AI Data Products: Blurring the Lines Between Engineering, Design, Data and Product

In the last 2 months, I haven’t been able to open social media, turn on TV, or even go to a meeting without hearing about AI. Hype? Sure. We’ve seen lots of hype cycles in the last decade. But this feels different. I think that’s partially because these Large Language Models and the data products built on top of them feel a bit like magic. It is also because I see the way these data products can change how we work. I’ve toyed around with creating working prototypes in the last couple of months, and its nothing short of remarkable.

Specifically, I think these AI products can shake up the way we think about product, design, data and engineering and more broadly, how product development works across tech and all industries. Specifically, I think that using these AI data products, we can all participate more actively in product building, creating a lot more grey area between what are traditionally distinct areas of practice.

Let’s look at a few company archetypes and the ways in which this change might take place.

First, consider existing product led companies. In tech, we typically have distinct areas of practice like engineering, design, product management and data. Each of those spaces have fairly well carved out roles for how they participate in the development of a new product or the iteration of an existing product. Because of the specialized knowledge involved in each domain and often the scale required, this works pretty well! In this space, these AI data products enable each of these domain areas to participate in product design. Ideas not only can start anywhere, but they can potentially get to a more impactful prototype before it requires involving cross functional teams. I think this can produce more ideas, more quickly and majorly impact the pace of innovation in already digitally innovative companies.

Second, consider companies that are still in the midst of a digital transformation. It is easy to forget that even though we talk a lot about digital products and optimization, there are many companies that have just started this journey. It is far more unlikely that they have well defined areas of practice that contribute to product led growth in the way companies in the first archetype might. Traditionally, these companies need to overcome inertia that comes from not having a mature product creation, iteration and improvement organization, at least when it comes to digital experiences. It can be hard to hire the right talent and retain them and finding consultants that can step it can be just as challenging and costly. Now imagine a world in which these companies can prototype their ideas without having to establish a full fledged team. I think it can flip the traditional idea from hire a team to build, to build a prototype before hiring. As with the first archetype, it can majorly impact the innovation in these companies and hopefully make trying out new ideas more cost effective.

Third, consider an individual who is just trying to get started. I know so many people with ideas and they see the blocker as needing to find a technical co-founder or a designer or product person to get started. Those roles are clearly still important, but the stage at which they need them may change. I’ve already had conversations with early stage founders who are trying to find product market fit without going the traditional route of hiring builders and designers first. They are starting simple, with basic prototypes built on top of these AI data products, and seeing how far they can get. Now, this doesn’t mean that they can entirely skip steps of a product development process. But I do think it means they have the potential to reorder the steps it takes to get a product idea off the ground with a single person with a good idea.

If you haven’t played around with some of these LLM data products, I encourage you to just give it a try. I’m still surprised on a daily basis on what they can do, and how I find myself thinking about my role in building things differently than I did just six months ago. As a people manager, I’ve been able to write more code and design more prototypes than I did in the previous few years combined, just because of the efficiency of getting ideas off the ground quickly.

I don’t really know where this is all headed, but I am confident that the lines we use to define product, engineering, design and data will blur, it is just a question of how much and how quickly.