Product Operating Model: The Crucial Enabler for AI Success

Speaker A: Foreign. Product Management Podcast over the past few years, many product teams have built AI fluency and have explored how AI can be applied to their work and products. Now, that's a great start, but it's not enough to fully leverage AI in product management. Why? Well, without the right product model in place, AI delivers only limited benefits. So in this episode, I'll explain how you can create a product management approach that is optimized for AI. To get started, let's briefly reflect on the current state of AI in product management. AI has certainly changed the way many product managers and product teams work. But to which extent exactly, that's hard to tell. The pace of technological developments and the diverse nature of our profession make it hard to come up with a definitive description. But my experience, combined with several research reports, suggests that AI adoption and product management has so far focused on three main areas. The first one is product discovery and product delivery. Product teams have used AI mainly in product discovery and delivery. For example, product managers have experimented with tools like Lovable and CLAUDE design to replace written requirements with prototypes. And they've used tools like OpenAI codecs and Claude workflow and Claude Code to automate tasks such as data analysis and document generation. The second area is tasks. Product managers have applied AI to their existing processes and used it to accelerate tasks. For example, teams use ChatGPT or Claude to come up with product ideas and Figma or Miro to generate user stories. This has led them to add AI to their current product operating model, leaving the way product management is practiced largely unchanged. And the third area is tools and technologies. AI has made design and coding much more accessible. And product management groups have invested in tool oriented AI training. And this includes helping people learn how to wipe code, design evals, use retrieval, augmented generation, and create and orchestrate agents using tools like Claude workflow and code, which I've already mentioned. Consequently, the product manager role has started to change. Some act now as product builders who carry out work traditionally associated with design and engineering, such as creating prototypes and in some cases, even production ready code. Now, applying AI to an existing product operating model and integrating it into the current ways of working can be helpful to get started. But it is not enough to truly leverage AI. You have to redesign your workflows and create AI native processes. Now, this makes total sense in my mind. If a company centralizes decision making, runs projects, uh, staffed with specialists, uses linear waterfall based processes, and employs detailed feature based plans, it's hard to see how AI can help product teams create significantly more value, faster and ideally at a Lower cost AI is no magic bullet. To reap all its uh, rewards we have to go further than just using the latest tools, learn how to wipe code and build agentic systems. AI innovation has to be complemented with process innovation. Product management work has to be redesigned for AI. However, there's more to it. By now most companies have established product management to some extent, but not all product transformations have been completed successfully. Recent research by product led alliance and Product Plan shows that only 20% of organizations are fully product led. All others are partially product led. That's 37%, 28% are uh, transitioning from project to product and 14% are still project based. Now this matches my experiences of working with a wide range of clients in different industries. And as a consequence many companies operate somewhere in between the new and the old world, between a traditional project feature and a modern product outcome based approach. And it's therefore no surprise that Pragmatic Marketing's uh, current State of product teams report found that ownership is still high for output, low for insight, and that product decisions are uh, most often truly driven by executive mandate. To put it differently, many product operating models, the way companies practice product management, aren't as good as they could and should be. In some cases, product management is just theater. Businesses pay lip service to product management but still operate very much in the project feature paradigm. Common signs that this is the case include the following seven Product development is still organized as projects and a uh, clear shared understanding of what a product, especially a digital product, is, does not exist. Product managers are project managers in disguise who lack the right skills and empowerment to be effective product professionals. Teams are loose collections of specialists, not tightly knit units who collaborate and take shared ownership of outcomes. Feature based plans dominate and teams are held accountable for delivering features instead of meeting outcomes. Senior stakeholders decide what happens with a product and what features are added, not the product manager and product team. Product development processes are uh, linear with milestones and handoffs instead of being integrated and adaptive. And finally, a head of product does not exist. And if the role is filled, the individual is not a member of the executive leadership team. But they might report, for instance to the cto, the Chief Technology officer, or they lack the relevant experience. Now if a company's operating model is ineffective, just bolting on AI will achieve limited benefits. Chances are that more output, but not more value is generated. In the worst case, it's garbage in, garbage out. Now let's make this more concrete and look at two strategy and user stories. If a company determines a product strategy for a 12 months time frame and expects it to be executed as it was originally defined, then this is no longer effective. In the age of AI, given the amount of uncertainty and change that we experience with regards to customer expectations, the competition and technologies strategy is best understood as being fluid and adaptive rather than fixed and definitive. Consequently, the product strategy has to be frequently reviewed and adjusted to address emerging opportunities and threats early on. And this requires the introduction of a new continuous strategizing workflow as well as empowering product teams to make strategic decisions for their products. And I explain this in more detail in the episode Continuous Strategizing. I'll put the link in the show notes similarly, if product managers currently write detailed user stories and then pass them on to development teams, using AI tools to automate story generation is just not effective. Why? Doing so would simply automate a wrong process. User stories were invented as a fast, lightweight and collaborative alternative to traditional requirements. A, uh, user story consists not only of a narrative, the story itself. It has to be complemented by a conversation that takes place between the person in charge of the product and the people building it. Together, they discuss each story and they may even co create them. If you now automate a wasteful, broken approach using elaborate stories without conversations, you might lighten the workload of the product people and free them from having to act as story scribes. But you are unlikely to create better products. A more effective approach is to change how features are discovered and described, having the product team members, including the product manager, carry out the work together. But this requires adapting behaviors and processes first before automating them with AI. As these examples show, there is a real opportunity and a real need to reflect on how you practice product management, including the roles and responsibilities and processes you employ, and to purposefully strengthen and adapt it so you can fully leverage AI. Now, where does this leave us? What can you do to build a product operating model that is truly optimized for AI? My recommendation is start by objectively assessing your current way of working. If your product model is only partially realized or if there are areas that need significant improvements. For example, product managers aren't adequately empowered, product teams aren't properly staffed, and product people and engineers don't collaborate, then address these issues first before you optimize the model for AI. Next, identify the areas that need to be redesigned to fully leverage AI. For instance, you might find that you have to change your strategy workflow and embrace continuous strategizing rather than fixing the product strategy for the next 12 months. Or you might consider evolving the product manager role towards A product builder to help you with this, guide your work and identify the right improvements, use my product operating model framework. The framework consists of four elements organization, people, processes and tools. And uh, here are some sample questions to get you started. For the element organization is the effort to establish an effective product operating model, uh, supported by executive management? Does a product management group exist and is the head of product role filled? Is there a clear shared definition of what a digital product is? And does the head of product lead the product operating model improvement or redesign effort to get the people aspect right? Address the following five questions. Are the product managers adequately empowered? Do they have the final say on decisions, especially strategic decisions for their products? Do they have the right skills to do a great job and maximize the value of their products? Do they have regular access to users and customers? Do product teams exist? Do they have the right members and decision making authority? And are uh stakeholders effectively engaged? For the processes element, ask are ah the right workflows in place? And this includes product strategy, product discovery and product delivery workflows, as well as a product portfolio management process. Then ask, are they free from waste including delays, handoffs, task switching and loss of knowledge? And finally, do the right people engage in the workflows? And for the tools element, answer the following. Are, uh, the right frameworks used? Especially are an effective goal outcome setting framework and a strategy framework in place? Are, uh, the right AI tools employed? Do they help the product people do a great job and are they uh aligned with company wide AI standards? Are the frameworks and tools used effectively? Are they fully integrated into the relevant processes to improve your current product operating model, address the organization and people aspects first. There is no point in optimizing processes and introducing new fancy tools if the product managers completely lack empowerment and the relevant skills and if there is no head of product. Note that the questions that I shared earlier are uh, intended just to be a rough guide and serve uh as a starting point to help you identify key issues. Product management is context dependent. That's very important to note. There is no one right way to practice it and different companies will employ different product operating models. Consequently, you need to determine what model is right for your organization and what specific measures are required to implement it. If you work for a large enterprise, you might even find that different business groups will implement the model differently depending on the type of products they build, the markets they serve and the organizational setup they use. And finally, don't forget that no product operating model is perfect. The work to improve it and maximize value creation is never truly done. You should therefore continuously aim to enhance your product management practices and regularly inspect adapts the model even after AI has become old news. I hope you found this episode helpful. To learn more, attend my workshops and read my books how to Lead in Product Management and Strategize. Reach out if you have any questions or need help assessing and improving your product operating model. And thanks for listening.