Over the past four to five years, there has been a push across industries to incorporate Artificial Intelligence technologies into enterprise scale product offerings. The gravity of this effort can feel overwhelming to organizations at first due to the complexity of the technology and the ever-pervasive problem of identifying which existing business problems can be best solved with AI. This leads to organizations being stuck in a long and repetitive cycle of constant experimentation, without clear results to show to stakeholders. At a fundamental level, this constant cycle is often the result of treating and managing enterprise AI efforts as projects, and not the products that they truly are. AI efforts aren’t things with a set timeline and budget, they’re either external or internal offerings to real users that could range from a fraud detection system to a supply chain optimization tool, and they will grow and evolve over time.
By appointing a product manager to lead these efforts, organizations can avoid the common pitfalls associated with the application of artificial intelligence to enterprise data. A product manager can ensure that AI applications are solving true business needs, and that they are engineered in a way that is usable across the organization, even by those who are not necessarily familiar with the underlying technology. This is especially important for AI projects due to the technical complexity of the implementations and the amount of requirement translation that needs to happen between business stakeholders and technical developers. Including a product manager will ensure that your organization realizes value quickly, and pursues MVP implementations such as basic regressions and classification algorithms to ensure your investment in both people and technology is appropriated correctly.
What Does a Product Manager Do?
The product manager is there to ensure that the user is represented both in the design of the overall experience and the performance of results, yet they also represent key stakeholders and ensure the solution fits the market’s needs and delivers on revenue goals. While project managers ensure solutions are delivered on-time according to specifications, the product manager will drive the solution to fit an ever-evolving market, apply new methods and technologies when appropriate, and optimize the value the business actually receives from the adoption of AI technology. In my time as a product manager, I’ve followed the above values to create solutions that optimize square footages for real estate, structure search content for research purposes, and leverage large datasets to determine optimal ESG recommendations.
Advantages of Having a Product Manager
Throughout this article, I will use my experience as a product manager within the enterprise data space to focus on the specific advantages a product manager gives to the application of AI technologies. I hope this experience gives you the ability to think about your enterprise AI with a more product-centric focus, and clearly understand what is needed to have a truly successful AI transformation within your organization.
Identifying Appropriate Problems
The first stumbling block most organizations run into when implementing AI at an enterprise scale is making sure they have the correct problem identified, which is a matter that can be solved in an appropriate amount of time. AI is complex, and isn’t something that can necessarily be achieved in a short period of time. However, a product manager can alleviate this problem and lead the process to define a minimum viable product (MVP). This MVP will only contain what is immediately valuable to your organization, and allow you to iterate quickly in order to gather feedback and tailor your solution to your user’s needs. Not only does this help keep initial costs down and cost overruns from occurring, it allows your organization to remain agile and pivot to other solutions and technologies if necessary. A good product manager will also help host demonstrations of development progress and create a clear roadmap for the future. This way, your organization can clearly communicate to stakeholders what features will be included in the future and how the technology will evolve over time.
The next step a product manager can take in helping your organization identify an appropriate problem is to lead the creation of a prototype that can be placed in front of end users. This prototype doesn’t necessarily need to be a full artificial intelligence solution, it could be something as simple as a linear regression. The main point of this prototyping effort is to ensure that what is being built, and the data that is being leveraged, is targeting the correct group of end-users. It is solving a clear issue for them in a simpler way than they have available today. This process will also allow your organization and the development team to identify if they have the appropriate data necessary to solve the problem and if it’s structured to be used in an AI application. This helps in projecting future costs, both in the time it will take to structure data for a full AI implementation and if there are any additional data sets that need to be purchased. If your organization is failing to succeed at the prototyping stage, or you’re projecting high costs for work to continue, this will give you the fastest and a less expensive way to learn early and adjust the product roadmap.
Correctly Applying Innovation to What’s Valuable
One of the biggest values a product manager can bring to an enterprise AI effort is to have a firm understanding of the organization’s data and to elevate conversations with different business groups that clearly demonstrate the organization’s problems. Without these understandings, an organization could put months or even years of development and investment into data exploration, taking up an unknown challenge. It’s not only important for an organization to be able to understand its problems, it’s absolutely critical for them to be able to prioritize them as well. Executives can often hear multiple different problems from different business groups, and have difficulty prioritizing which problems to solve. The product manager is the key member of your organization that will be able to help you identify which solutions will alleviate the most headaches, and generate additional value for your business, ensuring that your investments are being spent in the correct places.
In order for a product manager to make the correct decisions and prioritize appropriately, they need to be able to identify and communicate the organization’s strategic objectives, as well as establish metrics and KPIs that will be used to measure performance. These are imperative to establish ahead of time as it will allow the organization to filter a set of valuable ideas down to the ones that most align with existing objectives, and the ones most likely to meet performance goals. Another key method a product manager will use to prioritize valuable problems is to identify groups that are excited to pursue AI technologies and have data available to use. By filtering problems down to those that help your organization meet strategic goals, and allow collaboration with groups that are inherently excited to use novel technologies, the product manager enables your organization to shift priorities and pursue those challenges that are truly valuable, and make sure that innovative technologies are applied appropriately. This filtering strategy can be shown in a simple funnel as shown below.
Designing a Solution That is Familiar to Your End Users
In your organization’s pursuit of enterprise AI, the product manager will be the main individual who directs how the product grows and how it gets exposed to users. It’s extremely important to design a solution and user experience that will resonate with end users before development begins, to ensure long-term success and the adoption of the solution. Although AI may power the solution, end users who are not familiar with the technology may not care. At the end of the day, they want to see a solution that presents answers to their problems in a clear and concise way, no matter how the answers get created. Product managers can deliver this by leveraging teams of designers and front-end developers solely focused on providing the most appropriate end user experience. Their cross-functional knowledge on how to lead design sessions, define use cases, and create wireframes will prove invaluable to ensuring that the end product delivered to users is familiar and usable.
A product manager will also bring an inherent knowledge of tools that can be used to facilitate conversations and ensure everyone is in agreement on the end result. On one of my more recent projects, we were able to utilize Adobe XD to walk users through different wireframes and mockups, allowing them to see deliverables before they were produced and provide feedback early on in the process. Additionally, we used an online whiteboard application to present roadmaps and guide end-users to document the problems they were facing and the information they needed to solve them. By allowing your product manager to utilize these facilitation tools, you ensure every member of your organization is on the same page, and can address issues in communication of ideas early on in the process. This guarantees that as experimentation and development of your enterprise AI solution continues, the predictions or classifications that are being made are presented to your users in a way that is easily consumable and can be leveraged to generate additional revenue for your organization.
Handling Uncertainty and Communicating Product Value
Enterprises will often face problems not seen in smaller start-up companies when pursuing the creation of AI products. Having a product manager will make addressing all of the common enterprise sources of uncertainty easier, as your organization will have a key point of contact that is knowledgeable and responsible across the AI solution. That means that your product manager will be a go-to resource for solving problems with the size and complexity of data, as well as dealing with outdated or non-performant infrastructure. A product manager will be key in breaking down walls between departments to ensure alignment between business and IT stakeholders. Eventually, their efforts will lead to a more cohesive data strategy, with a focus on how AI will produce real change at your organization. As the product continues to evolve and grow, they’ll help mitigate future uncertainty by being an advocate for an agile software development lifecycle. In the long-term your organization’s engineering managers will continue to be focused on performance, and project managers will continue to focus on delivery, where in contrast the product manager will be focused on how your organization’s adoption of AI can continue to grow and the existing solution can be applied to new and varying use cases.
Once your enterprise AI solution has moved past the initial stages into development or production use, you will need a member of your team to communicate the solution across the organization. For example, the manager of a recommender system tracking possible changes to a company’s ESG policies will have all of the necessary information to make sure that accurate information about how the new AI system will influence upcoming policy decisions. They will make sure this information is conveyed throughout the company, as well as the true business relevancy of the AI product. They will help ensure that both your sales and marketing teams have a deep understanding of the power of the AI solution that has been created, so that your business can deliver clear outcomes to a variety of different users. They’ll also work toward creating rich documentation and training materials around the solution, so that your new users are enabled to get the most out of the AI application that has been developed. A good idea of where your AI product manager should fit amongst your different teams is shown below.
New AI technologies and their applications to business problems continue to evolve, especially as more and more large enterprises begin AI adoption. The challenges normally faced in scaling an AI solution to an enterprise level can be mitigated or even fully addressed by staffing experienced product managers as champions for these applications and the larger vision. They will help ensure the solution solves the right problems for the business, that the product grows and is adopted successfully, and that end-users are enabled with the knowledge to utilize and receive the maximum amount of value from the product. Ultimately, this will lead to the successful adoption of AI across many different business units and use cases. If enterprise AI is built with a product manager leading the way, it can truly deliver on helping your organization achieve its strategic objectives. Here are specific examples and case studies on how we have been approaching this:
- How to Take an AI Project Beyond a Prototype
- Building a Semantic Enterprise Architecture
- Enterprise Artificial Intelligence
Could your organization use a product manager to help guide the way for your enterprise AI transformation? Contact us today!