The Future of Product Management in The Era of Generative AI

It’s been a few months now since our last conversation on AI for product managers. As anyone on LinkedIn can attest, Large Language Models (LLMs) & Generative AI (GenAI) have peaked the interest of individuals and corporations alike. With 2024 on the horizon, we wanted to explore some of the current and future implications of GenAI for the work done by product managers. GenAI continues to attract investment and development, much like the evolution of natural language understanding and both supervised and unsupervised models in previous years. Notably, Google's recent unveiling of Gemini marks a significant leap in performance, surpassing that of Chat GPT4 on most tasks and? image input capabilities. X.ai's Grok’s real-time world knowledge appears promising. However we’re still in the early innings of GenAI. Much like historical improvements in computing, GenAI technologies will follow a similar trend and exponentially improve in the short term.?In this co-written blog, John L. & I discuss opportunities and changes GenAI will introduce specific to product managers.

For professionals in product management, the rise of GenAI marks a pivotal technological inflection point. This era demands not just awareness but a proactive stance in anticipating and preparing for rapid advancements in AI technologies. Being at the forefront of GenAI can be a career-defining opportunity. Embracing and mastering these technology positions opens doors to lead significant and positive change. In the near future, AI will eventually become a core PM competency and not just a differentiated expertise. In this article, we will explore the opportunities GenAI holds to improving PM productivity, product quality, as well as the dimensions and skills of empathy and innovation PMs need in this new world of working.

Considering the transformative potential of GenAI in product management, let's begin with with practical use cases?for integrating GenAI into your current product management practices. Additionally, we'll explore joint insights into how we think the evolution of GenAI will shape the future landscape of product management. Our hope is that this dual approach will equip you with both immediate applications and a forward-looking perspective, ensuring that you are well-prepared to harness the benefits of GenAI both today and in the years to come.

Practical GenAI Applications for Today's Product Managers

The most effective approach to understanding an emerging technology and its potential implications is through hands-on experimentation and exploration of its limits. For this reason, we’ve collected effective GenAI use cases that we’ve embedded into our own Product Management workflows focused on writing user stories. Using the newest Google Gemini version and other models that accept multi-modal inputs, here are six real-world tried and tested PM GenAI use cases? you can use today:??

1. As a PM leading initial discovery and research into a potential product, you can use GenAI To generate Industry/Persona Research Questions and even simulate answers. You can do this by having the GPT take the persona of the interviewee and pass it your questions for answering.

2. If you need help conceptualizing use cases for your product's roadmap, you can also use GenAI To Generate Roadmap Ideas (Including with new tech). As we’ll discuss later, this is effective for most items that are not disruptive innovation.

3. Even simple tactical work such as generating demo narratives and demo data for your pre sales environments is a great example where GenAI can provide more realistic looking data.

4. Writing user stories is a tactical task that takes a lot of time from PMs. With new multi-modal GenAI inputs, we can utilize GenAI To translate a screenshot of a design into written user story requirements. ?????????

5. To further expedite your story writing, you can also use GenAI utilizing a short text description to write detailed user story requirements. GenAI has proven effective in providing a starting point of different acceptance criteria given a simple short description. While it definitely still needs a human in the loop to validate it, it is a helpful starting point that can help you think of acceptance criteria you may have missed.

6. With supported multi-modal input, you can even use Utilize GenAI for ideas on how to optimize your existing designs. Working with your designer, you can upload screenshots of designs to identify common patterns for improvement. In this example we uploaded a UX design for an alert list and got reasonable feedback.

These simple examples show how we can improve PM productivity and drive quality in the way we build products today. While these reflect a few new use cases enabled by multi-modal GenAI, there are even more opportunities across the lifecycle that PMs should be mindful of, both today and as technology continues to improve.

Upcoming GenAI Opportunities for Product Management

Let’s analyze some of the additional key tasks in the product management lifecycle and the potential disruption of GenAI improvements on each:

Ideation/Discovery:

1. Primary industry/domain research with no additional insight or synthesis will become obsolete to GenAI derived insights. The knowledge gleaned from interviewing candidates will be readily available through GenAI and will offer persona, industry, and workflow level insights teams can leverage. It can simulate conversation from the perspective of relative personas, and interviews requiring finding specific practitioners will become a practice of the past for established industries or mature/late-stage products.
2. GenAI can help provide additional roadmap ideas based on current industry and competitor trends and will lead the PM to handle more of a prioritization and feasibility exercise than ideation. Innovative ideas, however, will still be on the PMs shoulders.
3. Product managers need to come up with GenAI use cases that improve productivity and usability of their apps, especially in software. GenAI itself can help uncover use cases for GenAI within your industry or domain. The highest value and easiest to implement activities however will still come from an AI familiar PM, which we forecast every PM will need to become.
4. GenAI will drive innovation and product development. Future Large Language Models, combined with Customer 360 data, will develop GenAI Digital Twins of the Customer (DToC). These digital representations will mirror human customers and identify needs and pain points.?
5. To train GenAI DToCs, customer interactions will increasingly be recorded and translated into text. This follows because Large Language Models need text format data for fine-tuning and training purposes. Given this, product management organizations may transition from PowerPoint-based product information slides to text-based product documents similar to Amazon’s PR/FAQ deliverables.

Execution/Development:

1. Engineering requirements will be translated from high level roadmap items to low level user stories through GenAI. PMs will clean up requirements as humans in the loop.
2. Initial designs will be generated, and existing designs will be optimized using GenAI recommendations. Designers will work alongside GenAI to produce brand/component limited designs.
3. Engineers themselves will write and modify code generated by AI and will see their own productivity increase by 20-30%.
4. Product documentation may become ‘self-documenting’ through multimodal image recognition that can write what it sees and add context on who its for. Doc writers can then simply modify and add to this baseline definition.

Product Market Fit/ GTM:

1. Dotcom language, value propositions, and any marketing text (display ads, search ads, etc.) will have an opportunity to be generated and optimized by GenAI, along brand guidelines and tone.
2. One core component of Product Market Fit is the value hypothesis, which can be defined by the following questions: What will you build? Who is desperate for it? How will you deliver this value? GenAI will help Product Managers accelerate the validation of the value hypothesis by:
3. What are you going to build??Technology industry homeruns, like Amazon AWS, OpenAI, Microsoft, Uber, Facebook, Salesforce, and Airbnb, stem from unique insights based on technological breakthroughs. Innovators typically notice new technologies, tinker with them, and then find a desperate market and suitable use case. We call this the Swing For The Fences approach. It results in many failures, but when it works, it works big.On the other hand, the most popular product development method begins with the customer’s problem. Afterwards, the product manager then finds the best technological solution for the customer’s pain. We call this the Working Backwards approach. This tends to result in incremental innovations, small but consistent successes, and can arguably scale more efficiently in large organizations since product managers don’t need to be experts in emerging technologies (the upshot is a considerably larger talent pool).I bring up these two approaches because the future of GenAI seems poised to make both methods radically more accessible. For example, in the Swing For The Fences approach, one significant blocker is having access to and understanding the latest technological breakthroughs. GenAI offers a solution here. Imagine a Large Language Model, continuously updated and trained on recent research papers and capable of summarizing new technological developments understandable by non-technical professionals. This would enable more Product Managers to replicate the Swing For The Fences method and increase the number of homerun innovations.Furthermore, GenAI can automate the Working Backwards approach. Namely, GenAI DToCs will scrape the internet automatically, incorporate customer support ticket conversations, ingest customer interaction transcripts, and conduct competitive product research to distill customer problems and recommend incremental feature innovations to product managers based on features that already exist in the market. The risk is that most products will have similar features, and a competitive product differentiation moat will be challenging to maintain.
4. Who is desperate for it / How will you deliver this value? Product Management and GTM organizations will create GenAI DToCs of customer personas separated by firmographics and industry idiosyncrasies. Product Managers will test new products, pricing, and value-delivery methods on these DToCs to identify overlooked customer segments, validate demand, and quickly test the value hypotheses.?
5. PMs will continue to focus on GTM and identify the story of what they are building, why, and for who. They will also influence pricing and packaging to identify the maximum value to extract.

Product Maintenance/ Expansion

1. After a product begins gaining traction within the organization, there generally comes a 'enablement bottleneck' where expertise is still found with the product org and not yet diffused to sellers, support, and marketing. GenAI has a tremendous opportunity here to generate 'virtual product managers' that have been trained on an added data set of all the webinar and demo text transcriptions, product strategy documents, and documentation to become a product expert and help answer questions from the internal org.
2. Once a product is launched and adopted by customers, GenAI can offer an interface to effectively surface insights on your products performance and data, taking different usage and analytics data and offering simple business summaries. Basic status updates on adoption, etc. should become irrelevant.
3. GenAI can recommend opportunities to expand a product into adjacencies based on the current capabilities. The PM will still need to identify market sizing and priority, as GenAI currently does not have a strong sense of numbers.
4. GenAI can offer summarized insight into customer feedback related to your product area or application.

While these and the prior use cases reflect “human in the loop” scenarios where GenAI can help improve product manager productivity, there are already interesting experiments around a fully AI run organization – where bots handle the design, requirements, development, and iteration, with no human in the loop. We can call this concept the “AI Software Factory”. While it’s unlikely such a future is in the next few years, GenAI assisted roles in each part of the product development (design, QE, eng, infodev, product, marketing) will absolutely manifest in the 2-3 year horizon. GenAI will not immediately replace roles, but it will begin driving productivity improvements that may lead some cost focused organizations to ‘do more with less’. You can find an interesting exploration of this concept in the following video about an ai workforce.

Navigating the Future: Implications of GenAI for Product Managers

With so much of the execution becoming automated and generated, what does this mean for our craft? With so much that can (and will) be automated in the product life cycle, it’s time to double down on three elements of the role.

Customer Empathy & Trust

Now more than ever, the notion of empathy becomes critical in customer meetings and conversations. As a product manager conducting discovery on a customer, or collecting feedback on a launched product, empathizing with the customer's concerns and pain is key. Whether it’s a roadmap request or a defect report, making them feel heard as a human being instills confidence that your product will improve their quality of life and that you are not a vendor trying to sell something for your own benefit, but rather a partner invested in your customers' success. Forming longer term customer relationships where you provide reference examples of roadmap or bug fixes you’ve delivered will help establish credibility and trust with the customer.

Additionally, the core responsibilities of the PM will double around identifying the strategic roadmap of “what” to build “when”, as well as identifying how to market it. The more tactical elements of execution in terms of “how” to build it will become automated. PMs will double down on storytelling for both customers and their teams, identifying the pain points and trends aligned to the needs of the market, at a price point the market would be willing to pay.

The same empathy and trust extend to the teams and cross functional design & engineering resources you’re working with as a PM. Your focus shifts to vocalizing the ‘why’ behind what you are doing, and ensuring the team stays engaged, excited, and efficient. It also focuses on transparency on process and continuing to collect feedback from your team to ensure your GenAI infused outputs are at a high enough quality for the team. In this matter, building a community within your customers, peers, and delivery teams is something you can do to encourage human connection and trust.

AI Upskilling & Competency

To remain competitive, Every PM will need to become, at a basic level, an AI PM. AI is no longer a specialty or differentiation point for products, but rather a table stakes requirement. AI fluency is a core competency of modern product management. AI reflects a new technology trend that companies will need to use to stay relevant in the market. The user benefits and outcomes enabled by Generative AI, and other AI technology promises are non negligible and past the point of simple proof of concepts. AI and data management is no longer a niche or moonshot but rather a core competency of every business. PMs are the champions of their domain, product, and user context, and should know where the largest pain points or the most valuable decisions are in their products' workflow that AI can automate. Understanding all the AI technologies available (supervised, unsupervised, neural nets, image recognition, generation, etc.) ensures you are using the best tool for your customers problems. If you aren’t doing it with AI, there is a startup out there that is.

Remember - AI will only get better from here. So while the ‘how’ may change technically, the fundamental needs of “Why” and “What” we are building is important to uncover as it will endure even as technology changes.

Disruptive Innovation & Critical Thinking: ?

Today, a PM oversees bringing a level of subject matter/domain expertise to the table to help ‘build the right thing, at the right time’. The reality is that Industry Knowledge and ‘domain expertise’ will become a commodity. GenAI has read every piece of text, every research article, and its expertise is bounded only by what is currently known. Engineering teams could probably leverage GenAI to stand in for low level decision making of a product manager already today.

With existing knowledge commoditized, disruptive innovation will rise in value. GenAI models are excellent at generating responses from known information & text, but it will be difficult to come up with responses for technologies or methods not yet known. An example around this is the classic horse manure problem of London in the late 1800s. Imagine being a PM currently and posing the question: How do we solve the city’s horse manure problem the next 100 years as the population increases?

As we can see from GenAIs reply, It would not be able to recognize or identify specifically that the automobile would be a potential solution, however it can propose solutions that will lead you down the path of developing the automobile (e.g. exploring alternative transportation).

Clayton Christensen, a Harvard Business School Professor, distinguishes between disruptive and incremental innovation. Disruptive innovation is when an inferior product (compared to the existing market-leading product) is introduced to an overlooked and desperate low-end market segment. Initially ignored by established companies and large customers, this inferior product is widely adopted by low-end market customers, continuously improves, and moves up-market, eventually overtaking leading firms and products. This occurs because incumbent firms prioritize incremental innovation. They refine their existing market-leading products based on the demands of their largest customers. Over time, this focus on improving the core product for only the high-end segment alienates lower market segments. The product becomes overbuilt, expensive, and sometimes bloated with features. Ultimately, this creates a beachhead for startups with disruptive innovations to emerge and expand.?

With the emergence of advanced GenAI DToCs, product managers specializing in incremental innovation will become less essential. This is due to GenAI DToCs' ability to continuously integrate new customer and competitor data, focusing on existing products. However, the skills needed for disruptive innovation will become more valuable. Thus, product managers should concentrate more on identifying opportunities for disruptive innovation.

While AI will commoditize knowledge and basic reasoning, the decision making associated with actual complex & abstracted understanding, comprehension, application, critical thinking, and problem solving related to the product and its go to market will remain with the PM in the near term. Which customer feedback to prioritize, navigating the politics, processes, and dependencies of corporate product management are nuances an AI won’t be able to perform today. How to best approach solving a problem will remain a team activity between engineering and product considering size, scope, and business priority. Focus will grow on hiring and developing PM talent that can efficiently access the 3rd and 4th levels of Webb’s depth of knowledge, as simple Recall and Concept can be handled by the AI today.

?

Conclusion – The Future of Product Management in the GenAI Era

To hypothesize on the future of the product manager role, it's valuable to reflect on its past and where product management gained its start. Product management as a formal business role and function has been widely attributed to the “Brand man” concept pioneered by Proctor & Gamble. These individuals were responsible for identifying product strategy, advertising, and promotion through field interactions with customers. Fundamentally, they were honing the GTM for consumer products.

These “brand men” were storytellers – connecting the needs and pain of consumers to the value their products provide. As more of the tactical portion of the product management process becomes automated, the strategic vision and storytelling associated with product market fit will once again become a top priority of every product manager.

If you look at the trends of organizations like AirBnb & Apple, they have already shifted their product management organizations back in this marketing direction. With GenAI use cases automating the execution of building a product, it’s important to reflect on the fact that the most of a PMs time will inadvertently shift time and focus right to the GTM function.

It feels the future of product management is a shift back to focusing on the 'what' & ‘why’, evangelizing the brand and building connection and trust with customers and teammates. Focus will recenter around roadmap prioritization, complex and abstracted problem solving, disruptive innovation, and GTM and storytelling. PM resources will become more efficient in automating product development and execution, and basic domain expertise will become a commodity. Teams and organizations will have a reckoning when they determine how to handle resourcing and job responsibilities in an age of increased productivity. Most importantly, PMs will need to upskill GenAI to understand ways they can add productivity, quality, and UX enhancements into their products to help their organizations remain competitive in the market.

Keep in mind these views are based on our understanding and the perceived capabilities of solutions like Artemis, Grok & ChatGPT 4 today and in the upcoming year. What does the world look like, and our role as PMS? in a world with ChatGPT 10? ChatGPT 100? Where and when should and will the hammer of regulation or caution swing? Are we to be worried or excited? Only time will tell.

?

Wishing you all the best to the new year,?

Darius Koohmarey & John Lee

Disclaimer: No AI Was Used in the Generation of This Post