Data Products Convergences !!

Data Products Convergences !!

Achieving an AI-enabled product: are we on track?


5 min read

Convergence of AI/ML models into product features has always been difficult for those of us working in the data and analytics space, thus we are unable to guarantee their incorporation into the business model. When we talk about data products, we mean apps that have been designed with data, analytics, and AI embedded into them, but all of them are restricted to operating inside the parameters of the training model and validating through the prediction pipeline. The business value development and capture is the point I intend to make here. Working with so many data scientists, engineers, architects, and AI/ML leaders (converged leaders post-data revolution), I've discovered that most people are primarily focused on developing solutions without a clear vision. It's not my intention to point fingers, and I don't want to assign blame, but the onset practice is a cliche when it comes to not considering the value of any effort.

I am aware that when it comes to analytics work, most sectors find that the data team ends up focusing on proof-of-concept experiments rather than ever having the solution implemented in a widely used application. According to a survey, the analytics/AI-ML solutions that are now in use fall under 0–20% of the successful models that have been deployed throughout the industry and are helping to improve the user experience within business applications. Leaving aside the newest LLM tools, it is still unclear whether they are generating commercial benefits. These are already in use, similar to another social network program, but we haven't yet discussed value creation and user experience. I restrained myself from using huge tech lingo because what I am talking about is obvious, at least to the audience present here...... :)

DPM "Data Product Manager"

If data product managers/leaders begin to dig deep, all of these may be improved. This process begins by keeping the vision and business value front and center and making the team aware of it. The user experience, the business model, and the workforce that is infused with the necessary technologies must be closely worked on by the team developing the enterprise AI solution to reach the production pipeline. With this in mind, the product manager or team members working on the analytics feature should be primarily aware of the needs and refrain from making assumptions or taking into account a generic analytics solution. We need to put the data product manager front and center when it comes to developing analytics features for any business data-enabled solution.

In my experience working with numerous product managers, I've seen that they envisage features that fall within the purview of the requirements backed up inside the application platform they are in charge of. However, when it comes to design features, the aim of analytics is not always clear. The data product manager must work closely with all the other product line managers and stakeholders from viable business units to consider the overall business goal and consider the analytics feature. For the next major feature convergence or enhancement in the enterprise software, DPM needs to align all the analytics priorities with other business unit priorities.

In a previous blog post, I reaffirmed that the task of data product management does not fit under the general ideas of product management, but that DPM must consider the analytics aspect for each integration being completed in the business application. DPM should start thinking proactively about the analytics strategy and data-enabled features and be more focused on important business use cases rather than waiting for product requirements from another business unit. They must continue speaking and taking part in future vision discussions while consulting with users and platform owners. The possibility of creating features both horizontally and vertically is now available thanks to analytics. On the aspects of horizontal and vertical analytics, I will write a separate draught. The emphasis should always be on value creation and capture for DPM, but this can only happen if we begin to map every single requirement related to data that can be used to provide analytical capabilities. DPM must also play a steward role in creating this success narrative. Always start with the customer when creating requirements, and then look at the current solution or application, how data is captured, what business processes are associated with the data, the overall workflow, and whether there is any opportunity to automate the current solution using analytics. DPM must have an analytical mindset; we cannot sound casual.

Any analytics solution's success story contains a solid and tested journey. A validated analytics model must be integrated into the current system, value creation must be verified in the business process, the user experience must be improved, and so on. If we want to deploy an analytics feature in the enterprise pipeline, this can only be done with meticulous preparation and supervision.

For any data leaders, it is equally crucial to remember that, despite the recent eye-catching developments in the analytics field, we shouldn't stray from our core competency of applying analytics to generate solutions. I've seen that recent leaders have a tendency to talk a lot without having a clear strategy or roadmap, but they always include a lot of flashy stuff from the analytics industry.

Big stories don't happen all at once, so we shouldn't be in a rush to complete everything at once. Instead, we should demonstrate little but significant value creation over time by integrating the analytics capability into commonplace applications. Leaders should avoid getting labeling their team as an experimental POC group by supporting their growth while ensuring that value generation is always a priority in all decisions made.

I can be reached on, happy to meet!!

Stay tuned.......