Using product management to bring a strategic view to data and AI operations

While early wins in AI and data are great, scaling across an enterprise requires a completely different approach.

As organizations begin to take advantage of the transformative power of data and AI, some bright spots are emerging. 56% of respondents in a study by McKinsey recently reported that AI had been successfully adopted in at least one area. This is up from 50% of respondents in 2020.

It’s a great thing. Yes, that’s great news. But now comes the difficult part: switching to a product-management mindset to effectively scale these functional quick wins at an enterprise level while establishing what Forrester refers to as the insight-driven business.

It is much easier to say than do, as experience has shown.

Signs that Your Data and AI Operation Needs a New Approach

The data and AI outputs do not cross organizational boundaries

Most organizations started their AI and data journeys within functional areas and continue to do so. More cross-functional data is needed for greater maturity.

Data and AI products that are mature tend to draw on data outside of the functional area to gain maximum insight. Similarly, reusable components like customer propensity scores or product-related pipelines are becoming more common across teams and at the corporate level.

A global SaaS company cited that its marketing department had three models in production. These models delivered significant value for lead generation, but they were not deployed, nor were the features of these models for any sales use case. Following some further discussions, it became apparent that these models could be used in sales, but sadly, this was a missed opportunity.

You are stuck in a dilemma of ‘build or maintain,’ and you have an excessive rotation toward prototypes.

Technical resources with high levels of skill tend to prioritize “new build” rather than “maintain.” Still, they lack enterprise-level priority mechanisms that optimize the value for an organization over time. It cannot be easy to overcome this challenge without solid data operations and model operations processes. However, building up that muscle takes time. To balance the two ends, new resources may be needed with a focus on AI and prototypes.

Enterprises have widely adopted AI models that are redundant or highly overlapping.

These tools are useful for gaining a better understanding of what is available today. However, they don’t help to reconcile the data and identify reusable components that benefit the whole enterprise.

For organizations who are beginning to see the results of their data and AI investments, a standard process that takes into account and rationalizes enterprise-wide needs is a common problem.

Enterprise ownership of AI and data products is not clearly defined.

Data governance councils, as well as individuals, focus on the quality of data and establishing necessary controls. However, there is still confusion about ownership and investment strategies for outputs.

CIOs and CDOs have a broader scope of responsibility, which means they are better positioned to lead change. However, their teams lack the skills and frameworks to do so.

Product Management: Bridging Gaps Between Operational Priorities and Enterprise Priorities

After establishing data operations and AI model ops capabilities in an organization, it is time to think strategically about these organizational assets.

Forrester defines data products as a “component that ingests data and delivers it to an insight solution used for decisions and actions.” This includes both the outputs of data ops and AI models ops practices.

Organizations that adopt a product-management mindset and apply tried-and-true principles of product management to these valuable assets will reap greater returns, maximize their technical resources, and quickly outpace their competitors.

To clarify, product management principles effectively…

  • Give a strategic perspective on the business. Shared services are commonplace in large and medium enterprises. However, the shift to product management puts the focus on the outputs from these services while also providing the ability to manage requests for information. It is important to align data and AI services with the business by using the concept of products.
  • Create a cross-functional mindset through products that lead to asset inventories and reuse. These, in turn, benefit the whole enterprise, not just single functional groups, while enabling more advanced outcomes. These connections often do not come naturally without a framework that unifies the work, such as a product.
  • Facilitate the use of a standard and consistent prioritization system. Many organizations today have radically different levels of maturity in terms of insight-drivenness across all functional areas. Companies that align these pieces into an advanced insight-driven company have an 8.5-fold higher chance of reporting at least 20% revenue growth year-over-year. Product management can help make decisions that will allow insight to be distributed more evenly or that focus on specific areas.

Activating Product Management Strategy

Follow these steps to fully reap the benefits of an approach that layers product management on top of model and data operations.

Use the Boston Consulting Group Growth-Share Matrix

This analytical tool, first introduced by Bruce D. Henderson in 1968, has been used for decades as a planning tool in branding, product management, and strategic management.

This matrix helps to illuminate the outputs of service teams and facilitates the correct discussions with leaders.

As part of their quarterly business reviews, one technology client included a review of the growth-share matrix of each functional group. This led to “dot connecting” conversations, which helped align everyone on enterprise-wide priorities and resource allocation.

Invest in product catalogs with road maps.

The next logical step is to ensure that products can be searched and that their investment plans for the near future are accessible through road maps.

Simple, engaging road maps can provide insight into timing and help bring products to life. These road maps should be a key part of annual and quarterly planning.

Create a training plan for product management.

Numerous local business schools provide product management courses in all shapes and sizes. It takes time and dedication at all levels to instill such a mentality. A formal training program signals leadership commitment to the framework and gives employees the skills to succeed.

Use the internal data product suite as an incubator for external data products.

Many organizations are finding new business opportunities with the products they use and create internally. Do not let the opportunity to monetize AI and data products pass you by while you’re busy running your business. CIOs who are innovative and CDOs who have a strong focus on business can often be the best-positioned to lead these types of transformations that will help companies scale and gain enduring competitive advantages.

Leave a Reply