her wealth of experiences and hard-earned wisdom. Thanks a bunch, Neha - it's been an absolute pleasure! ???